# Alloovium Documentation > Complete product documentation for Alloovium — AI-native document intelligence and > compliance for the construction industry. Index of individual pages: https://www.alloovium.com/llms.txt ## Documentation Source: https://www.alloovium.com/en/docs All systems operational # Help & guides Guides and answers for getting the most out of Alloovium — from your first upload to project-wide routines. [Getting Started](/en/docs/guides/getting-started)[AI Assistant](/en/docs/guides/ai-assistant)[Projects](/en/docs/guides/projects)[Changelog](/en/docs/changelog) ## Explore the platform [Working in AllooviumOne platform for the whole project: cited answers across your documents, routines that run your check-ins and reports, template filling and registers.Start with the user guide](/en/docs/guides/getting-started) [REV CREV DDocument controlVersion control and a defensible audit trail for every construction document. Clause-level diffs, trade markups, subcontractor packs and portal submissions.Read the document control guide](/en/docs/guides/exchange) [\[1\]AI AssistantAsk across every document and get answers with citations back to the exact clause. Draft documents, build routines, and act on the project record from one chat.Read the assistant guide](/en/docs/guides/ai-assistant) ## Quickstart paths Pick the role that fits and start in the right place. [10 minConstruction professionalYou manage documents, registers, and submissions. Get the AI assistant working across your project.Start with the User Guide](/en/docs/guides/getting-started) ## Popular guides What teams reach for most. [Document HubUploading documents](/en/docs/guides/projects/uploading) [AI AssistantAsking questions](/en/docs/guides/ai-assistant/asking-questions) [AI AssistantCitations](/en/docs/guides/ai-assistant/citations) [AI AssistantWorkflows](/en/docs/guides/workflows) [Document generationTemplate filling](/en/docs/guides/template-filling) [Document HubDrawing intelligence](/en/docs/guides/projects/drawing-intelligence) [IntegrationsSharePoint](/en/docs/guides/integrations/sharepoint) [Document controlClause diffs](/en/docs/guides/exchange/clause-diffs) Docs Assistant ## Get answers without reading the whole page. Ask anything about Alloovium — features, API endpoints, integrations, billing. Answers come from the docs directly, with clickable references that take you straight to the source. Ctrl/⌘ + K opens full search Ask the docs Try asking Every answer is drawn only from the documentation, and links to the exact page it came from. Press Enter to ask ## Community & support [Join the communityGet help from the Alloovium team and other users.Get in touch](mailto:zander@alloovium.com) [Need help?Get in touch with our support team directly.Contact Support](mailto:zander@alloovium.com) --- Canonical page: https://www.alloovium.com/en/docs --- ## Changelog Source: https://www.alloovium.com/en/docs/changelog Changelog # Changelog A high-level release history built from shipped product changes. It focuses on customer-visible progress, keeps sensitive implementation detail out of view, and makes it easier to see how quickly the product has been moving. April–July 2026Major release ## One product: Routines, Portal, and Conformance Alloovium consolidated into a single product. Exchange and Compliance were folded in as capability areas, Workflows became Routines, and the construction integration set expanded significantly. RoutinesIntegrationsCollaborationPlatform * +Merged Exchange (document version control and audit trail) and Conformance (compliance and attention) into the one Alloovium product — no more separate apps. * +Renamed Workflows to Routines: a linear, review-gated chain of trigger, action, and logic blocks that retrieve, reason, draft, and hand back for approval. * +Expanded construction integrations — a deeper Procore connection plus Autodesk APS, Aconex, Asite, Newforma, ProjectWise, Trimble Connect, Bluebeam, and Fieldwire, all via a self-hosted Nango. * +Introduced Action Items that turn a site update or meeting into a pre-filled variation or RFI draft. April 9, 2026Major release ## Docs, Developer Access, and Review Workflow Control Alloovium opened up a much broader public surface while tightening the workflows and review tools people use every day. Developer & DocsRoutinesAI & Analysis * +Refreshed the docs experience with a full guide, API reference, developer portal, and a rebuilt changelog. * +Expanded public developer access around API keys, MCP, capability docs, and clearer integration guidance. * +Upgraded the workflow library with filters, favourites, pins, run controls, grid/list view, and draft persistence. * +Improved review workflows with inline authorship cues, co-author support, and richer change review states. April 7-8, 2026Major release ## Notifications, Scheduling, and Customer-Facing Automation Automation moved from one-off runs to something teams can schedule, monitor, and trust. RoutinesCollaborationPlatform * +Shipped scheduled workflows with calendar preview, named runs, timezone support, and delivery notifications. * +Introduced a clearer notification center and stronger email notification coverage. * +Smoothed approval flows so review work is saved reliably and can trigger automatically from document revisions. * +Turned the new Workflows hub into a clearer home for personal, scheduled, and shared work. April 5-6, 2026Major release ## Workflow Studio, Monthly Reports, and Management Plans This release window pushed document automation much closer to a polished production workflow. RoutinesAI & Analysis * +Launched natural-language workflow creation with a redesigned create experience. * +Upgraded Monthly Report Fill with built-in templates, two-panel preview, editable results, and clearer progress stages. * +Shipped Management Plan generation end to end inside the product. * +Reworked workflow creation and history views to make ongoing automation easier to manage. Performance work in this window also reduced workflow launch overhead and improved sync throughput. April 4, 2026Release window ## Tender Reviews and Delivery Planning Construction-specific workflow coverage expanded beyond generic document analysis into delivery planning and tender support. RoutinesIntegrationsPlatform * +Added Tender Review, Tender Copy Forward, and Lookahead Schedule Generator flows. * +Moved advanced document generation toward a chat-style wizard experience. * +Made upload, retry, and processing states easier to understand during long-running work. * +Improved processing speed for fast-path document types and newly created workflow runs. April 1-3, 2026Major release ## Compliance Wizard, AI Brain, and Onboarding Overhaul A major product-facing release that connected guided compliance work, project intelligence, and first-run product experience. AI & AnalysisCollaborationPlatform * +Launched the ISO 9001 Gap Register and Compliance Wizard with guided evidence review, clause review, fix-gaps, and packaging steps. * +Turned AI Brain and Timeline into visible product surfaces for project intelligence and memory-style context. * +Rebuilt onboarding around role-aware steps, better demos, and real document previews. * +Moved change requests, versioning, project sharing, and reviews forward together as a more coherent workflow. March 29-31, 2026Major release ## Decision Intelligence and AI Brain The product moved from answering questions to surfacing higher-level project intelligence. AI & AnalysisIntegrations * +Built AI Brain into a dedicated intelligence workspace for memories, entities, and patterns. * +Expanded decision intelligence across timelines, entity relationships, briefings, scenarios, and predictive signals. * +Improved contradictions, risk scoring, and document-level intelligence so users can act on findings more easily. * +Broadened the intelligence story beyond documents by pulling in more communication and integration context. March 23-28, 2026Major release ## Template Fill, Document Generation, and Compliance A large release window that made structured output generation feel like a real product line rather than a single tool. AI & AnalysisRoutinesPlatform * +Matured template filling with search, filters, versioning, live progress, review flow, and source citations. * +Shipped evidence-backed document generation with stronger style fidelity and cleaner citation support. * +Introduced a fuller compliance platform with standards, assessments, evidence scanning, and ongoing health tracking. * +Added full-text search, multilingual marketing pages, and more polished editing flows. March 9-18, 2026Major release ## Sync Reliability and Processing Speed The platform spent this period getting much more reliable under real-world document and sync load. IntegrationsPlatformAI & Analysis * +Made SharePoint sync far more reliable across browsing, progress recovery, duplicate handling, and large-file imports. * +Improved bulk upload and document processing speed so large workspaces felt more responsive. * +Sharpened citation and highlight accuracy across PDFs, Word files, spreadsheets, and text content. * +Matured editing and review flows with better previewing, revision handling, and long-running progress feedback. March 2-8, 2026Major release ## Teams, Approvals, and Richer File Support Collaboration and file support moved closer to the shape needed for larger organisations. CollaborationRoutinesAI & Analysis * +Expanded teams, project memberships, sharing controls, and approval flows. * +Added voice input, stronger workflow templates, and more informative execution views. * +Improved support for spreadsheets and other file types across chat and workflows. * +Turned change impact and contradiction analysis into more concrete product experiences. February 27-28, 2026Release window ## Operational Polish This sprint focused on reliability and utility across high-touch operational surfaces. RoutinesPlatform * +Added duplicate upload detection and clearer confirmation flows to reduce accidental rework. * +Improved upload persistence, chat continuity, and stale-job recovery. * +Made image viewers and highlight alignment more dependable. February 14-17, 2026Major release ## Identity Refresh and Connected Workspaces Workspaces became easier to enter, easier to trust, and more tightly connected to the documents they contain. CollaborationIntegrationsPlatform * +Overhauled authentication with a cleaner sign-in experience and more stable workspace entry. * +Added project document attachments and workflow assistant chat to make conversations more context-rich. * +Expanded connected workspace behaviour around SharePoint and broader file coverage. * +Improved release speed to support faster product iteration. February 1-13, 2026Major release ## Workflows 2.0 Foundation This was a foundational release window: less visible than later launches, but important to many of them. RoutinesPlatform * +Rebuilt workflow execution around reusable execution plans, validation, evidence handling, and export steps. * +Started landing new workflow templates and stronger natural-language handoff into automation. * +Introduced more explicit safety, policy, and repair checks around generated outputs. * +Expanded the public site with a contact form, analytics, and broader product polish. January 8-31, 2026Major release ## Extraction Accuracy and Citation Precision Alloovium spent most of January getting much better at reading dense project documents and pointing users back to the right place. AI & AnalysisIntegrationsRoutines * +Overhauled table extraction, layout detection, and reading order for dense construction documents. * +Made citation mapping and highlight navigation much more precise inside project chat. * +Expanded the workflow canvas with zoom, pan, and clearer execution visuals. * +Improved SharePoint sync, folder filtering, MSG support, and DOCX viewing. December 19-31, 2025Major release ## Projects and Workflow Foundations By late December, the product had moved beyond a demo shell and started to resemble a usable document platform. RoutinesAI & AnalysisPlatform * +Moved workflow execution from basic prototypes toward reusable multi-step automation. * +Improved document highlighting and citation anchoring in a way users could immediately feel. * +Made project chat and assistant responses more grounded and more usable on real documents. * +Stabilised background processing enough to support larger ingestion flows. December 8-18, 2025Release window ## Cloud Transition and Team Foundations This was the move from a very early build toward a hosted product with real team workflows. PlatformCollaboration * +Shifted from an early local setup toward hosted deployment and cloud operations. * +Stabilised the dashboard, landing pages, upload flows, and assistant surface. * +Introduced teams and the first referencing pipeline. * +Started hardening security and release workflows. September to November 2025Release window ## Prototype Launch and First Cited Chat The earliest branch history shows the product taking shape from an initial prototype into a usable cited-chat experience. PlatformAI & Analysis * +Set up the initial product scaffold, early deployments, and first history views. * +Restored the first working cited chat loop with PDF extraction and conversation management. * +Started SharePoint-related groundwork and document history improvements. * +Closed out the first prototype cycle with better citation preview and history syncing. --- Canonical page: https://www.alloovium.com/en/docs/changelog --- ## Add-ins for Microsoft 365 — Overview Source: https://www.alloovium.com/en/docs/guides/add-ins Add-ins # Alloovium inside Microsoft 365 Three task panes — Word, Excel, and Outlook — that put your project's documents, registers, and AI assistant right next to whatever you're already working on. Screenshot The three add-ins, side-by-side Word answers questions about the open document, Excel builds and audits registers, Outlook reads the open email and its conversation. ## What you get The Alloovium add-ins are not separate products — they're the same project, documents, and AI assistant you use on the web, surfaced as a task pane inside Word, Excel, and Outlook. Sign in once and your active project follows you across all three. Each add-in is tuned to its host. Word focuses on what's selected in the document. Excel understands the active sheet's structure. Outlook reads the open email and the conversation it belongs to. One account, three surfaces The add-ins are deployed and signed in independently, but they share the same Alloovium account, the same projects, and the same Vault. Anything ingested in one shows up in the others. ## Choose an add-in | Add-in | You use it for | Best when | | ------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | | [Word](/en/docs/guides/add-ins/word) | Ask questions about the open document, review AI-suggested edits as native tracked changes, run "fill" workflows that draft contract clauses from your Vault. | You're reading or marking up a contract, scope, or specification. | | [Excel](/en/docs/guides/add-ins/excel) | Build obligation registers, variation registers, milestone schedules, and liquidated-damages calculators directly into a sheet with citations. | You're building or auditing a register, or comparing contracts side-by-side. | | [Outlook](/en/docs/guides/add-ins/outlook) | Summarise a whole conversation into decisions and commitments, review a draft before you send it, file email into the project record, draft grounded replies. | You're working correspondence — reading a long thread, or about to send something that matters. | ## How they all work Every add-in opens as a task pane on the right side of the host app. Click **Open Alloovium** in the ribbon and the pane slides in. The first time you open it, you'll be asked to sign in — the same login as alloovium.com. The shared building blocks across all three: * **Project picker** — every action runs against an active project. Every add-in asks you to pick one. * **Chat**— natural language Q&A scoped to your project plus the current document or email. Every answer cites the source files it was drawn from. * **Approvals queue** — anything the AI wants to write back (edits, register rows, workflow outputs) lands in a review queue first. You accept or reject before it touches the document. * **Workflows** — multi-step jobs your org has defined, like "extract obligations" or "draft RFI response", runnable from any add-in. Screenshot Shared layout Project picker at the top, content in the middle, approvals queue at the bottom — same shape in all three add-ins. ## Compatibility | Host | Required version | Notes | | ----------------- | ----------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ | | Microsoft Word | Microsoft 365 (any current channel) | Desktop (Windows / macOS) and Word on the web. Word 2019 perpetual is not supported — the add-in needs current Office.js APIs. | | Microsoft Excel | Microsoft 365 (any current channel) | Desktop and Excel on the web. | | Microsoft Outlook | Microsoft 365 (any current channel) | Outlook on the web, the new Outlook for Windows, and desktop Outlook (Windows / macOS) with a Microsoft 365 mailbox. | Microsoft 365 only The add-ins require Microsoft 365 (formerly Office 365). Standalone "Office 2021" and earlier perpetual builds do not have a recent enough Office.js runtime. ## Next steps * →[Install & deploy](/en/docs/guides/add-ins/install) — push the add-ins to your team via the Microsoft 365 admin centre, or sideload them on a single machine to try. * →[The Word add-in](/en/docs/guides/add-ins/word) — chat, tracked-change edits, and contract-fill workflows. * →[The Excel add-in](/en/docs/guides/add-ins/excel) — registers, variations, and contract comparison. * →[The Outlook add-in](/en/docs/guides/add-ins/outlook) — thread summaries, review-before-send, and email filing. --- Canonical page: https://www.alloovium.com/en/docs/guides/add-ins --- ## Excel add-in Source: https://www.alloovium.com/en/docs/guides/add-ins/excel Excel add-in # Alloovium in Microsoft Excel Turn contracts into auditable spreadsheets. Build registers, fill templates, and apply chat-driven edits through a single approval queue — every value tied to the source clause that produced it. Screenshot Alloovium task pane in Excel Three tabs in the bottom nav — Chat, Workflows, Edits. Header has the project picker and the overflow menu; ContextBar shows the live sheet read. ## Overview The Excel add-in plugs into Microsoft 365 Excel as a task pane. It understands the active sheet's structure, knows the active Alloovium project, and writes back to the workbook only through a single auditable path — every mutation runs inside one `Excel.run` transaction, so a partially-applied register is impossible. The pane has three tabs in the bottom nav: | Tab | What it does | | --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Chat | Q&A over the active project with the active workbook as context. Plain-English instructions that the planner turns into a preview-then-apply set of operations. | | Workflows | Hub of structured workflows — Fill Template at the top, Obligations Register and Variations Register below. Each opens its own sub-view; sub-tools (LD calculator, flow-down checker, milestone dates) live inside the Variations register. | | Edits | Chat-driven edits awaiting human review. Three segments — Awaiting me, Submitted, History. A static role-based authority table decides who can approve a given edit. | Video · \~3 min The Excel add-in in three minutes Open the task pane on a blank workbook, generate an obligations register from the project's parsed contracts, filter by severity and sync it to the sheet, then send one chat-driven edit through the Edits approval queue and watch it apply atomically. ## Open the task pane 1. 1 Install the add-in Centrally deployed via the M365 admin centre or sideloaded per machine. See the install guide. 2. 2 Open the workbook Any .xlsx — on disk, OneDrive, or SharePoint. New blank workbooks work too. 3. 3 Click "Open Alloovium" in the Home ribbon The task pane appears on the right. 4. 4 Sign in and pick the project First-run only. The same project picker as the web app — pick whichever contract or project this workbook is for. If the add-in recognises the workbook (see below), it'll suggest a project for you. ## Sheet detection & auto-match The first thing the add-in does is look at the active sheet. It samples the header row and a handful of sample rows, and asks the backend to classify the sheet — obligation register, variation register, milestone schedule, or something else. The detected sheet type appears in the context bar and tunes the suggested actions. Separately, when you open a workbook, the add-in tries to recognise it as a project you already have in Alloovium. The match scores filename, sheet names, and header/sample-row content against your accessible projects. A confident match shows a banner suggesting that project; you can accept or dismiss it per-document. Vault freshness Once a project is selected, a background poll checks for newer documents every 90 seconds. If anything's changed in the Vault, a **StalenessBanner** appears with a one-click Refresh. ## Chat tab The Chat tab is conversational Q&A over the active project, with the active workbook as live context. Every chat message sends a workbook context block — the header row, up to thirty sample rows, and the current selection — wrapped in XML delimiters that the LLM treats as untrusted data, not instructions. The intent classifier routes every message to one of six intents: | Intent | Behaviour | | -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | plan | The user described a structured change with values they already supplied or that are obvious from the sheet — "highlight row 5", "set B2 to 100", "colour overdue rows red". Backend returns a typed ExcelOperation plan (set\_value, set\_formula, highlight, insert\_row, clear\_range). Plan goes to the Edits tab for review; once accepted, ops apply atomically inside one Excel.run. | | populate | An edit whose values must first be retrieved from project documents — "fill column D with milestone dates from the contract". The AI searches the Vault, then proposes cell edits. On empty sheets the planner can skip the approval step and apply directly to fill a blank register. | | workbook\_qa | A question about the sheet itself — its data, structure, or totals. No document lookup, no mutations. Also the safe default for greetings and ambiguous input. | | document\_qa | A question about the content inside the project's uploaded documents. Retrieves from the Vault; citations come back as source chips under the assistant message. | | project\_meta | A question about the document collection itself — counts, inventory, what files exist — rather than what any one document says. "What documents do we have?", "how many files are uploaded?" | | out\_of\_scope | A substantive off-topic request (a recipe, the weather, a joke). The assistant declines politely. Bare greetings and acknowledgements are never treated as out of scope. | ## Workflows tab The Workflows tab is a hub. Clicking any card swaps the body to that workflow's sub-view; a back arrow in the sub-view header returns to the hub. Screenshot Workflows tab — hub state Fill Template (orange-bordered hero with AI badge) sits at the top; Obligations Register and Variations Register cards open their sub-views. The context bar at the top of the pane reads the active sheet so the planner can route suggestions correctly. * **Fill Template** — the hero card. Mirrors the Word add-in's Document Fill: point it at a template-shaped sheet, pick a Vault, and the planner drafts a value for each blank with a source citation, then writes the full batch into the workbook in one atomic transaction. * **Obligations Register** — opens the obligations sub-view (see below). Build the register from the project's parsed contracts, filter, and sync into the sheet. * **Variations Register** — opens the variations sub-view (see below). Houses the change-order register plus three sub-tools: LD Calculator, Flow-Down Checker, Milestone Dates. ## Obligations register The Obligations Register sub-view builds an auditable obligation register from the project's parsed contracts. Point it at a project Vault and it produces one row per obligation with party, action, due date, cadence, severity, clause reference, and the verbatim source text. ### Building the register 1. 1 Open a blank or partially-filled register sheet The add-in detects existing register layouts and switches the CTA from "Import" to "Update register" — it appends rows without disturbing what's there. 2. 2 Click "Generate from vault" Extraction kicks off across every processed document in the project. Production runs through a Celery queue; the pane subscribes to progress updates and lists obligations as they arrive. 3. 3 Filter and review Filter by severity or obligation type; bulk-select up to 500 rows; deep-link from a clause chip back into Chat to ask about it. 4. 4 Sync to sheet Writes the register into the workbook in a single atomic transaction. Citations live in a hidden column you can show on demand. Contradiction flags The Obligations view has a Contradictions toggle that surfaces flags from the LLM contradiction checker — places where two obligations on the same project disagree on party, value, or date. Each flag links back to both source clauses. ## Variations & calculators The Variations Register sub-view covers change-order workflows. Extract variations from notices, RFIs, and instructions into a structured register; or build a fresh variation log from scratch. Each row shows VO number, status pill, cost impact, time impact, source clause, approver, and submitted date. * **Send for approval** — kick a variation into the project approval workflow without leaving Excel. * **Sync to sheet** — write the full register into `A1:H{N+1}` with a dark header row. The same sub-view exposes three calculators, opened in an overlay: | Sub-tool | What it does | | ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | LD Calculator | Liquidated damages — rate × delay, with daily and weekly caps. "Look up" pre-fills daily rate and cap from the contract's LD clause via the project's clause-lookup backend. | | Flow-Down Checker | Head-vs-sub gap analysis. Assign Head / Sub / Skip per source document; the backend returns gaps with proposed subcontract wording for each missing obligation. | | Milestone Dates | Every dated obligation bucketed into Overdue / Due soon (≤14 days) / Upcoming / No specific date. | ## Edits tab Anything Chat is about to write to the sheet — register rows, formula changes, formatting — passes through the Edits tab first. It's the approvals queue: three segments at the top of the tab. * **Awaiting me** — edits assigned to you for review, with an op-kind chip summary, the first three operations, and a "Show all N operations" toggle for the rest. * **Submitted** — edits you've sent to someone else for review. * **History** — applied and rejected edits, filterable and exportable to CSV. Accepting an edit runs the operations inside one `Excel.run` transaction. A **StatusPill** at the top surfaces the outcome honestly — "Applied locally — Vault copy unchanged" on apply (the workbook changes; the Vault copy only updates when you sync it), "Edit rejected · Vault unchanged" on reject. Authority thresholds Approval authority follows a fixed, role-based table — the **Authority Thresholds**dialog shows it read-only: up to $10,000 sits with the Project Manager, $10,000–$100,000 with the Contract Manager, and anything over $100,000 with the Board. It’s a reference table, not a per-user setting. Screenshot Edits tab — Awaiting me Segmented control at the top — Awaiting me · Submitted · History. A StatusPill confirms the last apply ("Applied locally — Vault copy unchanged"). Each WorkbookEditCard shows a Pending chip, the title, op-kind chip row (Set / Highlight / Insert row), the first three operations as monospace previews, and a Show-all-N link for the rest. ## Plan validation Every edit plan — whether it came from Chat, Fill Template, or a register sync — is validated before it is allowed anywhere near your workbook. Validation runs at preview time and again at apply time: * **Range bounds** — every target must be a well-formed A1 reference inside the sheet's real bounds. * **Size caps** — a single edit is capped at 200 operations, each touching at most 10,000 cells; larger plans are rejected outright rather than partially applied. * **Formula safety** — proposed formulas run through a denylist of dangerous functions, and formulas that reference external workbooks are rejected. * **Precondition hashes** — when a plan is previewed, the add-in fingerprints the cells it is about to touch. At apply time it re-checks those fingerprints inside the `Excel.run` transaction — if the data changed underneath the plan (you kept editing, a colleague saved over it), the apply refuses instead of writing onto stale assumptions. Refuse, don't guess A plan that fails validation never half-applies. You'll see which check failed and can regenerate the plan against the current sheet state. ## Known limits * **Offline runs queue.** Online status is surfaced in the header. When offline, chat plan and populate calls go onto a durable queue (cap 50, 24 h TTL) and replay when you reconnect. Reads serve from React Query's cache where available. * **Sheet detection is heuristic.** Unusual layouts may misclassify — re-detect manually if the suggested actions don't fit. * **Sheet size matters.** Very large sheets (tens of thousands of rows) sample slowly. Filter or work on a smaller range when possible. * **Bulk delete capped at 500 rows.** The Obligations bulk-delete action processes up to 500 obligation IDs per request — paginate larger removals. * **Edit plans capped at 200 operations, 10,000 cells per operation.** Break very large writes into smaller plans — the validator rejects oversized plans rather than partially applying them. ## Next steps * →[Install the add-in](/en/docs/guides/add-ins/install) on your machine or push it to your team via the M365 admin centre. * →[The Word add-in](/en/docs/guides/add-ins/word) — same patterns, applied to documents instead of registers. * →[The Outlook add-in](/en/docs/guides/add-ins/outlook) — thread summaries, review-before-send, and email filing. --- Canonical page: https://www.alloovium.com/en/docs/guides/add-ins/excel --- ## Install & deploy the add-ins Source: https://www.alloovium.com/en/docs/guides/add-ins/install Add-ins # Install & deploy Two deployment paths — centralised admin push or single-machine sideload. ## Choose an install path | Path | Who it's for | Reach | | ------------------------------- | ---------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------- | | M365 admin centre (centralised) | IT admins rolling out to a team, department, or whole org. | All Microsoft 365 apps — desktop, web, mobile — for everyone you assign. | | Sideload (per machine) | Individuals trying the add-in, or developers testing. | One machine, current user only (Outlook sideloads install to the mailbox and follow the user across Outlook clients). | Recommendation For pilots, sideload on one machine to validate the add-in works in your tenant. For real rollout, use the M365 admin centre — it's the only path that survives machine rebuilds. ## Centralised deploy (M365 admin) The most robust path. Once deployed, the add-in shows up automatically for every user you've assigned it to, on every Microsoft 365 app they sign into. No per-machine work. Video · \~2 min Deploy to your whole org In the M365 admin centre: Settings → Integrated apps → Upload custom apps, paste the Word manifest URL, assign a pilot group, and deploy — ending with the Open Alloovium button appearing in a user's Word ribbon after restart. ### Prerequisites * You're a **Global Admin** or **Exchange Admin** in the M365 tenant * You have the add-in manifest URL from Alloovium (provided during onboarding) * Multi-factor authentication ready on the admin account ### Steps 1. 1 Sign in to admin.microsoft.com Use a Global Admin or Exchange Admin account. MFA is required for Integrated apps. 2. 2 Settings → Integrated apps → Get apps "Get apps" opens the catalogue. The Alloovium add-ins may not be in the public catalogue yet — use "Upload custom apps" instead and choose "Provide link to manifest file". 3. 3 Paste the manifest URL Word, Excel, and Outlook each have their own manifest URL. Repeat the upload step once per add-in. 4. 4 Assign users Pick "Specific users / groups" and select who gets it. Or pick "Entire organization" for a full rollout. "Test deployment" is also an option for a pilot group. 5. 5 Confirm and wait Microsoft propagates the deployment to users' Office apps within a few hours (sometimes up to 24h on the first push). Users will see "Open Alloovium" in the Home ribbon next time they restart. Screenshot M365 admin centre — Integrated apps The Integrated apps page where Alloovium add-ins are uploaded as custom apps. Manifest URLs Each add-in has its own host. The Word and Excel manifests live at `https://word-addin.alloovium.com/manifest.xml` and `https://excel-addin.alloovium.com/manifest.xml`; the Outlook manifest URL is provided during onboarding. Don't host a copy — point directly at the canonical URL so Microsoft picks up updates automatically. ## Sideload on one machine Sideloading installs the add-in for a single user on a single machine. Useful for quick trials before committing to a full rollout — or for development. ### Word and Excel on macOS 1. 1 Download the manifest Save manifest.xml from the URL Alloovium provides (Word and Excel have separate manifests). 2. 2 Drop the manifest into the Word/Excel WEF folder Word: \~/Library/Containers/com.microsoft.Word/Data/Documents/wef/. Excel: \~/Library/Containers/com.microsoft.Excel/Data/Documents/wef/. Create the wef folder if it doesn't exist. 3. 3 Restart Word or Excel The add-in appears in the Home ribbon as "Open Alloovium". ### Word and Excel on Windows 1. 1 Put the manifest on a network share or local folder Anywhere both the user and the trust setting can reach. 2. 2 Trust the folder in Office File → Options → Trust Center → Trust Center Settings → Trusted Add-in Catalogs. Add the folder path and tick "Show in Menu". Restart Office. 3. 3 Insert → My Add-ins → Shared Folder The add-in appears in the list. Click it once to install for this user; it stays installed until you uninstall. ### Word and Excel on the web (any OS) 1. 1 Open Word or Excel on the web office.com → Word or Excel → any document. 2. 2 Insert → Add-ins → Upload My Add-in Browse to the manifest.xml file and upload. The add-in installs for the current document and this user's account. ### Outlook (web, new Outlook, or desktop) 1. 1 Open the add-ins dialog In Outlook on the web or the new Outlook for Windows: open any message → the Apps icon → Get add-ins. In classic desktop Outlook, use Home → Get Add-ins. 2. 2 My add-ins → Add a custom add-in → Add from URL Paste the Outlook manifest URL Alloovium provides. Confirm the install warning. 3. 3 Open Alloovium from a message The add-in appears in the Apps menu on any open message (and in compose). Outlook add-ins install to the mailbox, so the same install follows you across Outlook clients. ## First-run sign-in The first time a user opens any add-in they'll be asked to sign in to Alloovium. The sign-in is a popup that uses the same identity provider as the web app — if your tenant uses SSO, the popup completes silently. Popup blockers The sign-in dialog opens as a popup. If users have a strict popup blocker, they may not see the prompt. Whitelist `*.alloovium.com` in your default browser's popup settings. ## Updates Add-in code and UI updates happen on the Alloovium side — there's nothing for you to redeploy. The manifest URL stays stable; Microsoft fetches the latest assets from it each time the pane opens. If the manifest itself changes (a new entry point, a new ribbon button, a new permission), users' Office apps will pick up the new manifest within a few hours. For centralised deploys, Microsoft pushes manifest changes through the same background sync as the initial deploy. ## Troubleshooting ### The Open Alloovium button doesn't appear * Centralised deploy: give it 24h after first push. Restart the Office app. * Sideload (Word/Excel macOS): make sure the manifest is in the right WEF folder for the host app — Word and Excel each have their own. * Sideload (Windows): re-check the Trusted Add-in Catalogs entry has "Show in Menu" ticked. ### Sign-in popup doesn't appear or fails * Whitelist `*.alloovium.com` in popup blocker settings. * If your tenant uses Conditional Access, ensure the Alloovium app is granted access. Talk to your Alloovium contact for the app ID. * Clear cached Office credentials on the machine, then re-open the pane. ### Add-in shows but actions hang * Check the network — the pane needs HTTPS access to `au-api.alloovium.com`. * For self-hosted Alloovium, confirm the API endpoint is HTTPS (HTTP is blocked by Office). * Open the add-in's developer tools (right-click inside the pane → Inspect) and check for blocked requests in the network tab. ## Next steps * →Back to [the Add-ins overview](/en/docs/guides/add-ins). --- Canonical page: https://www.alloovium.com/en/docs/guides/add-ins/install --- ## Outlook add-in Source: https://www.alloovium.com/en/docs/guides/add-ins/outlook Outlook add-in # Alloovium in Microsoft Outlook A task pane that reads the open email and its conversation. Summarise a long thread into decisions and commitments, get an advisory review of a draft before you hit Send, file correspondence into the project record, and draft grounded replies from your Vault. Screenshot Alloovium task pane in Outlook Bottom tab bar switches between Assistant, Files, Workflows, and Approvals. A context strip at the top shows which email the pane is reading; the header carries the project picker. ## Overview The Outlook add-in is a Microsoft 365 task pane wired into the same Alloovium backend as the web app. It always knows two things: your active project, and the email you have open — including, where the conversation has been synced to your project, the rest of the thread it belongs to. Four surfaces, switched from the bottom tab bar: | Surface | What it does | | --------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- | | Assistant | Chat over the active project with the open email as context. Hosts the thread-summary card in read mode and the review-before-send card in compose mode. | | Files | Browse the active project's documents without leaving Outlook — useful when a reply needs a fact from the record. | | Workflows | Built-in email workflows — contract scope comparison, email risk flagging, tender review, smart filing, and more. Runs land in Approvals first. | | Approvals | The review queue. Anything a workflow wants to produce or write back waits here for you to accept or reject. | Video · \~3 min The Outlook add-in in three minutes Open a long email thread, click Summarise conversation and read the decisions/commitments/unanswered-questions breakdown with its coverage note, then switch to a draft reply in compose mode, run Review before send, and act on a warning that the draft contradicts a figure in the project record. ## Open the task pane 1. 1 Install the add-in Centrally deployed via the M365 admin centre, or added per-user from a manifest URL. See the install guide for both paths. 2. 2 Open any email The add-in works from the reading pane and from an open message window — and in compose mode when you are writing a reply or a new message. 3. 3 Open Alloovium from the Apps menu The task pane slides in on the right. First time, sign in with the same account as alloovium.com. 4. 4 Pick an active project From the project picker at the top of the pane. Chat, filing, and workflows all run against that project. ## Email-aware chat The Assistant tab behaves like the web assistant with the open email attached as context. A context strip at the top of the pane shows exactly what the add-in is reading — the subject and sender of _this email_ — so there's never a question about what the answer was grounded in. Quick actions adapt to what's open. From a read email you can: * **Summarise** the email or its conversation (see below). * **Extract action items** — pull commitments, deadlines, and tasks out of the thread. * **Draft a reply** grounded in the active project's documents — ask for a response and the assistant composes one using the Vault as its source of facts. * **File the email** into the project so it becomes part of the searchable record. ## Thread summary Open any message that belongs to a conversation and a **Summarise conversation** card appears in the Assistant tab. One click produces a structured summary of the whole exchange — not prose, but the four things that matter on a project: * **Decisions** — what was agreed, and by whom. * **Commitments** — who promised what, with any dates attached. * **Unanswered questions** — asks still hanging in the thread. * **Changed positions** — where someone's stance moved over the course of the exchange. Honest coverage, always The summary card always states what it actually read: "Based on N synced messages" when the conversation has been synced to your project, or "Only this message — thread not synced" when it hasn't. It never claims thread coverage it doesn't have. Screenshot Thread summary card Decisions, Commitments, Unanswered questions, and Changed positions sections, with the coverage note underneath stating how many synced messages the summary is based on. From the summary you can keep asking — chat in the Assistant tab is conversation-scoped, so follow-up questions ("what did we say about the slab pour date?") are answered against the same thread plus the project record. ## Review before send When you're composing, the Assistant tab swaps in a **Review before send** card. Run it on a draft and Alloovium reads what you're about to send and reports back: * **Commitments** the draft makes — promises, dates, and amounts you're putting in writing. * **Warnings** — contradictions against the project's recorded facts, and things like attachments the text refers to but the message doesn't carry. Each warning is severity-tagged (Critical / Warning / Note). * **Suggestions** — smaller improvements worth a look before sending. Advisory, never blocking The review is user-invoked and purely advisory. Alloovium never blocks or delays the send — you stay in control of the Send button. The value is catching the figure that contradicts the register, or the missing attachment, while the fix costs nothing. Screenshot Review before send in compose mode A draft reply open in compose; the ReviewDraftCard lists one severity-tagged warning about a figure that contradicts the project record, the commitments the draft makes, and a suggestion — with Send untouched and available. ## Sent mail on the record Correspondence is only half a record if it only captures what arrives. When your Outlook account is connected to Alloovium, the sync covers your **Sent Items** too — so the reply you sent sits on the project record next to the email it answered, and thread summaries can cover both sides of the exchange. * Every synced message is stored direction-aware — inbound and outbound are distinguished, and your own outbound mail doesn't trigger the alerting that watches incoming correspondence. * Sent-items sync can be switched off per connection if you only want inbound capture. The Outlook _connection_ (mailbox sync) and the Outlook _add-in_ (this task pane) are separate pieces: the connection keeps the record flowing in the background, the add-in is the surface you work in. See [the Outlook integration guide](/en/docs/guides/integrations/outlook) for connecting the mailbox. ## Workflows & approvals The Workflows tab carries built-in email workflows — comparing an email and its attachments against the contracted scope, flagging risk in incoming correspondence, reviewing tender documents against bid criteria, routing approvals, and smart email filing. Anything a workflow produces goes to the **Approvals** tab first; a badge on the tab shows how many runs are waiting on you. Define workflows once, run everywhere Workflows are defined on the web. Any workflow tagged for Outlook shows up in this add-in — the same as Word- and Excel-tagged workflows in their hosts. ## Known limits * **Thread coverage depends on sync.** The conversation summary only covers messages that have been synced to your project — the coverage note on the card tells you exactly how many. An unsynced thread summarises this message only. * **Review is compose-mode only.** Review before send appears when you're writing a message, not when reading one. * **One project per session.** Like the other add-ins, the pane is scoped to a single active project at a time. * **Backend must be HTTPS.** The add-in cannot talk to plain http:// — for self-hosted Alloovium, terminate TLS before the API. ## Next steps * →[Install the add-in](/en/docs/guides/add-ins/install) on your mailbox or push it to your team via the M365 admin centre. * →[Connect the Outlook integration](/en/docs/guides/integrations/outlook) — mailbox sync is what gives thread summaries their coverage. * →[The Word add-in](/en/docs/guides/add-ins/word) — the same patterns, applied to the document you're drafting. --- Canonical page: https://www.alloovium.com/en/docs/guides/add-ins/outlook --- ## Word add-in Source: https://www.alloovium.com/en/docs/guides/add-ins/word Word add-in # Alloovium in Microsoft Word A task pane with three surfaces — Chat, Workflows, Edits — wired into the same backend as the web app. Ask about a selection, accept structured edits, fill a draft contract from your Vault. Screenshot Alloovium task pane in Word Bottom tab bar switches between Chat, Workflows, and Edits. The header carries the project picker and overflow menu; the ContextBar shows the live document read. ## Overview The Word add-in is a Microsoft 365 task pane wired into the same Alloovium backend as the web app. It always knows three things at once: your active project, the document you have open, and any text you have selected. That context drives three nav surfaces, switched from the bottom tab bar: | Surface | What it does | | --------- | ---------------------------------------------------------------------------------------------------------------------------------------------- | | Chat | Q&A over the active project, with the open document and selection as context. Every answer comes with citations back to the source clauses. | | Workflows | Two hero workflows — Document Fill and Generate Document — plus any org workflows tagged for Word. All run with live SSE progress. | | Edits | Reviewable structured edits, applied via Office.js. Accept, reject, or accept all; then push the new revision to the Vault with Save to Vault. | The header carries the project picker and a preferences panel. Sign-in uses Clerk; an offline banner appears at the top of the shell when the backend is unreachable so you know when actions will queue. Video · \~5 min The Word add-in in five minutes Open a contract in Word desktop, launch the task pane from the Home ribbon, ask a question about a highlighted clause and follow the citation back into the document, accept a suggested edit from the Edits panel, then run Document Fill on a draft and review the filled sections as tracked changes. ## Open the task pane 1. 1 Install the add-in Once-per-machine setup. M365 admins can push it to a whole team via the admin centre, or a single user can sideload it. See the install guide for both paths. 2. 2 Open any Word document Works on .docx files on your computer, in OneDrive, or on SharePoint. No special document format is needed. 3. 3 Click "Open Alloovium" in the Home ribbon The task pane slides in on the right. First time, you'll be asked to sign in with the same Clerk account as alloovium.com — the dialog opens in a popup window. 4. 4 Pick an active project From the project picker at the top of the pane. Everything you do from this point on is scoped to that project's Vault and chat history. Keyboard shortcuts The pane registers a few hotkeys: `Cmd/Ctrl+Enter` sends a chat, `/` opens the prompt picker, `Cmd/Ctrl+K` starts a new chat, `Esc` cancels the active stream, and `?` opens a shortcut help overlay. ## Document recognition When you open a document, the add-in classifies it heuristically into one of contract, specification, report, proposal, meeting minutes, RFI, or drawing register. The detected kind appears as a chip under the Quick Actions row and tunes the suggested prompts and the workflows that surface first. The chip is informational; it doesn’t auto-switch projects. To scope your work to a project, pick one in the project picker at the top of the pane. ## Chat & ask about a selection The Chat surface behaves like the web assistant, with one extra trick: selection-awareness. Whenever you highlight text in the document, a **HighlightAsk**card appears in the pane and quietly attaches the selection as context for your next message — there’s no button to press. Just highlight, type your question, and the reply is grounded in what you selected. A dismiss control drops the selection if you don’t want it carried along; clearing the highlight removes it automatically. Screenshot HighlightAsk in the task pane Selection-aware card — a "Selected text · in context" header with a dismiss control, and the live selection rendered as an italic preview below. It appears passively the moment you highlight; no quick-action buttons. Every answer comes with citations that open the source document. For in-document citations the add-in calls Word's `Range.select()`, which auto-scrolls; external citations open in the Vault viewer. As you type, an action-intent classifier decides whether the message is an ask, a fill, a workflow request, or an edit. If it smells like a workflow, a **WorkflowStatusCard** appears with a one-click "Run as workflow" affordance. If it's a compliance check, the findings render inline as severity-tagged cards under your message. What does it know? The add-in sees the document text, your active selection, and any document you've ingested into the active project. It does not see other Word documents you have open, and it never sees content from projects you haven't selected. ## Edits & Save to Vault When a workflow or agent produces structured edits, they show up in the **Edits** surface. Each suggestion shows the original passage, the proposed replacement, and a rationale. * **Accept** applies the change as a native Word tracked change, anchored to the exact passage it targets — so the edit shows up in Word's own review markup and can be accepted or rejected there like any human edit. * **Reject** dismisses the suggestion and removes it from the queue. * **Accept all** applies every remaining suggestion in one batch. Anchoring is deliberate, not fuzzy: the add-in locates the full original passage in the document (using surrounding context to disambiguate when the same text appears more than once) and verifies the match before writing. If the target text appears in several places and can't be disambiguated — or the passage has changed since the suggestion was made — the add-in refuses to apply that edit rather than guessing, and tells you why. The Vault is not updated automatically Accepting edits changes the live Word document only. To make the Vault reflect the new revision, click **Save to Vault** in the Edits panel — that uploads the current `.docx` as rev N+1 on the matched document. If someone else updates the document in the Vault while you have it open, a banner offers to reload so you don't accept findings against a stale copy. Screenshot Edits panel with one suggestion expanded Each suggestion card has a numbered badge, title, and chevron. Expanded: rationale, Before (red) and After (green) diff blocks, plus navigate / apply / dismiss controls. Save to Vault sits in the footer. ## Document Fill Document Fill is the flagship workflow for Word. Point it at a draft contract with blanks (placeholders, square brackets, empty heading rows) and pick a Vault of reference material. Alloovium drafts a fill for each blank, cites where the value came from, and lands them in a review pane before anything touches the document. ### Running a fill 1. 1 Open the draft document Any .docx with recognisable blanks. Common patterns auto-detect: "\[Insert\]" markers, "TBD", empty heading rows in registers, and standard contract placeholder syntax. 2. 2 Open the Workflows surface → Document Fill The pane lists all workflows tagged for Word, with Document Fill at the top. 3. 3 Pick a Vault in the VaultSelector The Vault is where the AI looks for source material — past contracts, project facts, party lookup tables. You can pick the active project's Vault or any other Vault you have access to. 4. 4 Watch FillProgress stream SSE-streamed progress per blank. You can keep editing the document while the fill runs in the background. 5. 5 Review each draft The FillReview pane shows every blank, the drafted value, and the source citation. Edit any value inline before accepting. Skip the ones you want to fill yourself. 6. 6 Insert as tracked changes Fills are written into the document at the correct anchor points as Word tracked changes — nothing is committed silently. You then accept or reject each section from its review card (or Accept all), so the fill lands section by section under your control. Regenerate clears a section's prior tracked change before reinserting. Screenshot Document Fill — running Select → Fill → Complete stepper at the top; the "Filling the template" card streams stage updates (Analysing template → Planning content → Filling placeholders → Filling tables → Writing sections → Finalising). The Edits tab badges any suggestions generated mid-run. Screenshot FillReview — once filling completes Summary at the top — filled / accepted / rejected / pending counts — plus an Accept-all-pending shortcut. Each section row has Accept (orange), Reject, and Regenerate controls; the expanded row reveals the drafted value and a source citation. ## Generate Document Generate Document is the second hero workflow in the Workflows tab. Unlike Document Fill — which pours content into a template you already have — Generate writes a brand-new document from a prompt, the project’s Vault, and whatever is already in the active document as context. ### Three steps 1. 1 Prompt Pick a project (or All projects), write a prompt describing the document you want, optionally narrow retrieval to specific source documents, and add tone/audience instructions. The form lives inside a single field-group layout so all inputs stay visible. 2. 2 Write The backend streams an outline first, then per-section content. Stage chips show progress live; sections land into the active Word document as they arrive — you can watch the document compose itself. 3. 3 Complete A summary card shows the generated title, section count, the project that supplied sources, and a Review hint asking you to fact-check before sending the document on. From here, Generate another resets the form. Screenshot Generate Document — Prompt step Project, Prompt, and Source documents field groups; the orange Generate CTA kicks off streaming. Stepper at the top shows Prompt active, with Write and Complete still ahead. Same Edits surface as Fill Sections produced by Generate land in the Edits panel as structured suggestions you can accept inline. Save to Vault on the Edits panel pushes the final document as a new revision once you’re happy. ## Running workflows Any workflow your team has built and tagged for Word shows up in the **Workflows** tab. Each row has a name, description, and a Run button that opens the workflow detail view with its input schema (documents, project, fields). Once running, progress streams live into the pane over SSE — every step (pending / running / done / failed) updates in order. You can keep editing the document while the workflow runs. Results land where they belong: edits in the Edits panel, fills in the FillReview pane, plain answers in the chat thread. Define workflows once, run everywhere Workflows are defined on the web. Any workflow tagged for Word is automatically available in this add-in — Excel-tagged workflows show up in the Excel add-in, Outlook-tagged in Outlook. ## Known limits * **Offline reads only.** An offline banner appears when the backend is unreachable; chat send is disabled with a tooltip. Mid-stream actions complete, but new requests have to wait for the connection to come back. * **One project per session.** The pane is scoped to a single project at a time. Switch projects with the picker at the top — there's no multi-project chat in the add-in. * **Document recognition is heuristic.**The classifier is informational only; it doesn’t auto-switch projects. Scope your work by picking a project in the project picker. * **Backend must be HTTPS.** The add-in cannot talk to plain http://. For self-hosted Alloovium, terminate TLS at a proxy before the request reaches the API. ## Next steps * →[Install the add-in](/en/docs/guides/add-ins/install) on your machine or push it to your team via the M365 admin centre. * →[The Excel add-in](/en/docs/guides/add-ins/excel) — same patterns, applied to registers, variations, and contract comparison. * →[The Outlook add-in](/en/docs/guides/add-ins/outlook) — thread summaries, review-before-send, and email filing. --- Canonical page: https://www.alloovium.com/en/docs/guides/add-ins/word --- ## Admin & Teams Setup Source: https://www.alloovium.com/en/docs/guides/admin Administration # Admin & teams setup Everything you need to stand up your organization in Alloovium: adding your colleagues, deciding who can do what, grouping people into teams, and connecting single sign-on. ## Overview An organization in Alloovium is your company workspace. It holds your projects, documents, templates and the people who work on them. This guide covers the one-time setup an administrator performs: inviting people, granting them the right access level, organizing them into teams, and connecting sign-on. Most of this is managed from the `Organization` area, under the `People` tab. For day-to-day sharing of individual projects and documents, see [Teams](/en/docs/guides/collaboration) and [Sharing](/en/docs/guides/collaboration/sharing). This page focuses on setting the organization up, not the everyday collaboration that follows. ## Your organization One company, one workspace. Everyone you invite becomes a member of it. When your company is provisioned, a single organization is created and you are set up as an administrator. Every person you invite joins this same organization and shares its projects, document library, templates and standards. A person belongs to exactly one organization. You manage the organization from the `Organization` area, which is divided into a few tabs: | Tab | What it holds | | -------- | --------------------------------------------------------- | | Library | Shared documents available across your projects. | | Company | Company profile, and the standards and codes you work to. | | Contacts | Your directory of companies and external contacts. | | Playbook | Default project structure and reusable lessons. | | People | Members, teams and roles — where most admin work happens. | ## Inviting people Send an email invitation and choose the person's role and teams up front. You add colleagues by sending an email invitation from the `People` tab. Each invitation carries a role and, optionally, one or more teams — so people land with the right access from their first sign-in. 1. 1 Open the People tab Go to the Organization area and select People, then choose to invite a member. 2. 2 Enter the email address Add the person’s work email. You can review existing members and recent invitees to avoid duplicates. 3. 3 Choose a role Pick a role from your organization’s role catalogue. The role determines the person’s access level. 4. 4 Assign teams (optional) Add the person to one or more teams so they immediately see the right projects. 5. 5 Send the invitation Alloovium emails a secure, time-limited link. When the person accepts, their account is provisioned on your organization. Pending invitations are listed on the People tab, where you can revoke one that was sent in error or is no longer needed. Invitations expire after a set period, so a stale link cannot be used later. Accepting an invitation The invited person clicks the link in the email, signs in, and is added to your organization automatically. They do not need to be added a second time once they accept. ## Roles & access levels A role sets how much a person can do across the organization. Roles map to a set of organization-wide access levels. The access level is what governs administrative capability; the role label (for example a job title such as Site Engineer or HSEQ Manager) is what you assign day to day, and it carries the access level with it. | Access level | Intended for | Can manage the organization | | ------------- | ---------------------------------------------------------------------------- | --------------------------- | | Org admin | People who administer the whole workspace — members, teams, settings. | Yes | | Project admin | People who lead individual projects and their teams. | Within their projects | | User | Everyday members who work on projects they are given access to. | No | | Viewer | People who need read-only visibility into projects they are given access to. | No | Your organization also keeps a catalogue of roles that reflect real construction job titles — for example Project Manager, Superintendent, Site Manager, Contracts Manager or Safety Officer. Each catalogue role is linked to one of the access levels above, so choosing a role for a person sets both their job title and their access in one step. Administrators can add, edit or archive catalogue roles to match how your teams are organized. Project- and folder-level access The access levels here are organization-wide. Access to a specific project or folder is set separately when you share it. For the granular, per-project permission model, see [Permissions](/en/docs/guides/collaboration/permissions). ## Teams Group people so you can grant access to projects in one move. A team is a group of people within your organization. Teams make access management simpler: grant a team access to a project and every member of that team inherits it, rather than adding each person one by one. Teams can be nested, so a division can contain the crews that report into it. ### Internal, external and company teams Teams come in a few flavours. An `internal` team is the usual grouping of your own members. An `external` team holds outside individuals — for example a client or subcontractor contact who does not have a full account — so you can share specific material with them. A `company` team links to an entry in your Contacts directory and reuses that company's identity. You manage teams and their members from the `People` tab. Within a team, members carry a team role, and you can grant the team access to a project as read-only, read-write, or admin. ## Sign-in & SSO Alloovium uses a managed identity layer for secure sign-in. Sign-in is handled through a managed authentication layer. People sign in with their credentials and, once verified, are placed into your organization based on the invitation they accepted. Their session is validated on every request, so access follows the role and teams you assigned. If your company requires a specific sign-in method for your rollout — such as connecting your identity provider — arrange this during onboarding so it is configured before you invite people at scale. An org admin can also enforce multi-factor authentication across the organization, requiring every member to complete a second verification step at sign-in. Invite first, then sign in A person must have an accepted invitation to your organization before they can reach your workspace. Signing in without an invitation does not grant access to your projects or documents. ## Organization settings Company profile, standards, playbook and workspace-level controls. Beyond people, the Organization area holds settings that shape how your whole workspace behaves. These are worth setting early so new projects and generated documents start from the right defaults. | Setting | Where | What it does | | ---------------- | ------------ | --------------------------------------------------------------------- | | Company profile | Company tab | Your company name and description, used across the workspace. | | Standards | Company tab | The codes and standards you work to, which inform reviews and checks. | | Project skeleton | Playbook tab | A default folder structure applied to new projects. | | Lessons | Playbook tab | Reusable lessons that resurface on relevant documents. | Administrators can also manage encryption settings for the organization, including the option to use a customer-managed key. If a member needs to leave the organization, they can do so from the Company tab; this removes them and revokes their team access, and it cannot be undone. Leaving is permanent Leaving an organization revokes access immediately and is not recoverable. If you are the only administrator, appoint another org admin before you leave so the workspace is not left unmanaged. --- Canonical page: https://www.alloovium.com/en/docs/guides/admin --- ## AI Assistant Source: https://www.alloovium.com/en/docs/guides/ai-assistant AI ASSISTANT # AI Assistant One conversation that knows your project — and can act on it. Ask questions and get answers with a citation on every claim; then ask for the follow-through: draft the notice, send the email, book the meeting, build the routine. Everything it does is grounded in your project record and gated on your approval. Video · \~3 min Ask your project anything End-to-end tour of the AI Assistant: typing a question about a project document, watching the answer stream in with numbered citations, clicking a citation to open the highlighted source passage, then using a follow-up suggestion chip to go deeper. ## Overview The AI Assistant is the front door to everything in Alloovium. It’s a conversational interface, but unlike a general-purpose chatbot it is always grounded in your workspace: your documents, your project’s live records — variations, RFIs, the programme, the timeline — your meetings and your connected email. Every answer links back to the passage or record it was derived from, so you can verify the source rather than take its word. And it doesn’t stop at answers. The same ask bar takes instructions: draft a document, email the subbie, invite someone to the project, share a file with another company, flag an activity at risk, set up a routine that does the checking for you. Actions are never fired silently — the assistant drafts what it intends to do and waits for your confirmation. Conversations are multi-turn: ask follow-ups, refine a draft, change your mind — the assistant keeps the thread’s context throughout. ## What you can ask Questions can be broad (“what are the key deliverables in this contract?”) or highly specific (“what’s the liquidated damages rate in clause 12?”) — and they aren’t limited to documents. The assistant answers from every source the project keeps: | Source | Example ask | | -------------------- | ------------------------------------------------------------------------------------------------------------------------ | | Documents | “What does the spec say about curing time?” · “Compare the two amendment versions — what changed?” | | Timeline & registers | “What happened with VO-118?” · “Which RFIs are still open?” | | Schedule & lookahead | “What’s planned on site next week?” · “Is anything at risk?” | | Meetings | “What did we decide about the facade?” | | People | “Who’s the services engineer on this job?” | | Email | “Did Karen send the facade photos?” · “Any new claims in the inbox this week?” | | Company knowledge | “Have we dealt with this kind of latent condition before?” — lessons and past projects, when you ask across the company. | Scope follows where you are: asking inside a project searches that project; asking from the company level — or saying “across all projects” — widens to everything you have access to. You can narrow too: name a folder and the answer draws only on that folder and everything beneath it. Retrieval is permission-aware throughout — the assistant only ever reads what you could open yourself. Query tip Phrase questions the way you’d ask a colleague. “Does this contract include a warranty period, and if so, how long is it?” works better than “warranty period duration”. ## Taking action When the follow-through is a write — a message, a meeting, a to-do, an invitation, a document share, a schedule risk flag — you can ask for it in the same sentence you’d use with a person: “text Wayne the crew count”, “invite Pat to the project”, “send the signed amendment to Lumino”. The assistant drafts the action, shows you exactly what will happen, and executes only when you confirm. See [Taking Action](/en/docs/guides/ai-assistant/actions) for the full list and how the approval flow works. ## Drafting documents Ask for a document — a delay notice, an RFI response, an evidence pack, a position paper — and the assistant writes it from the project record, with a citation on every factual claim and your library templates matched automatically. The draft opens beside the chat, where you refine it conversationally (“keep it to one page”, “add a cost table”) and save it to the project when it’s right. See [Drafting Documents](/en/docs/guides/ai-assistant/drafting). ## Creating routines Anything you ask repeatedly can become a routine. Describe the recurring job in one sentence — “every morning, check the inbox for new claims and send me a summary” — and the assistant builds the routine: trigger, steps, conditions and delivery, laid out for your approval before it’s switched on. It will even notice repeated asks and suggest automating them with an `Automate this` chip. See [Creating Routines from Chat](/en/docs/guides/ai-assistant/creating-routines). ## Citations Every answer carries numbered citations linking to the exact evidence behind it. Document citations open the file at the highlighted passage; `[T#]` citations point at live records — timeline events, action items, meetings — and deep-link to where those records live. One answer can carry both kinds. Multiple sources When an answer draws from more than one document, each citation is numbered separately so you can see exactly which portion of the answer came from which source — useful when comparing information across contracts or specifications. See [Citations](/en/docs/guides/ai-assistant/citations) for the full walkthrough. ## File attachments In addition to querying documents already in your project, you can attach files directly to a chat message — useful for a one-off question about a file without adding it permanently to the project. Supported formats include PDF, DOCX, XLSX, PPTX, CSV, TXT, PNG, JPG, and TIFF. Attached files are processed for the duration of the conversation and are not stored in the project unless you explicitly add them. File size limit Individual files attached directly in chat are limited to 20 MB. For larger documents, upload them to the project first and then query them through the assistant. ## Voice input Speak instead of typing: click the microphone in the ask bar, dictate your question or instruction, and the words appear as text you can edit before sending — particularly handy on site or for longer briefs. See [Voice Input](/en/docs/guides/ai-assistant/voice-input). For hands-free work away from the screen entirely, the assistant also answers [site calls and SMS](/en/docs/guides/site-calls). ## Sharing conversations Chat threads can be shared with other members of your team. Recipients see the full exchange — questions, answers, and citations — read-only, without re-running the queries. Use the `Share` button at the top of the conversation, or the conversation's three-dot menu in the sidebar. Permissions Sharing a conversation does not automatically share the underlying documents. Recipients who are not already members of the project will see the answers and citations, but may not be able to click through to view the source document. ## Background mode For complex queries that take longer — synthesis across a large number of documents, multi-step questions — you can run them in background mode. Submit the question and go work on other things (or close the tab) while the assistant processes it. You’ll get an in-app notification when it completes, and optionally an email, so you can pick up where you left off. When to use background mode Background mode is best for queries across 20 or more documents, or complex multi-step questions that synthesise a large amount of information. For quick factual lookups, standard mode is faster. --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant --- ## Taking Action Source: https://www.alloovium.com/en/docs/guides/ai-assistant/actions AI ASSISTANT # Taking Action The same ask bar that answers questions can do things: create to-dos, book meetings, send emails and texts, invite people to a project, share documents with another company, flag schedule risk. You describe the action in plain language; the assistant drafts it and waits for your approval. Video · \~2 min Do it from the ask bar Three actions in one sitting: asking the assistant to add a to-do (draft card → confirm → it appears in Today), then to email a subcontractor about a late claim (reviewing the drafted recipients and body, editing one line, sending), then to invite a new engineer to the project (the invitation draft, confirmation, and the pending invite appearing on the People page). ## Overview Most questions to the assistant are reads — “what does the contract say?”, “what happened this week?”. But plenty of the follow-through is a write: someone needs to be chased, a meeting needs to move, a document needs to go out. Rather than switching to another screen, you can ask for the action in the same sentence you’d use with a colleague — “text Wayne the crew count”, “invite Pat to the project”, “send the latest claim to Calabria”. The assistant works out what kind of action you mean, gathers what it needs from the project — the right person, the right document, the right time — and lays the whole thing out as a draft for you to review. It never acts first and tells you after. ## What it can do | Action | Example ask | | ------------------------------- | ---------------------------------------------------------------------------------------------------- | | Create a to-do | “Add a task to inspect the Level 14 pour prep on Friday.” | | Complete or move tasks | “Mark the pour-prep inspection done.” · “Push my tasks to Monday.” | | Book, move or cancel events | “Book a site walk at 9am Thursday with Reece.” · “Move Thursday’s meeting to Friday.” | | Send an email | “Reply to the client saying we’ll price the Level 10 bring-forward this week.” | | Send an SMS | “Text Wayne asking for tomorrow’s crew count.” | | Invite someone to the project | “Invite pat@meridian.com.au to Northgate.” — a real platform invitation, not just an email. | | Share a document via Exchange | “Send the signed amendment to Lumino.” — a tracked company-to-company share with receipts. | | Flag schedule risk | “Flag the Level 10 waterproofing as at risk — membrane delivery slipped.” | | Raise RFIs and start variations | “Raise an RFI about the facade shop drawings.” · “Start a variation for the Level 10 bring-forward.” | Two ways to send a document “Email the report to Sam” drafts an email with the file attached. “Send the report to Lumino” (another company) goes through [Exchange](/en/docs/guides/exchange) — a tracked share with delivery and read receipts. The assistant picks the right lane from how you phrase it, and the draft card always shows which one it chose. ## Nothing happens without you Every action follows the same pattern: the assistant prepares a draft card showing exactly what will happen — the recipient, the message body, the event time, the document being shared — and nothing executes until you confirm it. You can edit the draft directly, ask the assistant to change it, or dismiss it entirely. For messages, what goes out is what you approved — the assistant doesn’t rewrite your words after you’ve confirmed them. If the ask is missing something essential (a send needs a recipient, an event needs a time), the assistant asks rather than guessing. ## Refining a draft A pending draft stays in the conversation, so you can shape it the way you’d shape anything else in chat: “make it 8am instead”, “add Janelle”, “firmer tone”. The assistant updates the same draft rather than starting a new one, and the card re-renders with the change so you can see what you’re about to approve. ## Conditional asks Actions can be gated on a check: “Check my calendar for tomorrow morning — if I'm double-booked, move my site walk to Monday morning.” The assistant runs the read first and acts only if the condition actually holds — and the result carries a reference chip back to the thing it changed, so the outcome is on the record, not just in the reply. The same shape works for recurring routines — see [Creating Routines](/en/docs/guides/ai-assistant/creating-routines). ## Where actions land Confirmed actions create real records, not chat-only artifacts. A to-do appears on your [Today](/en/docs/guides/today) page, an event lands on your calendar, an invitation shows as pending on the People page, an Exchange share appears in the package register with its receipts — and each one writes an event to the project [timeline](/en/docs/guides/projects/timeline), so there’s always a record of what was done, by whom, and when. --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/actions --- ## Asking Questions Source: https://www.alloovium.com/en/docs/guides/ai-assistant/asking-questions AI Assistant # Asking Questions Ask anything about your project documents in plain language and get accurate, cited answers in seconds. ## Asking Questions You can ask the AI Assistant anything related to the documents in your project. Questions can be broad ("What are the key deliverables in this contract?") or highly specific ("What is the liquidated damages clause in section 12?"). The assistant handles both. The assistant is context-aware throughout a conversation. If you ask "What does it say about payment terms?" followed by "And what are the penalties for late payment?", it understands that the second question continues the same thread. Scope matters: ask from inside a project and answers draw on that project's documents; ask from the company-wide Assistant and every project you can access is in play. If a question needs a source you haven't uploaded, the answer will say so rather than guess. 1. 1 Open a project Navigate to a project with uploaded and processed documents. 2. 2 Open the AI Assistant Click the chat icon in the sidebar or use the keyboard shortcut Cmd+K to open the assistant panel. 3. 3 Type your question Enter your question in the text field at the bottom of the chat panel and press Enter or click Send. 4. 4 Review the answer and citations The assistant responds with an answer and numbered citations. Click any citation to jump to the source passage. 5. 5 Ask follow-up questions Continue the conversation by asking follow-ups — the assistant retains context from the entire thread. Query tip Phrasing your question as you would ask a colleague often produces the best results. For example: "Does this contract include a warranty period, and if so, how long is it?" works better than "warranty period duration". --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/asking-questions --- ## Citations Source: https://www.alloovium.com/en/docs/guides/ai-assistant/citations AI Assistant # Citations Every AI answer links back to the exact source passage it was drawn from. Click any citation to navigate directly to the document. ## Citations Every answer generated by the AI Assistant includes numbered citations that link directly to the source passages in your documents. Citations appear as superscript numbers within the answer text, and a full reference list appears below the response showing the document name, page number, and a short excerpt. Clicking a citation opens the referenced document in the document viewer and highlights the exact passage that the assistant used to construct its answer. Video · \~1 min From citation to source A question is asked, the answer arrives with numbered citations, and one citation is clicked to open the source document with the cited passage highlighted — proving where the answer came from. Multiple sources When an answer draws from more than one document, each citation is numbered separately so you can see exactly which portion of the answer came from which source. This is especially useful when comparing information across contracts or specifications. --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/citations --- ## Creating Routines from Chat Source: https://www.alloovium.com/en/docs/guides/ai-assistant/creating-routines AI ASSISTANT # Creating Routines from Chat Anything you find yourself asking the assistant repeatedly can become a routine — and the fastest way to build one is to just say it: “every morning, check the inbox for new claims and send me a summary.” The assistant turns the sentence into a routine and shows you exactly what it built before switching it on. Video · \~90 sec From a sentence to a routine Typing 'Every morning at 7, check the lookahead and message me anything at risk' into the ask bar, the assistant laying out the routine draft — trigger, steps, condition and delivery — a quick edit to the time, confirming, and then the routine appearing in the Routines list and its first run landing on the project timeline. ## Overview Routines are recurring jobs that check things, draft things, and chase things without you driving. There’s a full visual builder for them (see [Routines](/en/docs/guides/routines)), but you don’t need it to get started — the assistant can build a routine from a plain-language description, the same way you’d hand the job to a person. Like every assistant action, routine creation is draft-gated: the assistant lays out what it understood — when the routine runs, what each step does, any condition, and how the results reach you — and nothing is scheduled until you confirm it. ## Describe the job in one sentence The whole brief fits in one message: “every Monday at 7, draft the weekly site report from last week’s diaries”, “each morning, summarise anything new in the inbox about Northgate”, “every Friday, remind the subbies whose claims are due next week”. The assistant maps the sentence onto real routine steps — the same blocks the visual builder uses — so what you get is a first-class routine, not a chat-only reminder. If something essential is missing, the assistant asks instead of guessing — most commonly _how the result should reach you_: a notification, an email, or a post to your Today page. ## Conditions: “if…, then chase” Routines can carry a condition, and you can say it naturally: “every morning check the lookahead — if anything is flagged at risk, message the responsible sub.” The condition is evaluated on every run, in your words: when nothing is at risk, the routine stays quiet; when something is, the chase goes out. That’s the difference between a useful routine and noise. ## Schedules and event triggers Not every routine runs on a clock. Routines can also fire when something happens — when a document lands in the project, when someone joins, when a claim is raised. Say the trigger the way you’d say it to a person: “when someone joins the project, send them the site induction pack”, “when a new drawing set is uploaded, summarise what changed and post it to Today.” For things the platform can’t watch directly, the assistant proposes the closest reliable shape — usually a frequent scheduled check — and tells you that’s what it’s doing. ## Updating and cancelling from chat Existing routines are manageable from the same conversation: “change the morning summary to 6:30”, “pause the claim chaser over Christmas”, “delete the Friday reminder routine”. The assistant confirms which routine it matched before changing anything, and edits are drafted for approval just like creation. Every routine the assistant builds also appears in the Routines list, where you can open it in the visual builder — chat-created and builder-created routines are the same thing underneath. ## The “Automate this” suggestion The assistant notices patterns in what you ask it. If you’ve requested the same thing several mornings running, an `Automate this` chip appears under the answer — one click starts the routine draft with the schedule already inferred from when you actually ask. You review and confirm it like any other routine. Where routine runs show up Every run writes to the project [timeline](/en/docs/guides/projects/timeline) — what ran, what it found, what it sent — so routines stay as auditable as work done by hand. --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/creating-routines --- ## Drafting Documents Source: https://www.alloovium.com/en/docs/guides/ai-assistant/drafting AI ASSISTANT # Drafting Documents Ask for a document the way you'd brief a colleague — “draft a delay notice for the wet weather this month”, “put together an evidence pack for VO-118” — and the assistant writes it from the project record, with a citation on every claim. Then refine it in the same conversation: shorter, firmer, one page. Video · \~2 min Draft it in the conversation One document from ask to saved file: requesting a delay notice in the chat, the draft streaming into the side panel with citation chips accumulating, clicking a citation to check the source passage, asking 'keep it to one page' and watching the draft tighten, then saving it to the project documents. ## Overview Document generation isn’t a separate tool you have to go find — it lives in the same conversation as everything else. When an answer turns into “now I need to put this in writing”, the next message can be the brief: what the document is, who it’s for, what it should cover. The assistant plans the document, pulls the facts from the project’s documents and live records, and writes it while you watch. The draft opens in a panel beside the chat, so the conversation stays available while it writes — and generation keeps going even if you navigate away. A draft dock keeps in-progress documents within reach until you’re done with them. ## Asking for a document The brief can be as short or as specific as you like: “draft a response to RFI-214”, “a one-page position paper on the facade delay for the client”, “an evidence pack for VO-118”, “this month’s progress summary for the board”. Constraints in the ask are honoured — if you say one page, you get one page, not a ten-section report. Because the assistant drafts from the project corpus, the document arrives already knowing the contract clauses, the variation history, the correspondence and the programme — you review substance, not fill in blanks. Say who it's for “For the client”, “for the subbie”, “internal” — the audience changes the tone and what gets included, so naming it in the ask saves a refinement round. ## Evidence on every claim Generated documents carry the same citation discipline as answers: every factual claim in the draft — a date, a cost, a clause reference — links back to the source it came from. Click a citation chip and the source document opens at the cited passage, highlighted. That makes review fast: you’re checking the assistant’s receipts, not re-deriving the facts yourself. ## Your templates, matched automatically If your library holds a template for the kind of document you asked for — a delay notice, a site instruction, a monthly report format — the assistant matches it automatically and drafts into your format instead of a generic one. The draft shows which template it followed, and the saved document keeps that lineage. For form-style templates where values get filled into fixed fields, see [Template Filling](/en/docs/guides/template-filling). ## Refining without starting over The draft stays live in the conversation, so edits are just the next message: “make the summary shorter”, “add a cost breakdown table”, “firmer on the notice requirements”, “keep it to one page”. The assistant revises the existing draft — targeted changes touch only the sections they concern, and the citations stay attached through the edit. If a revision goes the wrong way, ask it to regenerate a section — or the whole document — with the new direction baked in. ## Saving and sending When the draft is right, save it to the project — it becomes a normal project document: filed, searchable, citable in future answers. From there it can go out like any other document: attached to an email, or shared with another company via [Exchange](/en/docs/guides/exchange) — both of which you can also ask for in the same conversation (see [Taking Action](/en/docs/guides/ai-assistant/actions)). Longer-form generation For the full drafting surface — reports, monthly reports, and generation outside the chat — see [Document Generation](/en/docs/guides/document-generation) and [Reports & Generation](/en/docs/guides/reports). --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/drafting --- ## Sharing Conversations Source: https://www.alloovium.com/en/docs/guides/ai-assistant/sharing AI Assistant # Sharing Conversations Share AI chat threads with team members so they can view questions, answers, and citations without re-running the queries. ## Sharing Conversations Chat threads in the AI Assistant can be shared with other members of your team. When you share a conversation, recipients can view the full exchange — questions, answers, and citations — without needing to re-run the queries themselves. To share a conversation, use the Share button at the top of the chat, or open the conversation's three-dot menu in the sidebar and select "Share". Add teammates by name; everyone with access is listed in the same dialog, where access can also be removed. Recipients find shared conversations in History, marked with who shared them. Permissions Sharing a conversation does not automatically share the underlying documents. Recipients who are not already members of the project will see the answers and citations, but may not be able to click through to view the source document. --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/sharing --- ## Voice Input Source: https://www.alloovium.com/en/docs/guides/ai-assistant/voice-input AI Assistant # Voice Input Speak your questions instead of typing them. Voice input is ideal for quick queries during site visits or when dictating longer questions. ## Voice Input The AI Assistant supports voice dictation, allowing you to speak your questions instead of typing them. This is particularly useful during site visits or when you want to quickly dictate a longer question. 1. 1 Click the microphone icon The microphone icon is located on the right side of the message input field. Click it to begin recording. 2. 2 Speak your question Speak clearly into your microphone. Your words appear as text in the input field in real time. 3. 3 Stop recording Click the microphone icon again to stop. Edit the transcription if needed. 4. 4 Send the message Press Enter or click Send to submit your question to the assistant. Video · \~1 min Ask by voice Tapping the microphone, speaking a full question while the words appear live in the ask bar, then stopping — the transcript is cleaned up server-side into a polished final question before it is sent. Where voice input works Voice input is available in every ask bar in the web app on Chrome and Edge. Your browser will ask for microphone permission the first time. --- Canonical page: https://www.alloovium.com/en/docs/guides/ai-assistant/voice-input --- ## Claim Evidence Source: https://www.alloovium.com/en/docs/guides/claims-evidence Routines # Claim evidence Build a claim evidence pack from the records the project already holds — by asking, with every answer cited to source — then draft the formal claim from it. ## Claim evidence A claim evidence pack is the set of project records that proves what happened, what it cost, and what the contract entitles you to: instructions, correspondence, site records and photos, delay records, cost records, and the governing clauses. You do not need to generate a document to see your evidence. Ask the assistant first — every answer is cited to the exact source record, so the answer itself is an evidence list with pointers to proof. Draft the formal claim only once the evidence is in front of you. ## Gather the evidence by asking Ask across the project's records one evidence category at a time. Each answer carries click-to-source citations, and each citation points straight back to the record that proves the item: * **Instructions** — "Show every site instruction and direction about the level 3 pour, with dates." * **Correspondence** — "Find the emails and letters about the June weather delays." * **Site records and photos** — "What do the site diaries record for 12–26 June — weather, crew, and progress?" * **Delay records** — "What changed in the programme around the level 3 pour, and when was it first flagged?" * **Cost records** — "List the cost records touching the level 3 pour — dayworks, plant standing time, supplier invoices." * **Contract clauses** — "Which clauses govern delay notice and extension of time, and what do they require?" ## Check the notice position Contractual commitments captured from the project land as action items with a notice clock — the governing clause attached and a countdown to the deadline. Check them before drafting: a strong evidence pack served after the notice window closed is a weak claim. ## Draft the claim from the evidence With the evidence in front of you, draft the formal document with Document Generation: describe the claim — the event, the date range, the affected works — and ask for the evidence organised by category. Pick the project so the draft grounds in the same records, and the result is an editable draft with citations. Evidence is built as the job runs The strongest packs are captured daily, not reconstructed later — site diaries, meeting notes, and a read inbox mean the record already exists when the claim comes. A human sends it, not the AI The draft is a starting point. Review every cited item, price it yourself, and check it against the notice clause before anything leaves the building. --- Canonical page: https://www.alloovium.com/en/docs/guides/claims-evidence --- ## Teams & Groups Source: https://www.alloovium.com/en/docs/guides/collaboration COLLABORATION # Teams & Groups Alloovium is built for teams. Access is granted through user groups — a group collects people and carries a set of permissions, and everything someone can see and do follows from the groups they belong to. This page covers the people directory, groups, and the two invite flows. ## Overview There are no per-person roles to assign when you bring someone in. Instead, you pick the groups they should join, and the groups do the rest: each group carries a permission matrix that says what its members can do across the workspace, and projects grant access to groups rather than to a long list of individuals. This keeps access manageable as the team grows. When a new engineer starts, you add them to the same groups as the last engineer — you do not reconstruct their permissions by hand. When a group’s permissions change, everyone in it changes at once. Video · \~2 min Invite your team the right way Inviting a new teammate from Organization → People: staging their email, keeping the preselected Read only group, adding a second group and reviewing its permission preview, then sending — followed by adding a group to a project so the whole crew gets access at once. ## How access works Three ideas carry the whole model: * **Groups grant access.** A person’s permissions are the union of the groups they belong to — someone in several groups gets the highest level any of them carries. There is no separate role field to keep in sync. * **Every member belongs to at least one group.** The organisation has a default `Read only` group that acts as the access floor — read-only everywhere — so nobody ever lands in the workspace with undefined access. * **Admin is a group too.** Membership of the `Organisation admins` group is what makes someone an organisation admin. Adding a person to that group grants admin; removing them revokes it. The detail of what each group can do — the permissions matrix of areas and levels — is covered on [Permissions](/en/docs/guides/collaboration/permissions). ## The people directory `Organization → People` is the home for everyone in your organisation. It lists members with their group memberships, lets admins see who is active or inactive, and holds the invite flow. Organisation admins also get a member detail view with each person’s recent activity and session controls. The same page carries a `Groups` tab where groups are created, renamed, and managed. Opening a group leads to its detail page: the roster of members on one side and the group’s permission matrix on the other. ## User groups A group is a named collection of people that carries permissions. Typical groups mirror how the business already works — _Site engineers_, _Contracts team_, _Subcontractors_ — and the same groups are reused everywhere: org invites, project access, and the Exchange addressing rail. ### Two groups exist by default `Read only` is the organisation’s access floor. Everyone in the organisation belongs to it, and it grants read-only access across the platform. It is created automatically the first time it is needed. `Organisation admins` is the admin tier. Its members manage people, groups and organisation settings. Because admin is expressed as group membership, you can see at a glance exactly who holds it. ### Creating a group Groups are created from the Groups tab, or inline in the middle of an invite. When you create one you can start blank or use `Copy permissions from…` to copy an existing group’s permission matrix as the starting point, then adjust it on the group’s detail page. ## Project membership Projects grant access to groups. Adding a group to a project gives every member of that group access to it, at a share tier the project controls — `View`, `Edit` or `Admin`. Each group on a project has its own detail page showing that access tier alongside the roster, so you can always answer “who can touch this project, and how?” People invited to a project directly also receive an individual membership at a view-only floor, so an invitee always has at least read access to the project — the groups they were invited through carry anything above that. Project access tiers are covered in more detail on [Permissions](/en/docs/guides/collaboration/permissions). ## Inviting someone to the organisation Org invites live in `Organization → People`. There is no member/admin choice to make — you stage the people, pick their groups, and send. 1. 1 Stage the recipients Type an email address (or several — they can be pasted as a batch) and stage each one. Existing members and recent invitees are suggested as you type. 2. 2 Pick their groups The Groups selector starts with the Read only group preselected. Keep it for read-only access, or add the groups that match their job. Selecting Organisation admins is how you invite an admin. 3. 3 Create a group inline if you need one Choose "New group with its own permissions" from the group menu to mint a group without leaving the invite — name it, optionally copy another group's permissions, and it joins the selection. 4. 4 Review what the groups grant The selected groups' actual permission matrices are shown in the invite panel, so you can verify what you are granting before you send. 5. 5 Send Each recipient gets an email invitation. When they accept, they land in the groups you chose — their access is already set. Admins are invited through a group There is no admin checkbox. Adding the **Organisation admins** group to an invite is what makes the invitee an organisation admin — the same group that controls admin for existing members. ## Inviting someone to a project Project invites work the same way, scoped to the project. There is no role dropdown — access comes from the project’s groups, and the invite asks which groups the person should join. 1. 1 Open the project invite From the project's People area, start an invite and stage the recipients. 2. 2 Pick from the project's groups A multi-select lists groups; pick at least one. Each group's share tier and permissions apply to the people who join it. 3. 3 Or mint a new group inline The same "New group with its own permissions" option is available here, with the option to copy permissions from an existing group. 4. 4 Send Invitees receive at minimum read access to the project (the view-only floor), with the picked groups carrying their real access. Prefer whole groups over one-off invites If a whole crew needs a project, add their group to the project instead of inviting each person. New members of that group inherit the project automatically, and removing the group removes the access in one move. ## What the Read-only default means The preselected `Read only` group is a deliberate safe default. Someone invited with only that group can open and read what is shared with them — documents, projects they are given, answers from the assistant — but cannot change anything, share anything, or manage anyone. That makes the low-effort path the safe path: sending an invite without thinking about groups never over-grants. When the person’s job calls for more, add them to a group that carries it — from their row in the people directory or from the group’s detail page — and the extra access applies immediately. --- Canonical page: https://www.alloovium.com/en/docs/guides/collaboration --- ## Permissions Source: https://www.alloovium.com/en/docs/guides/collaboration/permissions COLLABORATION # Permissions Permissions in Alloovium live on groups. Each group carries a matrix — areas of the platform down the side, permission levels across the top — and a person's effective access is the union of their groups. This page explains the matrix, the admin tier, and how project access layers on top. ## Overview There are no per-person permission settings. Everything someone can do is derived from the groups they belong to (see [Teams & Groups](/en/docs/guides/collaboration)), and a person in several groups gets the highest level any of them carries. That makes access reviewable: to know what a person can do, read their groups; to change what a whole role can do, edit one matrix. Permissions are enforced at the data layer, not just in the interface. The assistant only retrieves and cites documents the asker is allowed to see, and the same access rules gate uploads, sharing and administration. ## The permissions matrix Each group’s matrix is managed from the group’s detail page at `Organization → People → Groups`. The rows are areas of the platform — Documents, Projects, Company library, Routines, Compliance, AI assistant, Meetings, Exchange, and Administration. For each area you pick a level: | Level | What it means | | --------- | -------------------------------------------------------------------------------------------------------- | | None | The area is off for this group. | | Read only | Members can view the area but not change it. | | Standard | Members can do the everyday work — view plus create, edit, send or resolve, depending on the area. | | Admin | Members also get the area's management capabilities — approving, deleting, managing members or settings. | Levels are cumulative — each one includes everything below it. Under each area row the matrix can be expanded to the individual capabilities a level grants, and those capabilities can be adjusted directly; a group whose capabilities no longer match a preset level is marked `Custom`. ## An example matrix A typical mid-size builder might run groups like these. Each cell is the level that group holds for that area: | Area | Read only | Site engineers | Contracts team | Organisation admins | | -------------- | --------- | -------------- | -------------- | ------------------- | | Documents | Read only | Standard | Admin | Admin | | Projects | Read only | Standard | Standard | Admin | | Routines | Read only | Standard | Standard | Standard | | Compliance | Read only | Standard | Read only | Admin | | AI assistant | Read only | Standard | Standard | Standard | | Exchange | Read only | Standard | Admin | Admin | | Administration | None | None | None | Admin | A site engineer in both _Read only_ and _Site engineers_ gets the higher of the two in every area. Notice that only the admins group holds anything in Administration — that area carries company settings, access management, billing, integrations and data export, and it is what gates the organisation settings pages themselves. ## Organisation admins Organisation admin is not a flag on a user — it is membership of the `Organisation admins` group. Members of that group manage people, groups and organisation settings, and the group is maintained automatically so it always reflects exactly who holds admin. This has a practical consequence: granting or revoking admin is a roster change on one group, visible on that group’s detail page, rather than a setting scattered across user profiles. Inviting an admin is the same move — you include the admins group in the invite. You cannot lock yourself out You cannot remove yourself from the Organisation admins group — dropping your own membership would lock you out of the very pages used to manage access. ## Project access tiers Group permissions apply organisation-wide. On top of that, each project grants its own access tier to each group that has been added to it, set from the group’s detail page on that project: | Tier | What it grants | | ----- | ------------------------------------------------ | | View | Can open and read the project's files. | | Edit | Can also add, edit and organise files. | | Admin | Full control, including managing who has access. | The two layers work together: the project tier says how far a group can go on that project, and the org-wide matrix says what its members can do across the platform. A group can be Edit on one project and View on another without touching its matrix. ## The view-only floor Nobody in Alloovium has undefined access. Every member of the organisation belongs to the default `Read only` group — the organisation’s access floor, read-only everywhere — and a person invited to a project always receives at minimum read access to it, regardless of which groups the invite carried. The floor only sets the minimum. Real working access comes from the groups above it, and removing someone from those groups drops them back to the floor rather than into a broken state. To take someone out entirely, an organisation admin deactivates them from the people directory. --- Canonical page: https://www.alloovium.com/en/docs/guides/collaboration/permissions --- ## Sharing Source: https://www.alloovium.com/en/docs/guides/collaboration/sharing COLLABORATION # Sharing Three kinds of sharing cover most of the day: giving people access to a project, letting a teammate read an assistant conversation, and sending documents to parties outside your organisation. Each one has its own flow, and each one leaves a record. ## Overview Inside the organisation, sharing rides on groups — the same groups that carry permissions everywhere else (see [Teams & Groups](/en/docs/guides/collaboration)). Sharing a project means adding people or groups to it; there is no separate per-project role system to maintain. Outside the organisation, sharing goes through [Exchange](/en/docs/guides/exchange), which sends controlled document issues to external parties with provenance — who sent what, to whom, and when. ## Sharing a project A project is shared by inviting people or groups to it. The invite has no role dropdown: you pick from groups via a multi-select, and each group’s share tier and permissions determine what its members can do there. 1. 1 Stage people, groups, or both Add email addresses for individuals, or pick whole groups. A group is the better unit when a crew needs access — its future members inherit the project automatically. 2. 2 Pick at least one group Access comes from the project's groups. If none fits, create one inline with "New group with its own permissions", optionally copying an existing group's permissions as a starting point. 3. 3 Send Individual invitees receive at minimum read access to the project; the groups they join carry anything above that. Existing members are added straight away, new people get an email invitation. Once shared, each group on the project has a detail page showing its access tier — `View`, `Edit` or `Admin` — and its roster, so reviewing who can reach the project is one page, not an audit. Tiers are described on [Permissions](/en/docs/guides/collaboration/permissions). ## Sharing conversations Assistant conversations can be shared with teammates. Sharing a chat lets the recipient read the conversation — the question, the answer, and the citations that back it — which is often the fastest way to hand over a finding: instead of paraphrasing what the assistant said about a clause, share the conversation and let them follow the citations themselves. Conversations shared with you are collected under a shared view in your chat history, and a share can be revoked. Your chats stay yours until you share them — nothing is visible to anyone else by default. ## Sharing documents externally Sending documents to people outside your organisation — a subcontractor, the superintendent, a client — is Exchange’s job. An Exchange send is a controlled issue rather than an email attachment: you choose the documents, address the recipients, and the send is recorded so there is never a dispute about which revision went out, or when. Recipients can access what you sent them without needing a seat in your workspace, and documents you receive through Exchange become part of what the assistant can retrieve and cite. The full flow — the send picker, addressing, and received documents — is covered in [Exchange](/en/docs/guides/exchange). Internal vs external in one rule If the person works in your Alloovium organisation, share through groups and projects. If they do not, send through [Exchange](/en/docs/guides/exchange) — it exists precisely so external parties get the documents without getting your workspace. --- Canonical page: https://www.alloovium.com/en/docs/guides/collaboration/sharing --- ## The Commercial Control Loop Source: https://www.alloovium.com/en/docs/guides/commercial-control COMMERCIAL # The Commercial Control Loop Most entitlement isn't lost on the merits — it's lost in the gaps: the instruction nobody wrote up, the notice served a day late, the approved variation that never made a claim. Alloovium runs one loop across those gaps, from the moment something happens to the day the money is certified and paid. ## Overview A commercially significant event — a latent condition, a directed change, a delay — starts a race the moment it happens. The contract's notice clock is already running; the evidence is already going stale; and every day the event stays undocumented, it drifts closer to becoming free work. The features described across these docs are links in one chain built for that race: 1. 1 Capture The event is caught where it first appears — a meeting, a knock-off call, an email, a document received through Exchange — in the words it arrived in. 2. 2 Clock The contract clause it engages and the notice period it starts are identified while there is still time to act, and surface on Today with the deadline visible. 3. 3 Paperwork The notice or variation is drafted from the record, in your own templates, every factual line cited — and waits at a gate for a human. 4. 4 Proof The send leaves a receipt, the variation enters the register, and the record carries the event through claim, certification and payment — every transition stamped. Nothing in the loop is automatic past the gate: Alloovium prepares, you approve. What the loop guarantees is that nothing expires quietly — if a clock is running, you can see it, and if paperwork is owed, it is already drafted. Video · \~4 min No missed notice, no undocumented change, no free work One event walked through the whole loop — a latent condition surfaces on the evening site call, the notice clock lands on Today with the clause behind it, the cited notice is drafted and approved, the variation enters the register, and the reconciliation confirms it was never left as free work. ## Catch the event where it first appears Commercial events almost never arrive labelled. They surface as a sentence in a [meeting](/en/docs/guides/meetings), a remark on the [knock-off call](/en/docs/guides/site-calls), a request buried in an [email thread](/en/docs/guides/email-queries), or a document received through [Exchange](/en/docs/guides/exchange) before anyone has filed it. Each of those surfaces feeds the same record: the meeting review queue tags the instruction that is really a variation, the site call drafts the latent-condition entry while the evidence is hours old, and email and Exchange keep the cross-company facts searchable and citable even when they never reach the formal document system. The gap this closes Projects lose entitlement in the days between what the site knew at half past four and what the commercial record found out on Friday. Capture at the source means the loop starts the same day the event does. ## Know the clock you are on Once an event is on the record, the question is contractual: what does it engage, and how long do you have? Ask the contract directly — _“what must we notify within ten business days?”_ — and the clauses come back cited. When an event carries a notice obligation, the deadline surfaces on [Today](/en/docs/guides/today) as a card with the clause behind it and the days remaining in front of you, while there is still time to do something with them. ## Draft the paperwork with receipts The notice, the variation, the show-cause response — Alloovium [drafts them from the record](/en/docs/guides/document-generation): the facts from the diaries and correspondence, the clause references from this contract, the format from your own [library templates](/en/docs/guides/template-filling). Every factual line carries a citation, so reviewing the draft means checking receipts, not re-deriving facts. And the draft waits at a gate — nothing ships your words without you. ## Send it so it cannot be disputed A notice that cannot be proven served is barely a notice. [Exchange](/en/docs/guides/exchange)keeps the trail on every send: who received what, which version, when it was opened, who signed — hash-bound to the exact content, so “we never got that” stops being an argument anyone can make. ## Keep it on the record until it is paid Preserved entitlement still leaks at the back of the loop: the approved variation that never made a claim, the claim certified short that nobody chased. The [variations register](/en/docs/guides/variations) carries each event with its notice receipts attached, every transition writes to the [project timeline](/en/docs/guides/projects/timeline), and the [claim evidence](/en/docs/guides/claims-evidence) stays assembled as it happens rather than reconstructed under pressure. Ask for the reconciliation — approved versus claimed versus paid — and the gaps come back as a cited list, so the recovery starts the same afternoon. ## Let routines keep watch The loop holds because [routines](/en/docs/guides/routines) run it when nobody is looking: the daily check for events that start clocks, the subcontractor chased for the substantiation that is due, the month-end report assembled before the cutoff. Whatever a routine produces waits at the same approval gate — machines do the legwork, you keep the signature. Where to go next For the loop from a commercial manager's seat — ten workflows with a video each — see [Use cases for commercial managers](/en/docs/use-cases/commercial). For the individual links in the chain, start with [Variations](/en/docs/guides/variations) and [Claim Evidence](/en/docs/guides/claims-evidence). --- Canonical page: https://www.alloovium.com/en/docs/guides/commercial-control --- ## Claims register & cost position Source: https://www.alloovium.com/en/docs/guides/commercial-control/claims-register COMMERCIAL # Claims register & cost position The claims register is the one page that answers where the money is. Every progress claim on the project sits in it with a lifecycle state, the cost-position strip at the top totals what has been claimed, certified and paid, and a reconciliation panel surfaces the variations you won but never billed. ## Overview On most jobs the money position lives in three spreadsheets and one person’s head. The claims register replaces that with a typed record: each claim carries its number, counterparty, period, the amounts claimed, certified and paid, and a status that can only move along the legal edges of the claim lifecycle. The register reads directly from those records — it is a rollup, never a parallel source of truth. You reach it from the **Registers** band in the project sidebar, alongside the other registers for the project. The same band holds the [obligations register](/en/docs/guides/compliance/obligations-register) and the commercial change register described on [The Commercial Loop](/en/docs/guides/commercial-control). Why the position matters Construction pays a 33.5-day average against a 27.4-day all-industry average, and the slowest 15% of payers take 68 days against 34-day terms (Payment Times Reporting Scheme, Cycle 9, Jan–Jun 2025). A live certified-but-unpaid figure is how you see that exposure before it becomes a cashflow problem. The claims register for a project — the cost-position strip across the top showing approved variations, claimed-to-date and an orange-highlighted Owed-to-us figure, with the claims table below it mixing submitted, certified and paid status badges. ## The cost-position strip The strip pinned above the table is the project’s money position at a glance. It is computed from the claims and variations on the register, so it is always as current as the records underneath it. | Figure | What it totals | | --------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Approved variations | The value of variations the client has approved. | | Pending variations | The value of variations still in draft or submitted. | | Claimed to date | The sum claimed across your outgoing progress claims. | | Owed to us | Certified minus paid on your outgoing claims — the money you are entitled to that has not arrived. This is the cashflow exposure line, and it is highlighted whenever it is above zero. | | Subbie claims to date | The sum claimed against you on incoming claims. | | Owed to subbies | Certified minus paid on incoming claims — what you owe downstream and have not yet paid. | The strip tracks both directions of the money — what you have claimed up the chain and what subcontractors have claimed against you. A figure only earns its slot once it is real: zero-value counters stay hidden, and until the project has any live figure the strip does not render at all. ## Raising and detecting claims Claims arrive on the register two ways. ### Raise one yourself The `New claim` action opens a slide-over where you set the title, direction, counterparty, contract reference, claim period, the amount claimed, and — for an incoming claim — the date a response is due. Claim numbers such as `PC-005` are generated automatically, in sequence per project and direction. ### Let the documents register themselves When a payment claim document is uploaded to the project, Alloovium detects it during ingestion and registers the claim for you. The register refreshes itself while open, so a claim that arrives in the document library shows up in the table without you re-keying anything. Each claim opens into a workspace showing its line items and the evidence behind them, with line-level verification against the project record — a line is marked supported, contradicted, or flagged for you to verify, each with the citations it was judged on. ## The claim lifecycle A claim occupies exactly one of six states, and the register only offers the transitions that are legal from the state it is in. The same rules are enforced server-side, so the lifecycle cannot be skipped. | Status | What it means | Can move to | | --------- | ----------------------------------------- | ------------------------------ | | Draft | Under composition — not yet issued. | Submitted | | Submitted | Issued to the counterparty. | Certified, Scheduled, Disputed | | Certified | Assessed by the certifier. | Scheduled, Paid, Disputed | | Scheduled | Answered with an issued payment schedule. | Paid, Disputed | | Disputed | Contested by either side. | Submitted, Certified, Paid | | Paid | Money received — the end state. | — | The `submitted → scheduled` edge is the security-of-payment path: an incoming claim answered with an issued payment schedule needs no separate certification step. Every transition is recorded as an event on the project timeline, so the register doubles as an auditable history of when each claim moved and who moved it. Disputed is a state, not a dead end A disputed claim keeps its record — the amounts, the evidence and the correspondence — and can move back to submitted, certified or paid as the dispute resolves. See [Claim Evidence](/en/docs/guides/claims-evidence) for assembling the paper trail behind a contested figure. ## Incoming claims and the response clock The register tracks claims in both directions. An **outgoing** claim is money you are claiming; an **incoming** claim is one served on you — and an incoming claim starts a clock. Set the payment jurisdiction for the project (all eight Australian states and territories, plus New Zealand, are supported) and the register shows a response-clock chip against each incoming claim, counting down the statutory window in which a payment schedule must be served. Assessing the claim line by line and generating the schedule itself is covered on [Payment schedules](/en/docs/guides/commercial-control/payment-schedules). ## Approved but never billed The quietest loss on a project is a variation that was argued, won, and then never appeared on a progress claim. The register runs that reconciliation for you: it checks each approved variation against the claims on the register and surfaces the ones that appear on none of them, with a total. 1. 1 Approved variations are matched against claims Each approved variation is checked for a reference on the progress claims of the project. 2. 2 Unmatched variations are surfaced Variations that appear on no claim are listed with their value, summed into an approved-but-never-billed total. 3. 3 Weak matches are flagged, not counted Where the match is uncertain, the row is labelled for you to check rather than silently included in the total. It is a text match — read the method note There is no stored link between a variation and the claim that billed it, so this reconciliation matches by reference text and says so on the panel. Treat the “check” rows as prompts, and confirm before you re-bill. When nothing is unbilled, the panel stays silent. ## Exporting the register The register downloads as a spreadsheet (XLSX) or CSV from the toolbar, so the position can go into a month-end report or a client meeting without re-keying. For a narrative report built from the same records, see [Monthly Report](/en/docs/guides/monthly-report). ## Related pages [The Commercial Loop](/en/docs/guides/commercial-control) — how change signals become typed commercial records. [Variations](/en/docs/guides/variations) — capturing a change and tracking it to approval. [Claim Evidence](/en/docs/guides/claims-evidence) — assembling the cited paper trail behind a claim. [Payment schedules](/en/docs/guides/commercial-control/payment-schedules) — answering an incoming claim inside the statutory window. --- Canonical page: https://www.alloovium.com/en/docs/guides/commercial-control/claims-register --- ## Delay substantiation & evidence packs Source: https://www.alloovium.com/en/docs/guides/commercial-control/delay-substantiation COMMERCIAL # Delay substantiation & evidence packs An extension-of-time claim lives or dies on its particulars — what happened, when, what the contract says about it, and what it did to the programme. Alloovium assembles that case from the records you already keep: site diaries, weather, contract clauses and the schedule, every factual line cited back to its source. ## Overview The traditional way to substantiate a delay is a forensic report: a consultant reconstructs the record months or years after the event, at a cost that routinely runs to tens of thousands of dollars per report — and the reconstruction is only as good as the paperwork that survived. Alloovium inverts that. Because the diaries, weather records, contract and schedule are already in the project corpus, the substantiation is assembled from contemporaneous records while they are still fresh, and every claim in the pack carries a citation you can click through to the source document. The output is an evidence pack for a specific delay event: the narrative of what happened, the diary entries and weather records that prove it, the contract clauses that entitle you to time, and the schedule analysis that shows the delay actually consumed the programme rather than available float. Alongside the pack, every notice you issue for the event carries a served-on trail, so timeliness is provable, not remembered. Where this fits in the loop Delay substantiation is one link in [the commercial control loop](/en/docs/guides/commercial-control): the event is captured, the notice clock surfaces, the notice ships with receipts — and this page covers the part where the entitlement gets proven, not just preserved. ## The Delays & EOT register Everything on this page starts in one place: the **Delays & EOT** register, in the **Registers** band of the project sidebar alongside the project’s [other registers](/en/docs/guides/projects/registers). Its subtitle states the job plainly — _“Each delay’s cited evidence pack, critical-path impact and served-notice trail.”_ One row per delay event, and the primary action on the page is **Log a delay**. | Column | What it carries | | ------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Delay | The event itself — what happened, as one row on the record. | | Cause | A controlled list, so the register groups and filters honestly: Weather, Design / RFI, Principal instruction, Access, Subcontractor, Materials, Industrial, Latent condition, Other. | | Days claimed | The extension of time being sought for this event. | | Notice clock | Time left to serve — "Due in Nd", "Due today", "Overdue", or "—" where no deadline has been computed. The row detail spells it out as "Notice due {date}". | | Status | Where the event sits in its lifecycle, from draft through to granted or rejected. | ### The status lifecycle A delay event moves through four states, and the register is ordered around the fact that the middle two are where the deadlines live. | Status | What it means | | ------------------ | ------------------------------------------------------------------------------------------------ | | Draft | The event is on the record but nothing has been served — the notice clock is the thing to watch. | | Notified | A notice has gone out; the entitlement is preserved and the served-on trail exists. | | Claimed | The extension of time has been claimed, with the evidence pack behind it. | | Granted / Rejected | The assessed outcome, kept against the event rather than in an inbox. | ### The evidence pack in the row detail Opening a row shows the delay’s evidence pack and when it was put together — `Assembled {date}`. The two empty states are deliberately honest rather than reassuring. _“Not assembled yet — assemble the evidence to ground the notice.”_ means the case has not been built. The second one matters more: _“Assembled {date} — no supporting records were found for this window.”_ That is a pack that ran and came back empty, and it is the most useful thing the register can tell you — the record for those days is thin, and you know it **before** you serve rather than under challenge at adjudication. The Delays & EOT register for a project — the subtitle 'Each delay's cited evidence pack, critical-path impact and served-notice trail', a 'Log a delay' action, and rows across Delay, Cause, Days claimed, Notice clock and Status columns, with notice clocks reading 'Due in 4d', 'Due today' and 'Overdue', and an expanded row showing its assembled evidence pack. On a project with nothing logged the register says so — _“No delays recorded”_ / _“Log the first delay to start assembling its evidence pack and notice.”_ Ask about delays in chat The register is readable by the assistant, read-only and filtered to what you are permitted to see. It can list the delay events on a project and, separately, the notices that were actually served — so _“which delays are noticed but not yet claimed?”_ is answered from these rows rather than reconstructed. ## What an evidence pack cites A delay evidence pack pulls from four kinds of record, each contributing a different part of the case. | Source | What it proves | | ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------- | | Site diaries | The contemporaneous account — what happened on the affected days, in the words recorded at the time, including voice diaries filed from site calls | | Weather records | The conditions on the days claimed, matched to the delay window rather than asserted from memory | | Contract clauses | The entitlement itself — the extension-of-time clause engaged, the notice regime it imposes, cited to the executed contract for this project | | Schedule | The programme impact — where the delay landed, what it pushed, and how it reconciles against available float | Every factual line in the assembled pack carries a citation back to the source document, so reviewing the pack means checking receipts rather than re-deriving the case. The same records that answer _“what happened in week 14?”_ in the assistant are the ones the pack cites — one corpus, one set of facts. ## Float reconciliation A delay that fits inside available float is not an extension-of-time entitlement, and an assessor will check. The pack does that check first: each delay is reconciled against the schedule to establish whether it exceeds the float available on the affected work, and the reconciliation is part of the pack rather than an argument you have to construct under challenge. That cuts both ways deliberately. A pack that shows a delay consumed only float is a claim you chose not to send — which is cheaper than a claim that came back rejected with your credibility spent. ## The served-on trail Entitlement to time usually hangs on a notice served within a contractual window. For every notice issued through Alloovium, the record keeps a served-on trail: what was served, to whom, and when. When the claim is assessed months later, timeliness is a record you point to, not a recollection you defend. ### Notice first, substantiation behind it The notice preserves the entitlement; the pack proves it. Because both ride the same record, the notice you served and the evidence behind it stay attached to the same delay event — the assessor sees one coherent trail from the day it happened to the day it was claimed. ## The delays page The **Delays & EOT** register described above is also the working view for the month-end pass: the delay events on the record, their status, the notices served against them and the evidence assembled so far. It answers the question the question every commercial manager asks at month end — _“which delays are live, which are noticed, and which are still unsubstantiated?”_ 1. 1 A delay event lands on the record From a site call, a diary entry, an email or a direct entry — captured while the evidence is hours old, not months. 2. 2 The notice goes out with a served-on trail The entitlement is preserved inside the contractual window, and service is provable. 3. 3 The evidence pack assembles the case Diaries, weather, clauses and schedule cited into one pack, with the float reconciliation done. 4. 4 The claim goes in substantiated Particulars, not assertions — assembled at claim time, when they can still be considered. ## Why particulars decide EOT claims The case law is blunt about what kills extension-of-time claims, and it is rarely the merits. * In _Rimfire Energy v BSF Co (No 2)_\[2025\] FCA 384, six EOT claims totalling 865 days combined across two respondents were all held invalid — not because they were late, but because the notices lacked the “detailed particulars of the Extension Event” the contract required. Liquidated damages of $1.4M per respondent stood. * In _Builtcom v VSD Investments (No 2)_ \[2025\] NSWCA 134, $30.6M was claimed and $8.47M awarded — roughly $22.1M went unvalued because the supporting documents first arrived with the adjudication application. Evidence served after the claim may never be considered, and the court did not fix it. * Across NSW adjudications in FY24-25, $170.2M was adjudicated against $252.4M claimed on released determinations (\~67%), and only about a third of determinations awarded the full amount — evidence quality is the delta between the claim and the cheque. Substantiation is not optional paperwork A timely notice with thin particulars can be as dead as a late one. The point of assembling the pack at claim time — diaries, weather, clauses, schedule, cited — is that the particulars exist when the contract demands them, not two years later in a consultant's reconstruction. ## Related pages * [The Commercial Control Loop](/en/docs/guides/commercial-control) — the full chain from event to payment. * [Payment schedules](/en/docs/guides/commercial-control/payment-schedules) — assessing incoming claims and responding inside the statutory window. * [Claim Evidence](/en/docs/guides/claims-evidence) — evidence assembly for variations and disputes more broadly. * [Site calls & voice diaries](/en/docs/guides/site-calls) — where the contemporaneous record starts. * [Project registers](/en/docs/guides/projects/registers) — the Registers band the Delays & EOT register lives in, and the other registers beside it. --- Canonical page: https://www.alloovium.com/en/docs/guides/commercial-control/delay-substantiation --- ## SOPA payment schedules Source: https://www.alloovium.com/en/docs/guides/commercial-control/payment-schedules COMMERCIAL # SOPA payment schedules Under Australian security-of-payment legislation, a payment claim you don't answer with a valid payment schedule inside the statutory window becomes a debt for the full amount — disputed parts included. Alloovium takes the incoming claim, walks you through a line-by-line assessment against the record, and generates the statutory schedule with the deadline computed correctly. ## Overview A payment schedule is the respondent's side of the security-of-payment machine: the formal response that says what you will pay, what you won't, and why. It is one of the few documents in construction where the deadline itself is the entitlement — serve it a business day late, or not at all, and the legislation deems you liable for the whole claim regardless of its merits. Alloovium's generator treats it that way. When a payment claim comes in, the clock is computed from the statutory rules for the state — in business days, holiday-aware — the claim is assessed line by line against your project record, and the schedule is generated as a formal document with your assessment and reasons on it, ready to serve inside the window. ## What missing the window costs The consequences are not theoretical; they are the point of the legislation, and the courts apply them mechanically. * In _Manariti Plumbing v Universal Property Group_ \[2025\] NSWCA 135, no payment schedule was served — the result was summary judgment for the full $221,901 plus interest and costs. The claim was held valid despite alleged defects, and s 15(4) barred every defence. * In _Sharvain Facades v Roberts Co_\[2025\] NSWCA 161, the head contractor's payment schedule was one business day late — its contract clause deeming after-5pm service to land the next day was void under the Act — and judgment followed for $3,278,043.27 including interest. * The failure mode is common enough to show up in the regulator's numbers: of 990 NSW adjudication applications lodged in FY24-25, 212 were withdrawn and 44 timed out — “timed out” being pure record-and-deadline failure, before any adjudicator reached the merits. Liability is automatic, not argued Miss the window and the disputed parts of the claim are owed alongside the undisputed ones. There is no do-over: the schedule either exists, on time, with reasons — or the claim becomes a debt. ## From incoming claim to served schedule 1. 1 The claim lands The incoming payment claim is captured against the project, and the statutory response clock starts — visible, in business days, with the due date computed for the relevant state. 2. 2 Line-by-line assessment Each claimed line is assessed against the record: what was instructed, what was delivered, what the contract says. Lines that don’t hold up — dayworks never instructed, a variation never approved, quantities that don’t match delivery records — surface for your decision. 3. 3 Scheduled amounts and reasons For every line you certify below the claimed amount, the schedule records the scheduled amount and the reason for withholding — the content the Act requires for the schedule to do its job. 4. 4 Generate and serve The statutory payment schedule is generated as a formal document. You review it, then serve it the way you serve anything — and if you send it through Exchange, the served-on receipts come with it. The assessment is yours; the generator's job is that nothing required by the statute is missing, nothing is served late because someone was counting days on a wall calendar, and every reason you give is one you can stand behind at adjudication because it points at the record. ## The statutory window, computed correctly Security-of-payment deadlines run in business days, and business days are jurisdiction matters: public holidays shift real due dates in ways a naive calendar count gets wrong. The generator computes the response deadline holiday-aware for the state the project sits in — a deadline that would land on Good Friday does not fall due until the next business day, and the clock you see reflects that. ### Why a day matters _Sharvain_ turned on exactly one business day. When the margin for error is a single day boundary and the downside is the full claimed amount, the day arithmetic is not an administrative detail — it is the entitlement. ## Certifying down on the record Certifying a line below the claimed amount is a commercial decision that should leave a trail. When you certify down, the schedule records the claimed amount, the scheduled amount, and the reason — and that assessment lives on the project record, not in a spreadsheet on someone's desktop. If the claimant takes the line to adjudication, your position and its basis are already written down, dated, and consistent with what you served. Reasons now, not reasons later Under the East Coast security-of-payment model, an adjudication response is generally confined to the reasons given in the payment schedule. A reason you didn't state when you served the schedule is a reason you may not get to argue later — which is why the generator asks for reasons at assessment time, line by line. ## Coverage | Jurisdiction | Status | | --------------------------------------------------------------------- | ----------------- | | NSW — Building and Construction Industry Security of Payment Act 1999 | Supported | | QLD — Building Industry Fairness (Security of Payment) Act 2017 | Supported | | Other Australian states and territories | Not yet supported | NSW and QLD today The generator currently covers New South Wales and Queensland. Projects in other jurisdictions can still assess claims against the record and draft responses through [document generation](/en/docs/guides/document-generation) — the statutory clock computation and schedule form are what the coverage above refers to. ## Related pages * [The Commercial Control Loop](/en/docs/guides/commercial-control) — where claims and schedules sit in the wider money loop. * [Delay substantiation & evidence packs](/en/docs/guides/commercial-control/delay-substantiation) — proving the time side of the entitlement. * [Exchange](/en/docs/guides/exchange) — served-on receipts for the schedule you send. * [Variations](/en/docs/guides/variations) — the register the line-by-line assessment checks against. --- Canonical page: https://www.alloovium.com/en/docs/guides/commercial-control/payment-schedules --- ## Security register — retention & bank guarantees Source: https://www.alloovium.com/en/docs/guides/commercial-control/security-register COMMERCIAL # Security register Every retention and bank guarantee a counterparty holds against you, when it must come back, and the chase that gets it back. Release dates are computed from the trigger in the contract, not remembered — and when one falls due the register drafts the chase for your review. ## Overview Retention is your money sitting in somebody else’s account. It was withheld from progress claims you already earned, and it comes back on a trigger — practical completion, the end of the defects liability period, a date in the contract. Nobody upstream has a reminder set for that trigger. If you don’t either, the money quietly stays where it is. Unclaimed retention is the classic silent write-off: not disputed, not refused, just never asked for. The security register exists to make that trigger visible. Record what is held, record how it is released, and the register computes the release date, moves the item to **Due** when it arrives, and offers you a chase. You reach it from the **Registers** band in the project sidebar under **Security**, at `/projects/{id}/security`. Every user on the project has it. Across the top sits the hero strip: **Money held against you** with a count of items, **Coming back** with the number of open release windows, **Past release** with the number of items that should already have been returned, and **Next release** — the next date anything is due, or _Not scheduled_if nothing on the register has a computable trigger. That last value is the honest one to read first: “Not scheduled” usually means the release terms were never captured, not that nothing is owed. The Security register for a project — the hero strip across the top showing Money held against you with an item count, Coming back, Past release and Next release; beneath it the lens strip (All, Held, Coming back, Past release, Closed) and the register table with columns for Security, Kind, Counterparty, Amount, Release, Release trigger, Release due, State and Chase, mixing cash retention and bank guarantee rows with Held and Due state badges. ## What the register records Two kinds of security sit on the register, and they behave differently when they come back. | Kind | What it is | What coming back looks like | | -------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------- | | Cash retention | Money withheld from your progress claims and held by the counterparty. On Australian jobs this is typically a percentage of each claim up to a cap, with part released at practical completion and the balance at the end of the defects liability period. | A payment back to you — so it belongs on your cashflow forecast, and it is the kind that most often goes unasked-for. | | Bank guarantee | An instrument lodged with the counterparty as security instead of cash. It costs you facility capacity for as long as it is out there, even though no money moved. | The original instrument returned so it can be discharged — not a payment. Until it comes back it keeps consuming your facility. | Each row on the register carries: **Security** (what it is), **Kind**, **Counterparty** (who holds it), **Amount**, **Release** (the release terms), **Release trigger**, **Release due** (the computed date), **State**, and **Chase**. The drawer groups the same fields as _What’s held_, _When it comes back_ and _More details_. The lens strip above the table cuts the register down: **All**, **Held**, **Coming back**, **Past release** (items that passed their release date and have not come back) and **Closed**. Two of those are the ones you live in — _Coming back_ for the month ahead, _Past release_ for the money that is late. Different money, different register Retention withheld from a claim is not the claim. The [claims register](/en/docs/guides/commercial-control/claims-register) tracks what you claimed, certified and were paid; the security register tracks the slice that was deliberately held back and has its own, much later, return date. A claim can be fully paid while retention against it stays outstanding for years. ## How the release date is computed Nobody should be doing this arithmetic in their head at month end. You record the trigger and the offset, and the register computes **Release due (computed)** from them. Three triggers are available. | Release trigger | What it anchors to | Typical use | | -------------------- | ---------------------------------------- | ------------------------------------------------------------------------------------------------ | | Practical completion | The project’s practical completion date. | The first tranche of cash retention, commonly released at or shortly after PC. | | DLP end | The end of the defects liability period. | The balance of retention, and bank guarantees held through the defects period. | | Date | An explicit date you enter. | Security with a hard expiry or a negotiated return date that is not tied to a project milestone. | On top of the trigger you set _days after trigger_ and a **day type**, because “14 days” in a contract rarely means fourteen squares on a calendar. | Day type | How the offset is counted | | ------------- | ------------------------------------------------------------------ | | Calendar days | Every day counts, including weekends and holidays. | | Business days | Weekends and public holidays are skipped. | | Working days | Counted as working days, per the day convention the contract uses. | Copy the day type off the clause, not from habit The gap between calendar and business days on a 90-day release is weeks, and the whole point of the register is that the date is right without anyone re-reading the contract. Take the trigger, the offset and the day type straight from the release clause when you record the item — that is the one moment the accuracy is decided. ## The lifecycle Every item sits in exactly one state, and the state is what the hero strip and the lenses count. | State | What it means | | -------- | --------------------------------------------------------------------------------------------------------------- | | Held | The counterparty holds it and the release date is still in the future. Nothing is required of you yet. | | Due | The computed release date has arrived. This is the moment retention normally starts being forgotten. | | Demanded | You have asked for it back — the chase is under way and the ask is on the record. | | Released | The money came back, or the guarantee was returned and discharged. The end state. | | Expired | The security lapsed — closed without being returned. Recorded rather than deleted, so the history stays honest. | **Held** and **Due** are open; **Released** and **Expired** are closed and fall into the _Closed_ lens. **Demanded** is the state that proves somebody asked — which is exactly the fact that is missing when retention goes unrecovered. ## The release chase When items reach or pass their release date the register raises a _needs you_ band above the table — _“N items are past or near their release date”_ — with a single action: **Start release chase**. That is the register doing the part people don’t: noticing. Nothing is sent automatically Release requests are drafted for your review through the chase — nothing is sent automatically. The chase composes the ask and puts it in front of you; issuing it to the counterparty is always your action. The register states this on the drafts themselves: drafts are prepared for your review, nothing is sent. A chase runs as an escalation ladder rather than a single email. Each chase carries a status — **Not chased** before it starts, **Chasing** once it is running — a **step level** showing how far up the ladder it has climbed, and a **next step date** for when the following beat falls due. As it climbs it accumulates **drafts**, each one a prepared release request you can read and copy out. An empty draft list is not a failure Start a chase and you will often see _“No draft yet — the first one is minted on the next escalation beat”_. Drafting happens on the ladder’s schedule, not the instant you press the button. The chase is live; the first draft is simply pending its beat. The release chase on a cash retention item — the amber needs-you band reading that several items are past or near their release date with the Start release chase action, and the chase panel below showing status Chasing, the current step level, the next step date, and a list of prepared release-request drafts each with a copy action and the notice that drafts are prepared for review and nothing is sent. ## Getting security onto the register An empty register is the normal starting point: _“No security on the register”_ — _“Record the retentions and bank guarantees held against you — the register computes each release date and chases it when it falls due.”_ There are two ways to fill it, and on a project with an executed contract the first is faster. ### Suggest from contract **Suggest from contract**reads the project’s contract and proposes the security it finds — the retention percentage and cap, the release tranches, any guarantee requirement — each proposal carrying its clause reference, shown as `Clause {ref}` so you can check the source before you accept it. On the empty state the same action reads **Find them in the contract**. 1. 1 Run Suggest from contract The register reads the executed contract for the project and proposes the security it finds. 2. 2 Check each proposal against its clause Every suggestion shows the clause reference it came from. Read the clause for the trigger, the offset and the day type before accepting — those three fields decide every date the register computes afterwards. 3. 3 Accept what is real Accepted items land on the register as normal rows with computed release dates. Anything the contract does not actually impose you simply do not accept. ### Add security manually **Add security** opens the drawer for anything the contract does not cover — a guarantee lodged under a side deed, retention held by a client on a variation-only arrangement, security carried over from a predecessor contract. 1. 1 What’s held The kind (cash retention or bank guarantee), the counterparty holding it, and the amount. 2. 2 When it comes back The release trigger (practical completion, DLP end or an explicit date), the days after that trigger, and the day type. Release due is computed from these and shown in the drawer as you set them. 3. 3 More details The supporting context — references and notes that make the row readable to somebody who was not in the negotiation. Record it when it is withheld, not when you want it back The cheapest moment to capture a retention is the claim it was withheld from, while the percentage and the release clause are in front of you. Capturing it two years later, when somebody wonders whether anything is still outstanding, is the expensive version — and it is the version where the money is already lost. ## Counterparty exposure Beneath the register sits the **counterparty exposure** panel: first-party facts about how each counterparty actually pays, drawn from your own record with them. It answers the concentration question the table alone cannot — _who is holding most of my money, and what is their track record of giving it back?_ It is a panel, not a second register. Read it when you are deciding how hard to chase, or how much security you are comfortable leaving with one party across several jobs. ## Ask about it in chat The assistant can read the security register directly through a read-only, permission-filtered lookup, so questions like _“how much retention is still held on Northgate and when does it come back?”_ are answerable in chat. The answer comes from the register rows, not re-derived from documents — so what you hear in chat and what you see on the page are the same numbers. The lookup respects your permissions: it only ever returns security on projects you can already see, and it cannot change anything. Recording, accepting suggestions and starting a chase all stay on the register. ## Related pages * [Claims register & cost position](/en/docs/guides/commercial-control/claims-register) — what was claimed, certified and paid, and the retention’s parent claim. * [The Commercial Loop](/en/docs/guides/commercial-control) — how change signals become typed commercial records. * [Variations](/en/docs/guides/variations) — capturing a change and tracking it to approval. * [Registers](/en/docs/guides/projects/registers) — the full Registers band and how the registers relate to each other. --- Canonical page: https://www.alloovium.com/en/docs/guides/commercial-control/security-register --- ## Compliance Source: https://www.alloovium.com/en/docs/guides/compliance COMPLIANCE # Compliance in Alloovium Compliance is not a separate product. It is a risk lens that reads the documents already in your project — spotting expiring credentials, unsigned records and evidence gaps, then surfacing them alongside your day's work. ## Overview Construction teams already hold the evidence an auditor wants: safety plans, ITPs, permits, inductions, registrations and certificates. The problem is rarely that the evidence is missing — it is that a plan quietly expired, a record was never signed, or a commitment made in a specification has no proof behind it. Compliance in Alloovium is the lens that reads for exactly those signals. By default the lens is deliberately light. It watches documents as they arrive, detects a small set of deterministic signals — expiry dates, incomplete or unsigned records, blank templates — and raises a plain, trackable finding for each. No configuration and no separate console are required to get value on day one. Video · \~2 min Compliance as a risk lens From signal to closed loop: a document arrives with an expiry date, the lens raises a finding in Requires attention, we open the citation on the exact page, then assign it, add resolution notes and resolve it out of the queue. ## A lens, not a separate app Everything in Alloovium sits inside one product built around your project documents. Compliance is one way of looking at that same corpus — a risk lens — rather than a place you go to do separate work. The findings it raises become items you act on in the flow you already use, and they link straight back to the source document and the exact page. Two levels of depth are available. The light lens described above is always on. A heavier conformance engine — which extracts the commitments a project makes, searches the corpus for proof, and judges whether each is satisfied — is available when the compliance engine is enabled for your workspace. Light by default, deep when enabled The credential-expiry and unsigned-record checks run for everyone. The full promise-versus-proof engine, war-game scenarios and audit-pack assembly are available when the compliance engine is enabled for your workspace. Where a capability below is engine-gated, it says so. ## The credential-expiry backbone The backbone of the light lens is expiry. When a document arrives, Alloovium reads it for expiry language — phrases such as `expires`, `valid until` or a recertification date — and parses the date next to it. If that date has passed, or is close, the lens raises a finding attributed to whoever uploaded the document. Registered professionals are tracked the same way: a registration expiry on an engineer or other licensed role becomes a signal the lens can watch. The point is not to build a separate register you have to maintain, but to notice the dates that are already written into the documents you hold. | Signal | What it catches | How it surfaces | | --------------------- | ------------------------------------------------------------------------ | -------------------------------------------------------- | | Expiry | A plan, permit, certificate or registration past or near its expiry date | A finding attributed to the uploader, linked to the page | | Unsigned / incomplete | A record left unsigned or with required fields blank | A finding flagging the incomplete document | | Blank template | A template placed but never filled in | A finding prompting completion | Deterministic, so it is quiet Light signals are detected without any language model. They surface only when a real date or an unsigned marker is present, which keeps the lens from crying wolf. Dismiss a finding enough times and it stops resurfacing. ## Findings and gaps A finding is the unit of work the lens produces. Whether it comes from a light signal or from the deeper engine, it carries the same shape: a short title, a severity, a status, and — critically — a citation back to the document and page it came from. You can assign it, add resolution notes, resolve it, or dismiss it. ### What a finding records | Field | Meaning | | -------- | --------------------------------------------------------------- | | Title | A short, human-editable description of the gap | | Severity | Low, medium, high or critical | | Verdict | Verified, partial, unverified or contradicted (from the engine) | | Citation | The source document, page and excerpt the finding rests on | | Status | Open, resolved or dismissed, with who is assigned | When the compliance engine is enabled, findings are computed as gaps between what a project promises and what the corpus proves. That deeper analysis — conformance gaps, contradictions and the Attention surface that routes each gap to the person who can close it — is covered on the [conformance and attention](/en/docs/guides/compliance/conformance) page. ## Standards it understands When the compliance engine is enabled, findings can be organised against management-system standards. The built-in rubrics cover the three ISO systems most construction and engineering teams carry. | Standard | System | | --------- | ------------------------ | | ISO 9001 | Quality management | | ISO 14001 | Environmental management | | ISO 45001 | Health and safety (WHS) | Each standard is broken into hierarchical clauses with evidence hints, so a gap can point at the specific clause it relates to. The light lens does not require any of this — it works on raw document signals alone. ## Working the queue When the compliance engine is enabled, **Compliance** in the sidebar is the place to work: it lands on **Requires attention** — the queue of every open compliance item, with nothing to configure or filter. Each project also carries its own **Findings** board (open the project, then Findings) for the gaps raised on that job. Light-lens signals — an expiring credential, an unsigned record — also flow into the feed of things needing you on [Today](/en/docs/guides/today), so they meet you where you already work. 1. 1 A signal is detected A document arrives with an expiry date or an unsigned record, and the lens raises a finding. 2. 2 It lands in Requires attention The finding is attributed to a person and joins the queue, cited to the document and page it rests on. 3. 3 You close the loop Provide the evidence, then assign, resolve or dismiss. Resolved items leave the queue; dismissed ones stop resurfacing. ## Where to go next Two companion pages go deeper into the engine-gated capabilities: * [Conformance and attention](/en/docs/guides/compliance/conformance) — how project conformance gaps are computed, routed and war-gamed. * [Audit pack and evidence](/en/docs/guides/compliance/audit-pack) — assembling a point-in-time evidence dossier for an auditor. --- Canonical page: https://www.alloovium.com/en/docs/guides/compliance --- ## Audit pack & evidence Source: https://www.alloovium.com/en/docs/guides/compliance/audit-pack COMPLIANCE # Audit pack & evidence An audit pack is a curated, point-in-time bundle of the evidence behind your compliance findings — the documents that map to each clause, frozen at a moment in time and shareable read-only with an auditor. Available when the compliance engine is enabled Audit-pack assembly builds on the conformance engine and is available when the compliance engine is enabled for your workspace. If you are new to compliance in Alloovium, start with the [compliance overview](/en/docs/guides/compliance). ## Overview When an audit arrives, the work is rarely finding the evidence — it is choosing which records to hand over, making sure they map cleanly to the clauses being assessed, and being able to prove nothing was changed after the fact. An audit pack does that assembly for you. A pack is a container you build inside a project. You decide which documents cross to the auditor, map them to the clauses they satisfy, then finalize the pack so its contents are frozen and can be shared read-only. It turns the evidence already sitting in your corpus into a defensible dossier. Video · \~2 min Assemble an audit pack Building a pack from scratch inside a project: create the draft, add evidence documents and map each to its clause, gate what the auditor sees with the inclusion toggles, finalize to snapshot everything, then open the read-only share link the auditor gets. ## What a pack collects A pack is a small set of parts, each chosen deliberately so the auditor sees exactly what you intend and nothing more. | Part | What it is | | ---------------- | -------------------------------------------------------------------------------- | | Evidence items | The documents you have chosen to include, each mapped to the clause it satisfies | | Inclusion gate | A per-item toggle that decides whether an auditor sees a given document | | Snapshots | A frozen version of each included document, captured when the pack is finalized | | Ruleset snapshot | A record of the rule documents in force plus a hash, so changes are detectable | | Share token | A read-only link the auditor uses to view the finalized pack | You choose what crosses over Only items with inclusion switched on are visible to the auditor. Building the pack and gating it are separate steps, so you can stage everything and then decide what to hand over. ## Assembling a pack Assembly happens inside the project whose evidence you are packaging. The typical flow is short. 1. 1 Create a pack Start a new pack in the project. It begins as a draft you can edit freely. 2. 2 Add evidence and map to clauses Pull in the documents that support your findings and map each to the clause it satisfies. 3. 3 Gate what the auditor sees Toggle inclusion on each item so the pack shows only what you intend to share. 4. 4 Finalize Lock the pack. Its included documents are snapshotted and it becomes shareable. Because a pack draws on the same findings the conformance engine produces, the mapping between evidence and clauses usually reflects work you have already done closing gaps — see [conformance and attention](/en/docs/guides/compliance/conformance). ## Freezing and finalizing Finalizing is the step that makes a pack defensible. A draft can change; a finalized pack cannot. ### Document snapshots When you finalize, each included document is captured at its current version. If the live document changes afterwards, the pack still shows what the auditor was given, which prevents any question about tampering after the review began. ### Ruleset hash Alongside the evidence, a pack records the rule documents in force at finalize time together with a `SHA-256` hash. An auditor can use that hash to confirm the rules a project was assessed against were not quietly changed after the fact. Finalize is a point in time A finalized pack reflects the corpus as it stood at that moment. If evidence changes later, assemble a fresh pack rather than editing the finalized one. ## Sharing with an auditor A finalized pack is shared through a read-only share token. The auditor opens the link and sees only the included evidence, mapped to clauses, with the snapshots and ruleset record intact. They cannot alter anything, and nothing outside the pack is exposed. This keeps the audit on your terms: you controlled what went in, the contents are frozen, and the pack proves its own integrity. ## Related pages * [Compliance overview](/en/docs/guides/compliance) — the risk lens, credential expiry and findings. * [Conformance and attention](/en/docs/guides/compliance/conformance) — how the gaps behind your evidence are computed and routed. * [Today](/en/docs/guides/today) — where compliance items surface in your daily flow. --- Canonical page: https://www.alloovium.com/en/docs/guides/compliance/audit-pack --- ## Conformance & attention Source: https://www.alloovium.com/en/docs/guides/compliance/conformance COMPLIANCE # Conformance & attention The conformance engine reads the commitments a project makes, searches the corpus for proof, and judges whether each is satisfied. Where proof is thin, it raises a gap and routes it to the person who can close it. Available when the compliance engine is enabled The capabilities on this page — commitment extraction, evidence verification, conformance gaps, the Attention surface and war-game scenarios — are available when the compliance engine is enabled for your workspace. The always-on light lens is described on the [compliance overview](/en/docs/guides/compliance). ## Overview Conformance is the deeper half of the compliance lens. Where the light lens reads for simple signals like expiry dates, the conformance engine asks a harder question: for every commitment this project has made, is there evidence in the documents that it was actually met? A commitment with no proof, weak proof, or contradicting proof becomes a gap. The engine runs as a sweep over the project corpus. It extracts commitments, searches for the evidence that would satisfy each, judges the match, and publishes the results as findings. Every finding keeps the citation it was judged on, so you can always trace a gap back to the exact page. ## Promise versus proof The engine models two things. A **commitment** is something the project promised — a control in a safety plan, an inspection required by an ITP, a clause in a specification. A **verdict** is the engine’s judgment of whether the corpus proves that commitment was met, together with the citations it relied on. 1. 1 Extract commitments Commitments are pulled from intention documents — plans, specifications and ITPs — and from document signals such as required inspections. 2. 2 Retrieve evidence For each commitment, Alloovium searches the project corpus for the records that would satisfy it, filtered to what you have permission to see. 3. 3 Judge the match The engine decides whether the evidence verifies the commitment, and records the reasoning and citations behind that verdict. 4. 4 Publish findings Anything short of verified is published as a gap and routed to a person through the Attention surface. ## What counts as a gap A gap is any verdict that falls short of fully verified. The engine grades each commitment so you can tell the difference between missing evidence and a genuine contradiction. | Verdict | What it means | Treated as a gap? | | ------------ | ----------------------------------------- | ------------------- | | Verified | The corpus proves the commitment was met | No | | Partial | Some evidence exists but it is incomplete | Yes | | Unverified | No evidence was found for the commitment | Yes | | Contradicted | Evidence conflicts with the commitment | Yes (high priority) | Findings carry a severity from low to critical, and roll up into a per-standard health picture — how many commitments are verified, partial, missing or contradicted. That summary is what tells you whether a project is broadly audit-ready or has critical gaps to close first. A contradiction is louder than a gap Missing evidence often just means a record has not been filed yet. A contradiction — evidence that conflicts with what was promised — points at a real problem and is surfaced with higher priority. ## The Attention surface Finding a gap is only useful if it reaches the person who can close it. The Attention surface is how gaps become work. When a finding is published, it is attributed to the person tied to the source document and raised as an item they can act on, with a notification so it is not lost. ### What keeps it quiet To avoid a wall of noise, a finding surfaces only when it earns its place. Deterministic light signals surface directly. Engine verdicts must clear a relevance check and carry a verbatim excerpt — a citation with a page — before they appear. And a finding dismissed enough times is suppressed so it does not keep coming back. Attention items appear in the same feed as the rest of your day on [Today](/en/docs/guides/today), tagged as compliance, so you triage them alongside schedule and cost work rather than in a separate silo. ## War-game scenarios A war-game is a forward-looking check. Instead of grading commitments already made, it asks what would happen if a specific situation arose — and whether the project’s documents show the team is ready for it. You enter a scenario in plain language, for example `bushfire on site` or `a subcontractor’s insurance lapses mid-works`. Alloovium pulls the most relevant document excerpts and asks the model where the response would fall short, returning a short summary and a list of gaps, each with a severity and a recommended action. 1. 1 Describe the scenario Write what could go wrong in plain language. 2. 2 Alloovium reads the corpus It retrieves the most relevant excerpts from the project documents to ground the analysis. 3. 3 Review the gaps You get a summary plus specific gaps. Accept the ones worth tracking and route them to the risk register; dismiss the rest. Grounded, not guessed War-game gaps are drawn from what your documents actually say, not generic advice. Treat them as prompts for a human review — accept the ones that matter and dismiss the rest. ## Working a finding to close Every finding — light signal, conformance gap or accepted war-game gap — moves through the same simple lifecycle so nothing is left half-open. | State | Meaning | | --------- | ----------------------------------------------------- | | Open | Raised and waiting to be worked | | Assigned | Given to a specific person to close | | Resolved | Closed with resolution notes; leaves the feed | | Dismissed | Judged not relevant; repeated dismissal suppresses it | Once your findings are in good shape, you can package the supporting evidence for an auditor. That is covered on the [audit pack and evidence](/en/docs/guides/compliance/audit-pack) page. --- Canonical page: https://www.alloovium.com/en/docs/guides/compliance/conformance --- ## Obligations register & chase campaigns Source: https://www.alloovium.com/en/docs/guides/compliance/obligations-register COMPLIANCE # Obligations register & chase campaigns The contract is a list of promises with deadlines. The obligations register extracts that list — one row per obligation, cited back to its clause — and keeps it in front of you with due dates and status. Chase campaigns then do the follow-up: one launch chases every subcontractor with a missing or expiring document, and each response closes its own chase. ## Overview Failure to understand or comply with contractual obligations has been the number-one cause of construction disputes for three years running (Arcadis Global Construction Disputes Report 2025). The obligations register exists so the contract’s promises are tracked as records rather than remembered: what must be done, by which party, triggered by what, due when, and where in the contract it says so. You reach the register from the **Registers** band in the project sidebar, next to the [claims register](/en/docs/guides/commercial-control/claims-register) and the project’s [other registers](/en/docs/guides/projects/registers). You do not have to build it: contracts are read as they arrive, and every row keeps the citation it came from. The obligations register for a project — rows of obligations with clause reference, party, type, due date tinted red for overdue and amber for due soon, severity and status columns, and a source chip on each row linking back to the contract clause. ## Extraction is automatic — and re-runnable The register is not something you have to go and generate. The picker states it in its own words: _“Contracts are read automatically as they arrive — every deliverable, submittal and obligation becomes a register row, cited back to its clause. Use this to re-run a contract, or search to point the AI at a document it did not pick up.”_ Upload the contract and the rows appear behind it. So the extraction picker exists for the two cases automatic reading does not cover: * **Re-running a contract** — the contract has been revised, or an earlier read was thin and you want the rows rebuilt. * **Pointing the extractor at a document it did not read as a contract** — a deed, a variation of contract, a special-conditions annexure or a scope document the classifier did not treat as contractual. ### Using the picker By default the picker lists only the project’s contracts, under a **Contracts in this project** heading — the short list is the common case. Searching widens the picker to every document in the project, which is how you reach a document that was never classified as a contract. A run is capped at **10 documents** (`Up to 10 documents per run.`), each read as its own extraction job, and the register shows per-document progress while the runs are underway. Documents that have already been read are skipped unless you tick **Re-extract even if already extracted** — that is the option to reach for after a contract revision. The picker’s empty states tell you which situation you are in: _“Nothing in this project reads as a contract yet — search above to pick any document.”_ means nothing was classified as contractual, while _“No documents in this project yet — upload the contract first.”_ means there is nothing to read at all. Grounded in the clause, not a summary of it Every row keeps its source: the document, the clause reference, and the exact passage the obligation was extracted from. The source chip on a row opens the contract in the preview pane at that clause, so checking an obligation against the words that created it is one click. ## What each row carries | Field | Meaning | | ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Clause | The contract clause reference, for example cl 12.3(a). | | Obligation | What must be done, in plain language. | | Party | Who owes the obligation — Principal, Contractor, or both. | | Trigger | The event or date that starts the obligation running. | | Due | The due date where one can be resolved, or the contractual expression where it cannot — for example "30 days after practical completion". Days-to-due are calendar days. | | Recurrence | One-off, recurring (for example monthly), or triggered by an event. | | Type | Reporting, payment, notice, insurance, indemnity, milestone, or other. | | Severity | Critical, high, medium or low — how much it matters if this one is missed. | | Status | Active, fulfilled, overdue, waived or superseded. | | Source | The document and clause the row was extracted from. | Due dates are tinted so the register reads at a glance: overdue rows in red, rows due within fourteen days in amber. Rows are ordered soonest-due first. Where a trigger was extracted, it is shown beneath the obligation text, so the row reads as the promise and the thing that starts it running. Severity is deliberately quiet Only **Critical** and **High** carry a coloured chip; Medium and Low render as plain text. Colour marks the exception, so on a register of two hundred rows the ones that will hurt you are the loudest thing on the page. ### Every row drills into its clause The source chip on a row opens the contract in a preview panel, jumped to the passage that produced the obligation. The jump is matched on the verbatim snippet the extraction kept, falling back to the clause reference when the snippet cannot be located — so a row is never an assertion you have to take on trust. ## Corrections survive re-extraction Extraction gets most rows right and some rows nearly right. When you correct a field on a row — the due date, the party, the trigger, the wording of the obligation — that field is recorded as **confirmed**, and the row carries a small green **Confirmed** badge. Hover it and you see the provenance of the fix: what the machine originally read, what you confirmed instead, and the date you did it. A confirmed field is not overwritten when the contract is re-extracted. Re-running a revised contract refreshes the rows around your corrections rather than through them. No, the AI will not clobber your fix This is the first question a commercial manager asks of any extracted register, and the answer is the reason it is safe to correct rows as you go. Human judgement is treated as the higher-confidence source: once a field is confirmed, extraction defers to it and the original reading is kept alongside as history, not discarded. ## Working the register The register is a working list, not a one-off report. As obligations are met, mark them fulfilled — the record keeps who fulfilled it and when. An obligation that no longer applies can be waived, and one replaced by a revised contract is marked superseded, so the register stays honest without deleting history. ### Status actions on a row From the row itself you can **Mark fulfilled**, **Mark waived**, or **Reopen** one that was closed in error or has come back around. Fulfilled and waived rows render struck through and greyed rather than disappearing, so the history stays intact and an open register reads as the work that is left. ### Back-to-back gaps Above the register sits a panel showing time-bar exposure across the contract chain: where the window you have under the head contract is shorter than the window you have given a subcontractor. It reads the same obligations you are looking at and shows where the windows do not line up — the places where a claim can reach you after your own right to pass it on has closed. The assistant exposes the same view, so you can ask for it in chat rather than opening the page. ### Ask about it in chat The register is readable by the assistant, read-only and filtered to what you are permitted to see. So `what’s due on Northgate in the next fortnight?` — or `what notices do we owe this month` — is answered from these rows, not re-derived from the contract on the spot. Same rows you see on the page, same citations back to the clause. ## Compliance-chase campaigns Knowing what is owed is half the job; collecting it is the other half. A compliance-chase campaign follows up an entire list of subcontractors for a document you need from each of them — certificates of currency are the classic case — in one launch, instead of one email thread per sub. ### Who gets chased When you create a campaign, Alloovium builds the candidate list from the counterparties already known on the project and works out each one’s compliance state from the documents they have previously submitted through document requests: **current** when their best document is in date, **expiring** when it lapses within the window (30 days by default), and **missing** when no matching document exists — an expired document counts as missing. You review that preview and choose who to include before anything is sent. ### What a launch does You give the campaign a name, the document kinds you are asking for — certificate of currency, insurance certificate, SWMS, ITP, permit, licence, induction record or JSA — and an escalation ladder. Launching it creates, for each target, a document request (the ask) and an escalation case (the chase) that walks the ladder on its own clock. One response never stops another chase Every target runs as its own independent chase. When a subcontractor submits the requested document and it is accepted, their chase stops — everyone else’s keeps walking the ladder. That is what makes a 200-target campaign practical: you launch once and only the stragglers keep consuming attention. Candidate states are computed from requested documents A counterparty’s compliance state is derived from documents received through document requests on the project. A sub whose current certificate sits in your library but was never submitted through a request will read as missing — check the preview before launch and exclude anyone you know to be current. ## Following a campaign The campaign page is the rollup: how many targets are resolved, how many are still being chased, and where each one sits. | Target status | What it means | | ------------- | ------------------------------------------------------ | | Pending | Queued — the campaign has not sent to this target yet. | | Sent | The request has gone out. | | Opened | The target has opened the request. | | Responded | The target has submitted a document. | | Resolved | The submission was accepted — this chase is closed. | | Escalated | The chase has moved up the ladder without a response. | | Failed | The chase could not proceed for this target. | The campaign itself moves from `draft` to `active` on launch, and can be paused, resumed, or completed. A target can also be resolved manually — for example when a certificate arrives by another channel. ## Related pages [Compliance overview](/en/docs/guides/compliance) — the always-on lens that reads expiry dates and other light signals. [Conformance & attention](/en/docs/guides/compliance/conformance)— proving the project’s commitments are actually met. [Audit pack](/en/docs/guides/compliance/audit-pack) — packaging the evidence for an auditor. [Working with Trades](/en/docs/guides/portal) — how document requests reach subcontractors. [Project registers](/en/docs/guides/projects/registers) — the Registers band the obligations register sits in, and the other registers beside it. --- Canonical page: https://www.alloovium.com/en/docs/guides/compliance/obligations-register --- ## Document Generation Source: https://www.alloovium.com/en/docs/guides/document-generation Routines # Document Generation Document Generation drafts a new document from scratch when you don't have a template. Describe what you need — a variation notice for a scope change, an RFI about a drawing clash, a site instruction for a setting-out adjustment — and Alloovium produces a structured draft using your project documents as context. ## Overview Document Generation drafts a new document from scratch when you don't have a template. Describe what you need — a variation notice for a scope change, an RFI about a drawing clash, a site instruction for a setting-out adjustment — and Alloovium produces a structured draft using your project documents as context. When to use which If you have a blank DOCX template with placeholders, use **Template Filling**. If you need a brand-new document and don't have a template, use **Document Generation**. ## Document Generation 1. 1 Open Document Generation Open Routines in the left sidebar, then switch to the Actions tab and click "Document Generation". 2. 2 Describe what you need Tell Alloovium the document type and the situation — for example, "Draft an RFI about the clash between the structural drawing S-201 Rev C and the hydraulic drawing H-104 Rev B on Level 2." 3. 3 Pick the project context Select the project Alloovium should pull information from. 4. 4 Review and export The draft appears with citations. Edit it, regenerate sections you want rewritten, or export as DOCX. Video · \~3 min Draft a document with receipts Describe the document you need, watch the draft stream in grounded in project files, open an inline citation to check its source passage, then export — or sign it in place with the built-in e-signature. ## Common document types * **Variation notice** — notification of a scope or cost change. * **RFI** — a request for information to a designer or consultant. * **Monthly report** — a progress report covering programme, commercial, risk, and safety. * **ITP** — an inspection and test plan generated from the relevant specification. * **Site instruction** — an instruction to a subcontractor on site. ## History Every filled template and generated document is saved under the **History** tab on the Routines page. From there you can re-open, re-export, or regenerate any past document. Running the same template again on an updated project pulls the latest information without losing the edits you made to earlier versions. Citations stay with the document When you re-open a past document from History, the original citations remain intact. You can click any cited value to jump straight to the source document and page. --- Canonical page: https://www.alloovium.com/en/docs/guides/document-generation --- ## Email Queries Source: https://www.alloovium.com/en/docs/guides/email-queries DAILY WORKFLOW # Email Queries Project email is where commitments hide — approvals given in a reply, dates agreed in a thread, an attachment nobody filed. Email queries let you ask across your synced inbox the same way you ask across your documents, and get a cited answer back. ## Overview Once your mailbox is connected through the [Outlook](/en/docs/guides/integrations/outlook) or [Gmail](/en/docs/guides/integrations/gmail) integration, Alloovium reads project-relevant threads and their attachments into the same searchable corpus as your documents. Ask “did the superintendent ever approve the revised propping sequence?” and the answer comes back with the exact email cited — sender, date, and the line that says so. One question, every source You do not have to decide up front whether the answer lives in a document or an email. The assistant searches both, and each claim in the answer is cited to wherever it was found. Video · \~2 min Asking your inbox a question A real question — has the supplier confirmed a delivery date? — asked on the Email Queries page, answered with cited lines from the thread, then traced back to the original email via its citation. ## Asking a question From the Email Queries page, or anywhere you talk to the assistant. The Email Queries page gives you a focused surface for mail-heavy questions — chasing a commitment through a long thread, reconstructing who said what and when, or finding the attachment that never made it to the project files. The same questions also work in the ordinary [AI Assistant](/en/docs/guides/ai-assistant), where email simply participates as one more source. | Ask something like | What you get back | | ------------------------------------------------------------------ | ------------------------------------------------------------- | | “What did the client say about the variation in their last email?” | The relevant reply, quoted and cited to the thread. | | “Has the switchboard supplier confirmed a delivery date?” | The latest commitment found across the thread, with its date. | | “Summarise the back-and-forth on the façade panels this month.” | A short timeline of the thread with each turn cited. | ## What it reads Threads, attachments, and the project context around them. Email queries draw on the mail your integration has synced: message bodies, thread structure (who replied to what, and when), and attachments — which are processed like any other document, so a PDF sent by email is as searchable and citable as one uploaded by hand. Where a thread clearly belongs to a project, it is linked there, and shows up in the project’s Emails tab alongside its files. ## Action items from email Commitments in your inbox surface on Today, not in your memory. Beyond answering questions, Alloovium reads incoming project mail for things that need you: a request waiting on your reply, a date somebody committed you to, a document you were asked to provide. These surface as action items on your [Today](/en/docs/guides/today) page, each linked back to the email it came from — so the inbox stops being a place where obligations quietly age. ## Privacy Your mailbox is yours. A connected mailbox is private to the person who connected it. Your teammates’ questions never search your mail, and yours never search theirs — email joins the shared project corpus only through the threads linked to a project. See our [data handling commitments](/en/docs/legal/data-handling) for how synced content is stored and protected. --- Canonical page: https://www.alloovium.com/en/docs/guides/email-queries --- ## Exchange — Overview Source: https://www.alloovium.com/en/docs/guides/exchange Exchange # Version control for construction documents Alloovium Exchange treats a contract, specification, or management plan the way a version-control system treats source code: every revision is captured, every party's edits are attributable, and every read and acknowledgement lands in a write-once audit trail. Video · \~3 min Send, sign, and keep the receipts One document's full round trip through Exchange: pick it from project files with the slide-over preview, send it across companies, watch the recipient open the share link, accept and sign in place, then land on the hash-chained audit trail showing every event. ## What it is Exchange is the document layer of the Alloovium suite. It is built around the revision — not the folder, not the email attachment, not the SharePoint timestamp. It runs on DOCX as a first-class format, with PDFs supported in the viewer for read-and-acknowledge flows. Who it is for Head contractors running subcontractor document packs on Australian infrastructure projects. The three-layer audit trail (person, organisation, credential) is what separates Exchange from a filename-suffix workflow. ## Capabilities Four product surfaces, all anchored to the same write-once audit log: * **Clause-level diffs** — see what changed between two DOCX revisions, clause by clause — not glyph by glyph. * **Trade markups, merged** — every sub's redlines roll back onto the head contractor's master with conflicts flagged at the clause they touch. ## Receipts The point of sending through Exchange is that the send leaves a receipt. Every step in a shared document’s life is written to an append-only trail — recorded once, never edited, never deleted — so “we never got that” stops being an argument and becomes a lookup. * **Sent and opened**— each share records when it went out and when the recipient first opened it, giving you a timestamped record of notification per counterparty. These facts also feed the project’s Distribution register. * **Every state transition** — shared, submitted, sent back with a note, resubmitted, approved: each move is logged with who made it, when, and the note that travelled with it, including the review outcome stamped on the transition (approved, approved-as-noted, revise-and-resubmit, or rejected). * **Signatures bound to the bytes** — when a document is signed at approval, the trail records who signed and a `SHA-256` hash of the exact file as published. The signature row survives as evidence even if the document itself is later deleted, and the hash proves the signed file was not swapped afterwards. * **Honest tracking only** — open receipts are shown only for channels that actually support them. A document that went out by plain email is never dressed up with a read receipt it does not have. Timestamps decide real money In _Sharvain Facades v Roberts Co_\[2025\] NSWSC 606 (upheld \[2025\] NSWCA 161), a subcontractor uploaded a $3,207,999.03 payment claim at 7:18pm on a Friday; the head contractor’s after-5pm deeming clause was held void and its payment schedule was one business day late — judgment for $3,278,043.27 including interest. When service and response windows are in play, a timestamped trail of what was sent and when is the whole ballgame. For the who-holds-which-revision view built on these receipts — including winding the record back to any past date — see [the Distribution register](/en/docs/guides/projects/registers). ## Where to go next * →[Clause-level diffs](/en/docs/guides/exchange/clause-diffs) — the DOCX diff engine and how comment threads survive across revisions. * →[Trade markups, merged](/en/docs/guides/exchange/markups-merge) — merging redlines from every sub back onto the master. * →[Working with Trades](/en/docs/guides/portal) — run the submission and transmittal loop with your subcontractors. --- Canonical page: https://www.alloovium.com/en/docs/guides/exchange --- ## Clause-level diffs Source: https://www.alloovium.com/en/docs/guides/exchange/clause-diffs Exchange # Clause-level diffs DOCX is the format Exchange is built around. When you issue a new revision, the document is parsed clause by clause and the diff is computed against the previous revision in the same branch. ## How it works The viewer surfaces three things on every revision: * **Side-by-side and inline views** — toggle in the viewer toolbar. Reviewers default to side-by-side; site teams default to inline. * **Per-clause status** — added, removed, modified, or unchanged — colour-coded in the gutter. * **Jump-to-change** — keyboard navigation through every change in the document, in order. ## Why DOCX (and not PDF-only) PDF diffs operate on glyph positions. They tell you something changed on page 14 — not that the indemnity cap moved from $5m to $25m. Exchange parses DOCX into the structural tree (paragraphs, runs, tables) so diffs land on the actual contract language. This is the same reason auditors prefer red-line DOCX over PDF mark-ups for contract negotiation. PDFs in Exchange PDFs can be issued and read against in the viewer with comment threads, but the clause diff and markup merge only run on DOCX. Convert from PDF if you need the structural diff. ## Comment anchoring Comment threads anchor to a position inside the document so they survive editing in subsequent revisions. There are three anchor kinds: * **paragraph** — the comment is attached to a structural paragraph; survives unless the paragraph is deleted. * **text\_range** — the comment quotes a literal text span and stores the page number and bounding box; useful for clause-level disputes. * **region** — the comment is pinned to a screen region, normalised 0..1 across the page. Useful for tables and diagrams that don't map cleanly to a paragraph. --- Canonical page: https://www.alloovium.com/en/docs/guides/exchange/clause-diffs --- ## Trade markups, merged Source: https://www.alloovium.com/en/docs/guides/exchange/markups-merge Exchange # Trade markups, merged The inverse of one-revision-every-party. Where propagation fans a single revision out to many subs, markup-merge collects clause-level edits from every sub and rolls them back onto the master. ## How it works Each sub redlines their own copy in their own branch. When the head contractor opens a merge, Exchange overlays each branch's clause-level edits onto the master and walks the head contractor through every change in order. ## Conflict resolution Where two subs touched the same clause, the merge UI surfaces a conflict inline. The head contractor sees both edits side by side along with the original clause and resolves once — accepting one, accepting both, or writing a new version. Resolving in context Resolve conflicts inside the viewer rather than rerouting back to the subs. The audit log captures the resolution decision and the principal can see exactly why a clause reads the way it does. ## Why this matters Without a merge primitive, sub edits arrive as commented PDFs or red-line DOCX attachments by email. The head contractor retypes them into the master, mistakes accumulate, and the audit trail is reconstructed from forwarded email headers. Exchange replaces all of that with one merge action. --- Canonical page: https://www.alloovium.com/en/docs/guides/exchange/markups-merge --- ## Getting Started Source: https://www.alloovium.com/en/docs/guides/getting-started Getting started # Getting started with Alloovium Alloovium turns your project documents, emails, and site records into a living record you can question, draft from, and prove things with. This page takes you from a fresh account to your first cited answer. Alloovium in ten minutes · 13 min ## What Alloovium is Alloovium is an AI platform for construction and AEC teams. You bring the project record — contracts, specifications, drawings, emails, minutes, site diaries — and Alloovium reads all of it, so you can ask questions in plain language and get answers with citations back to the exact clause, drawing, or email they came from. Everything is organised into **Projects**: one workspace per contract, tender, or job site, each with its own documents, people, and history. Above projects sits your **Organization** — the company-wide library of templates, standards, contacts, and lessons that every project can draw on. Who is Alloovium for? Project managers, contract administrators, engineers, and estimators — anyone who spends their day inside dense technical and legal documents and needs answers they can defend. ## Your first hour Five steps from a fresh account to a working project. Most teams get their first cited answer within minutes of uploading. 1. 1 Sign in — or accept your invite Sign up at alloovium.com with your work email, or follow the invite link a teammate sent you. Invited users land directly in the projects they were given access to — no setup required. 2. 2 Create your first project Click "New Project", give it a name like "Northgate Tower — Main Works", and add an optional description. Projects are private until you invite people. 3. 3 Upload your documents Drag PDFs, Word, Excel, and PowerPoint files — or whole folders — into the project. Processing (OCR, classification, indexing) runs automatically, and Auto-sort proposes a folder for every file, for your review, before anything moves. 4. 4 Ask your first question Type into the ask bar on the project's Files tab — "What's the defects liability period?" — and get an answer with numbered citations back to the source. 5. 5 Open Today Today is your daily cockpit: the calendar spine, the "Needs you" queue, and the activity feed. It fills up as your projects come alive. ### Ask your first question Every answer carries numbered citations. Click one and the source document opens with the supporting passage highlighted — so you never have to take the AI's word for it. This is the core promise of Alloovium: **answers with receipts**. Your first question, start to finish · 33 sec ## Finding your way around The left sidebar is the whole map. Seven surfaces cover everything. | Surface | What it does | | ------------ | ---------------------------------------------------------------------------------------- | | Today | Your daily cockpit — calendar, the Needs-you queue, action items, and the activity feed. | | Assistant | Company-wide chat. Ask across every project you can access, with citations. | | Projects | One workspace per job — files, meetings, people, claims, and the project timeline. | | Organization | The company layer — template library, playbook, lessons, contacts, and people. | | Conformance | Compliance signals — expiring credentials, findings, and audit packs. | | Exchange | Send, receive, and sign documents across companies, with a full audit trail. | | Integrations | Connect Outlook, SharePoint, Procore, Autodesk, and 15+ other systems. | Jump anywhere with ⌘K The command palette finds documents, projects, and people from anywhere in the app. Press ⌘K (Ctrl+K on Windows) and start typing. ## Inviting your team Access in Alloovium is granted through **user groups**, not per-person role picking. When you invite someone you choose which groups they join; each group carries a permission matrix that decides what its members can see and do. A **Read-only** group is preselected for new invitees, so the safe default is one click. Invite people to the whole organisation from **Organization → People**, or to a single project from the project's **People** tab. You can create a new group inline during the invite — including copying an existing group's permissions as a starting point. The full access model — groups, the permission matrix, and project membership — is covered in [Teams & Groups](/en/docs/guides/collaboration). ## Next steps You have a working project. From here, go deeper wherever your day takes you: * →[Key Features](/en/docs/guides/getting-started/features) — the complete capability tour, from drafting to routines to Exchange. * →[Today](/en/docs/guides/today) — make the daily cockpit yours. * →[Documents & Projects](/en/docs/guides/projects) — uploading, the AI Organizer, and drawing intelligence. * →[AI Assistant](/en/docs/guides/ai-assistant) — querying, citations, voice input, and sharing. * →[Routines](/en/docs/guides/routines) — put the busywork on a schedule with approval gates. * →[Use Cases](/en/docs/use-cases) — ten real workflows for your role, from project managers to executives, each with a walkthrough video. --- Canonical page: https://www.alloovium.com/en/docs/guides/getting-started --- ## Key Features Source: https://www.alloovium.com/en/docs/guides/getting-started/features Getting started # Key features Everything Alloovium can do, in one tour. Each capability links to its full guide. ## Ask and verify The core promise: plain-language answers with citations back to the exact source. ### AI Assistant Ask about a single project or across your whole company. Every answer carries numbered citations — click one and the source opens with the passage highlighted. Attach files to a question, dictate by voice, share conversations with teammates, and let long answers keep generating in the background while you work. [AI Assistant guide →](/en/docs/guides/ai-assistant) ### Command palette Press ⌘K anywhere to jump to a document, project, or person. It's the fastest way around the app once your record grows. ## The project record Upload once. Alloovium reads, files, and indexes everything — and keeps the history. ### Projects and files One workspace per job. Drag in PDFs, Word, Excel, PowerPoint, and whole folders; OCR, classification, and indexing run automatically, and **Auto-sort** proposes the right folder for every file — reviewed and editable before anything moves. [Documents & Projects guide →](/en/docs/guides/projects) ### Drawing intelligence Drawings get their title blocks read automatically — sheet number, revision, discipline, date — so "latest revision of A-101" is an answerable question. [Drawing Intelligence guide →](/en/docs/guides/projects/drawing-intelligence) ### Project timeline Every event — uploads, emails, claims, variations, AI actions — lands on the project timeline. Assistant answers cite timeline records with \[T#\] chips that deep-link straight to the event. [Timeline & Activity guide →](/en/docs/guides/projects/timeline) ## Drafting with receipts Documents drafted from your record, with every claim traceable to its source. ### Document generation Describe what you need — a delay notice, an RFI response, a WHS plan — and Alloovium drafts it grounded in your project files, auto-matching your library templates. Inline citations, export to Word, and native e-signature to sign in place. [Document Generation guide →](/en/docs/guides/document-generation) ### Template filling Drop in a blank Word template and Alloovium fills every field from your documents. The output is tracked changes — accept or reject each fill, each one citing its source. [Template Filling guide →](/en/docs/guides/template-filling) ### Reports Monthly reports, management plans, tender reviews, and infrastructure reports — assembled from diaries, registers, and minutes, in your own template, reviewed before anything ships. [Monthly Report guide →](/en/docs/guides/monthly-report) ## Routines Routines do the recurring work — on a schedule, on conditions, with approval gates. ### Routines Build a chain of blocks: a trigger, then actions — check documents, ask the team, draft a document, send it. Routines can pause for your approval, wait for an external reply and chase it if none comes, and run on schedules or conditions. Repeated asks in chat offer an **Automate this** chip that builds the routine for you. Video · \~2 min A routine that waits and chases A routine sends a request, pauses on 'await reply', chases the recipient automatically after the deadline, and resumes when the reply lands. [Routines guide →](/en/docs/guides/routines) ## The daily loop Alloovium runs alongside your day — the cockpit, your meetings, your inbox, the site. ### Today The daily cockpit: your calendar spine, the "Needs you" queue of approvals and flags, action items, and the activity feed across your projects. [Today guide →](/en/docs/guides/today) ### Meetings Record or upload a meeting and get the transcript, summary, and extracted action items, decisions, and risks — each traceable to the moment it was said. [Meetings guide →](/en/docs/guides/meetings) ### Email and site Connected mailboxes make email part of the record — ask questions across threads. From site, a knock-off phone call or SMS becomes a diary entry, a drafted variation, an assigned action. [Site Calls & SMS guide →](/en/docs/guides/site-calls) ## Commercial The money loop — variations, claims, and the evidence behind them. ### Variations Changes get caught where they first appear — a site call, an email, a minute — and walked through intake to a drafted variation and a live register with cost and time impact. [Variations guide →](/en/docs/guides/variations) ### Claim evidence When money is contested, assemble the evidence pack — notices, site records, correspondence — every entry cited to its source document. [Claim Evidence guide →](/en/docs/guides/claims-evidence) ## Across companies Documents move between organisations with provenance, approvals, and signatures. ### Exchange Send document packages to other companies with a hash-chained audit trail. Recipients preview, accept or reject, and sign in place — the signature bound to the exact content. Paragraph-level DOCX diffs show precisely what changed between revisions. [Exchange guide →](/en/docs/guides/exchange) ### Working with Trades The subcontractor document loop through Exchange — submissions with review outcomes, document requests, and tracked transmittals. [Working with Trades guide →](/en/docs/guides/portal) ## Assurance Compliance signals, captured lessons, and audit-ready evidence. ### Compliance Expiring credentials, unsigned records, and conformance gaps surface before they bite — graded against the evidence actually on file. Assemble a frozen, shareable audit pack when the auditor calls. [Compliance guide →](/en/docs/guides/compliance) ### Lessons Capture what a mistake cost once, and Alloovium watches for the same pattern on every later project — resurfacing the lesson before the order goes out. [Lessons guide →](/en/docs/guides/lessons) ## In the tools you already use Your mailbox and 15+ connected systems. ### Integrations Outlook, Gmail, Google Calendar, SharePoint, OneDrive, Teams, Google Drive, Procore, and Autodesk — connected once at the organisation level, attached per project. Plus an MCP server so your own AI agents can query the record. [Integrations guide →](/en/docs/guides/integrations) Want the guided version? The [Getting Started overview](/en/docs/guides/getting-started) walks the first hour, and [Use Cases](/en/docs/use-cases) applies these features to pre-tender, claims, disputes, and handover. --- Canonical page: https://www.alloovium.com/en/docs/guides/getting-started/features --- ## Setting Up Your Workspace Source: https://www.alloovium.com/en/docs/guides/getting-started/workspace Getting started # Setting up your workspace One organisation, many projects, and user groups that decide who can do what. Understand the three layers and set them up in minutes. ## How the workspace fits together | Layer | What it holds | Where to manage it | | ------------ | ------------------------------------------------------------------------------------------------ | ------------------------------ | | Organization | Your company: the template library, playbook, lessons, contacts, and every member. | Organization in the sidebar | | Projects | One workspace per contract, tender, or site — documents, people, meetings, claims, timeline. | Projects in the sidebar | | User groups | Named sets of people with a permission matrix. Groups grant access — to the org and to projects. | Organization → People → Groups | Your account is scoped to your organisation — all colleagues who sign up with your company share one tenant, one library, and one set of groups. Projects are private by default; a project is visible only to the people whose groups grant access to it. ## Getting set up 1. 1 Create your account Sign up at alloovium.com with your work email. If your company already uses Alloovium, ask for an invite instead — you'll inherit the right access automatically. 2. 2 Create your first project Click "New Project" from the Projects page. Name it after the job — "Calabria Substation D&C" — and add a description so teammates recognise it. 3. 3 Upload your documents Drag files or folders into the project. PDF, DOCX, XLSX, and PPTX are all understood; drawings get title-block extraction automatically. 4. 4 Connect your systems Head to Integrations and connect Outlook or Gmail, SharePoint or Google Drive, and any construction platforms you run — the record stays current without re-uploading. ## Inviting your team Alloovium grants access through **user groups**. Each group carries a permission matrix — which areas its members can see and which actions they can take. When you invite someone, you pick their groups; the safe **Read-only** group is preselected, and adding someone to your admins group is what makes them an org admin. 1. 1 Open the invite panel Organization → People → Invite for company-wide access, or a project's People tab to invite someone to just that job. 2. 2 Enter their email and pick groups Choose one or more groups. Create a new group inline if none fits — you can copy an existing group's permissions as a starting point. 3. 3 Send it They get an email link, see who invited them and to what, and land in the right projects on first sign-in. Coming from roles? There are no per-person Viewer/Editor/Admin roles to assign anymore — group membership is the single source of truth for what someone can do. See [Permissions](/en/docs/guides/collaboration/permissions) for the full matrix. --- Canonical page: https://www.alloovium.com/en/docs/guides/getting-started/workspace --- ## History Source: https://www.alloovium.com/en/docs/guides/history DAILY WORKFLOW # History The answer you got last Tuesday is worth nothing if you can't find it on Thursday. History is the one page that holds everything you've touched — conversations, documents, projects, usage — searchable, filterable, and a click from picking up where you left off. ## Overview History lives in the sidebar under **Manage** and splits into four tabs: **Chats**, **Documents**, **Projects** and **Usage**. Each tab keeps its own scope controls — search, date, project — in the same header band, so narrowing any of them feels the same. ## Chats Every assistant conversation you've had, grouped by day, with search and a date filter above the list. Click any conversation to reopen it exactly where it ended — answers, citations and all. Conversations that belong to a project reopen inside that project; the rest open in the assistant. Rows can be selected in bulk (shift-click for a range) and deleted together when it's time to tidy up. ### Shared with me When a teammate [shares a conversation](/en/docs/guides/ai-assistant/sharing) with you, a **Shared with me**toggle appears beside the filters. That scope lists what's been shared to you — each row carrying the sharer's name — and opening one shows the full thread with its citations, read-only. ## Documents and Projects The **Documents**tab is every document you can access, searchable server-side — the search covers all pages of results, not just the rows on screen — and scopeable to a single project. It's often the fastest route back to a file when you remember the document but not where it was filed. The **Projects** tab is the same idea one level up: your projects with their activity, scoped and paged the same way. ## Usage The **Usage**tab shows what your workspace has consumed over a chosen period — useful for understanding how the team actually uses Alloovium, and for spotting when a project's activity spikes. History is personal Chats in History are yours — teammates don't see your conversations unless you share them, and admin-level visibility lives in the [admin console](/en/docs/guides/admin), not here. --- Canonical page: https://www.alloovium.com/en/docs/guides/history --- ## Project Inbox Source: https://www.alloovium.com/en/docs/guides/inbox DAILY WORKFLOW # Project Inbox Construction runs on email, so Today gives it a lane. The Inbox view shows recent messages from your connected mailbox routed to their projects and classified by what they need from you — without leaving the page where the rest of your day lives. ## Overview On [Today](/en/docs/guides/today), the wide left column toggles between **Calendar** and **Inbox**. Switch to Inbox and you get recent email routed to your projects from your connected Outlook or Gmail mailbox — each message tagged with the project it belongs to, listed with its sender and age, and sitting right beside the Needs attention queue so triage and action share one screen. ## Triage: Decision, Approval, FYI Every routed message is classified by what it actually needs from you: **Decision** when someone is waiting on a call, **Approval** when something needs signing off, **FYI**when it's information you should have but nothing more. Filter chips above the list cut it to one kind — so "what needs a decision from me this morning" is one click, not an inbox scroll. This is triage, not a mail client The Inbox view is deliberately thin: it exists so project email lands next to your calendar and your queue, classified. Your mailbox stays the system of record — Alloovium reads, it doesn't move or file your mail. ## From email to action item The same reading pass that classifies a message also extracts what it asks for. Requests buried in email — "can you confirm the exclusion zone by Friday" — surface as action items in the [Needs attention queue](/en/docs/guides/tasks), each linked back to the message it came from. Email stops being a second, private to-do list that only you can see. This is different from [Email Queries](/en/docs/guides/email-queries), which lets you _ask Alloovium questions by email_. The Inbox is about mail that arrives; Email Queries is about answers you request. ## Connecting a mailbox The Inbox fills from a connected mailbox — Outlook or Gmail, set up under [Integrations](/en/docs/guides/integrations). Until one is connected, the pane offers the connect step instead. Routing is per-project and automatic: messages are matched to the projects they concern, so a three-project week doesn't mean three inboxes. --- Canonical page: https://www.alloovium.com/en/docs/guides/inbox --- ## Integrations Source: https://www.alloovium.com/en/docs/guides/integrations Integrations # Integrations Connect Alloovium to your existing tools and document repositories to bring your documents in automatically. ## Overview Alloovium integrates with the tools your team already uses, so you can bring documents in from SharePoint, Google Drive, Procore and more without manual uploading. Connecting a source is a tenant-level action: synced files land in your tenant's file library once, and you attach them to projects whenever you want — instantly, without re-processing. Settings → Integrations page with three services connected — SharePoint, Google Drive, and Procore — each row showing a green Synced status, a file count, and a last-synced time, with the Add integration button top-right. Video · \~2 min Connect your first integration Start on an empty Integrations page, connect Google Drive via OAuth, pick two folders, watch the first sync ingest files into the library, then attach a synced file to a project with Add from library. ### How sync and project attachment work Connecting an integration is **tenant-level**, not project-level. When you sync a SharePoint site, Google Drive folder, Procore company, or any other source, the files are ingested into your tenant's file library — extracted, chunked, and embedded — exactly once. Attaching a synced file to a project is a separate, optional step that takes effect **instantly** because nothing needs to be re-processed. This means you can: * **Just sync.** Connect a source and let it ingest in the background. Decide later — or never — which projects each file belongs to. * **Attach from a project.** Inside any project, use _Add from library_ to search the tenant library, multi-select, and attach. Attach is a database insert, not an upload. * **Attach during sync (optional).** When starting a sync, you can optionally pick one or more projects to auto-attach files to as they finish ingesting. This is a shortcut for the "this folder is the project" case — it is not required. * **Reuse across projects.** The same synced spec or drawing can be attached to as many projects as you like. There is one copy of the file and one set of embeddings; project attachments are just pointers. Updates to a source file in SharePoint, Drive, or Procore re-sync automatically and propagate to every project the file is attached to — no manual re-attach needed. Read-only by default Integrations default to read-only — Alloovium reads and syncs your documents but does not modify or delete data in the connected service. A few integrations (such as Procore) offer optional write-back for specific actions; where available, it is opt-in and off by default. [Read our data handling commitments →](/en/docs/legal/data-handling) ## Microsoft * [SharePoint](/en/docs/guides/integrations/sharepoint) — sync document libraries and site folders into Alloovium projects. * [Microsoft Outlook](/en/docs/guides/integrations/outlook) — import project emails and attachments from your M365 mailbox. * [Microsoft Teams](/en/docs/guides/integrations/microsoft-teams) — sync channel messages and meeting recordings. Requires admin consent. * [Microsoft OneDrive](/en/docs/guides/integrations/one-drive) — sync personal and shared OneDrive folders via Microsoft OAuth. ## Google * [Google Drive](/en/docs/guides/integrations/google-drive) — sync shared drives and folders; supports Docs, Sheets, and Slides. * [Gmail](/en/docs/guides/integrations/gmail) — import project emails and attachments from selected Gmail labels. * [Google Calendar](/en/docs/guides/integrations/google-calendar) — sync meeting events and notes from selected calendars. ## Construction Platforms * [Procore](/en/docs/guides/integrations/procore) — import drawings, specs, submittals, and RFIs from your Procore projects. * [Autodesk APS](/en/docs/guides/integrations/autodesk) — connect BIM 360 and ACC via the Autodesk Platform Services OAuth app. * [Primavera P6](/en/docs/guides/integrations/primavera-p6) — schedule intelligence from the XER exports your scheduler already produces. File-based, no sign-in. ## Other Platforms * [Egnyte](/en/docs/guides/integrations/egnyte) — sync files and folders with Egnyte permissions carried through to document access. ## Tools * [MCP Server](/en/docs/guides/integrations/mcp) — connect Claude Desktop, Cursor, and other MCP-compatible AI tools to Alloovium. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations --- ## Oracle Aconex Source: https://www.alloovium.com/en/docs/guides/integrations/aconex docs.guides.aconex.eyebrow # docs.guides.aconex.title docs.guides.aconex.intro ## docs.guides.aconex.section\_overview docs.guides.aconex.overview\_p1 Read-only access Alloovium requests read-only access to Aconex. We never create, modify, or delete documents or transmittals. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.aconex.section\_what\_we\_access docs.guides.aconex.what\_we\_access\_body ## docs.guides.aconex.section\_setup docs.guides.aconex.setup\_p1 1. 1 docs.guides.aconex.setup\_step1\_title docs.guides.aconex.setup\_step1\_desc 2. 2 docs.guides.aconex.setup\_step2\_title docs.guides.aconex.setup\_step2\_desc 3. 3 docs.guides.aconex.setup\_step3\_title docs.guides.aconex.setup\_step3\_desc 4. 4 docs.guides.aconex.setup\_step4\_title docs.guides.aconex.setup\_step4\_desc Oracle support required Aconex API access requires Oracle to provision credentials for your organisation. This process typically takes 1–3 business days. Contact your Oracle account manager to start. ## docs.guides.aconex.section\_permissions docs.guides.aconex.permissions\_body ## docs.guides.aconex.section\_troubleshooting ### docs.guides.aconex.trouble\_credentials\_title docs.guides.aconex.trouble\_credentials\_desc ### docs.guides.aconex.trouble\_setup\_title docs.guides.aconex.trouble\_setup\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/aconex --- ## Asite Adoddle Source: https://www.alloovium.com/en/docs/guides/integrations/asite docs.guides.asite.eyebrow # docs.guides.asite.title docs.guides.asite.intro ## docs.guides.asite.section\_overview docs.guides.asite.overview\_p1 Read-only access Alloovium requests read-only access to Asite. We do not modify or delete any project documents. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.asite.section\_what\_we\_access docs.guides.asite.what\_we\_access\_body ## docs.guides.asite.section\_setup docs.guides.asite.setup\_p1 1. 1 docs.guides.asite.setup\_step1\_title docs.guides.asite.setup\_step1\_desc 2. 2 docs.guides.asite.setup\_step2\_title docs.guides.asite.setup\_step2\_desc 3. 3 docs.guides.asite.setup\_step3\_title docs.guides.asite.setup\_step3\_desc 4. 4 docs.guides.asite.setup\_step4\_title docs.guides.asite.setup\_step4\_desc ## docs.guides.asite.section\_permissions docs.guides.asite.permissions\_body ## docs.guides.asite.section\_troubleshooting ### docs.guides.asite.trouble\_title docs.guides.asite.trouble\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/asite --- ## Autodesk Forma (ACC / BIM 360) Source: https://www.alloovium.com/en/docs/guides/integrations/autodesk Integrations # Autodesk Forma (ACC / BIM 360) Connect Autodesk Forma (formerly Autodesk Construction Cloud) to sync drawings and documents from your projects into Alloovium. Setup takes about 10 minutes and only needs a Hub Admin — no developer account required. ## Overview The Autodesk integration connects Alloovium to your Forma (ACC) or BIM 360 hub through a dedicated Alloovium service account. Once connected, Alloovium syncs drawings and documents from the Files area of the projects you choose, picks up new versions automatically, and makes everything queryable alongside the rest of your project documents. Support for structured records — RFIs, submittals, and issues — is on the roadmap. Read-only access Alloovium requests read-only access to your Autodesk data. We never modify files, RFIs, or submittals in BIM 360 or ACC. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: documents and drawing files (PDFs) from the Files area of projects you add the Alloovium service account to, plus folder and version metadata. Access is scoped per project — we only see projects you explicitly invite the service account to. We do not read billing, payment, or account-level settings. ## Setup You don't need to create anything in Autodesk's developer portal. A Hub Admin authorises the Alloovium app once, then invites the Alloovium service account to each project you want to sync. The Client ID and service account email are shown on the Integrations → Autodesk page in Alloovium. 1. 1 Open Custom integrations In Forma, open Hub Admin → Settings → Custom integrations. If the menu item is missing, open any project's Files module once and check again — Autodesk enables it on first use of Docs. 2. 2 Add the Alloovium integration Click "Add custom integration", paste the Alloovium Client ID (SVaDGugX8w9INIwG5abapL5Jrq691TxeuYSuhI72wDSOADmj) and name it "Alloovium" exactly. On the next step, keep Document Management (and Account Administration, if offered) enabled, then save. The Alloovium service account is added to your hub automatically. 3. 3 Invite the service account to your projects Open each project you want to sync → Members → add the Alloovium service account (alloovium-sync@SVaDGugX8w9INIwG5abapL5Jrq691TxeuYSuhI72wDSOADmj.adskserviceaccount.com) as a Project member with the Data Management (Docs) product enabled — plus Build if your projects use it. Service accounts activate automatically; no invite acceptance needed. 4. 4 Connect in Alloovium In Alloovium, go to Integrations → Autodesk and click Connect. Your hub becomes visible within a couple of minutes of the first project invitation. 5. 5 Choose folders to sync Browse to the project's Files folders, pick the ones you want, and start the sync. New files and new versions in those folders are picked up automatically from then on. Autodesk Forma Hub Admin — Custom integrations page with the Add custom integration dialog open, the Alloovium Client ID pasted into the Client ID field, and the app name filled in above it. ACC is now Forma Autodesk renamed Autodesk Construction Cloud (ACC) to Forma in March 2026 — Autodesk Docs is now "Forma Data Management" and Autodesk Build is "Forma Build". The steps are the same on either branding. Legacy BIM 360 hubs are also supported. ## Permissions The Alloovium service account authenticates with read-only scopes and can only see projects it has been invited to: * **data:read** — read documents, drawings, and model data * **account:read** — read project members and project metadata ## Troubleshooting ### Custom integrations is missing from Hub Admin Open any project's Files module once, then re-check Settings → Custom integrations. If it still doesn't appear, email bim360appsactivations@autodesk.com with your account ID (Settings → Profile → View Account ID), account name, and admin email, requesting Docs API activation. ### Hub not appearing in Alloovium The hub only becomes visible to Alloovium once the service account is a member of at least one project. Confirm the custom integration shows as Active, the service account is a project member with Data Management enabled, and allow a few minutes for Autodesk to propagate the change. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/autodesk --- ## Autodesk APS — Bring Your Own App Source: https://www.alloovium.com/en/docs/guides/integrations/autodesk/byo-app Integrations · Autodesk APS # Bring your own Autodesk APS app Create an APS application in your own Autodesk account and install it in your ACC Hub as a Custom Integration. No App Store approval required — the app lives in your Autodesk account and only your Hub can use it. ## Why this route exists Autodesk’s App Store requires a formal review process. This route skips it: you create the APS app yourself, your ACC Hub Admin installs it as a Custom Integration, and you hand Alloovium the credentials. Nothing is shared with other Alloovium customers. Custom Integrations are an official Autodesk feature for exactly this use case — connecting approved third-party tools to a specific Hub without a public listing. ## Requirements * An Autodesk account with access to APS (developer.autodesk.com). Free account works. * ACC Hub Admin access (to install the Custom Integration in your Hub). * About 10–15 minutes. BIM 360 vs ACC Custom Integrations are an ACC feature. If your organisation is still on legacy BIM 360, upgrade to ACC first or use the standard APS OAuth flow via Alloovium support. ## 1\. Create an APS application 1. 1 Sign in to the APS Developer Portal Go to developer.autodesk.com/myapps and sign in with your Autodesk account. 2. 2 Create a new application Click "Create application". Give it a name like "Alloovium — Internal". Select BIM 360 and/or Autodesk Construction Cloud as the target APIs. 3. 3 Add the redirect URI In the app settings, add this redirect URI: https://app.alloovium.com/integrations/autodesk/callback. For staging, also add https://staging.alloovium.com/integrations/autodesk/callback. 4. 4 Copy Client ID and Client Secret After saving, copy the Client ID and Client Secret from the app detail page. You will need both in later steps. ## 2\. Install as a Custom Integration in ACC 1. 1 Open ACC Account Admin In Autodesk Construction Cloud, go to your Hub’s Account Admin (top-right menu → Account Admin). 2. 2 Navigate to Custom Integrations In Account Admin, click "Custom Integrations" in the left sidebar (under the Integrations section). 3. 3 Add your APS Client ID Click "Add Custom Integration". Paste your APS Client ID and give the integration a name. Save. 4. 4 Confirm the integration is listed The integration will appear in the Custom Integrations list with a status of Active. Alloovium can now authenticate against your Hub using your credentials. Hub Admin permission required Only ACC Hub Admins can add Custom Integrations. If you are not a Hub Admin, ask your project owner or IT administrator to add the integration. ## 3\. Connect in Alloovium 1. 1 Open Alloovium → Integrations → Autodesk Navigate to the Autodesk integration page. 2. 2 Enter your credentials Paste your APS Client ID and Client Secret into the credential fields. 3. 3 Authorise and select Hub Click Connect. You will be redirected to Autodesk to authorise access. After authorising, select the ACC Hub and projects you want to sync. ## Troubleshooting ### Custom Integrations tab not visible This tab only appears for ACC Hub Admins. Have an admin log in and check Account Admin → Integrations. ### Hub not appearing in Alloovium Ensure the Custom Integration was added to the correct ACC Hub. Each Hub is a separate environment. If your org has multiple Hubs, repeat the Custom Integration step for each. ### Redirect URI mismatch Verify the redirect URI in your APS app exactly matches https://app.alloovium.com/integrations/autodesk/callback (no trailing slash, exact domain). --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/autodesk/byo-app --- ## Bluebeam Studio Source: https://www.alloovium.com/en/docs/guides/integrations/bluebeam docs.guides.bluebeam.eyebrow # docs.guides.bluebeam.title docs.guides.bluebeam.intro ## docs.guides.bluebeam.section\_overview docs.guides.bluebeam.overview\_p1 Read-only access Alloovium requests read-only access to Bluebeam Studio sessions. We do not modify markups or session documents. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.bluebeam.section\_what\_we\_access docs.guides.bluebeam.what\_we\_access\_body ## docs.guides.bluebeam.section\_setup docs.guides.bluebeam.setup\_p1 1. 1 docs.guides.bluebeam.setup\_step1\_title docs.guides.bluebeam.setup\_step1\_desc 2. 2 docs.guides.bluebeam.setup\_step2\_title docs.guides.bluebeam.setup\_step2\_desc 3. 3 docs.guides.bluebeam.setup\_step3\_title docs.guides.bluebeam.setup\_step3\_desc 4. 4 docs.guides.bluebeam.setup\_step4\_title docs.guides.bluebeam.setup\_step4\_desc Bluebeam Developer Console application settings showing redirect URI field ## docs.guides.bluebeam.section\_permissions docs.guides.bluebeam.permissions\_body ## docs.guides.bluebeam.section\_troubleshooting ### docs.guides.bluebeam.trouble\_auth\_title docs.guides.bluebeam.trouble\_auth\_desc ### docs.guides.bluebeam.trouble\_sessions\_title docs.guides.bluebeam.trouble\_sessions\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/bluebeam --- ## Bluebeam Studio — Private Integration Source: https://www.alloovium.com/en/docs/guides/integrations/bluebeam/private-app docs.guides.bluebeam\_private\_app.eyebrow # docs.guides.bluebeam\_private\_app.title docs.guides.bluebeam\_private\_app.intro ## docs.guides.bluebeam\_private\_app.section\_why docs.guides.bluebeam\_private\_app.why\_p1 docs.guides.bluebeam\_private\_app.why\_p2 ## docs.guides.bluebeam\_private\_app.section\_requirements * docs.guides.bluebeam\_private\_app.req\_item1 * docs.guides.bluebeam\_private\_app.req\_item2 * docs.guides.bluebeam\_private\_app.req\_item3 docs.guides.bluebeam\_private\_app.callout\_studio\_prime\_title docs.guides.bluebeam\_private\_app.callout\_studio\_prime\_body ## docs.guides.bluebeam\_private\_app.section\_register 1. 1 docs.guides.bluebeam\_private\_app.register\_step1\_title docs.guides.bluebeam\_private\_app.register\_step1\_desc 2. 2 docs.guides.bluebeam\_private\_app.register\_step2\_title docs.guides.bluebeam\_private\_app.register\_step2\_desc 3. 3 docs.guides.bluebeam\_private\_app.register\_step3\_title docs.guides.bluebeam\_private\_app.register\_step3\_desc ## docs.guides.bluebeam\_private\_app.section\_create\_app 1. 1 docs.guides.bluebeam\_private\_app.create\_step1\_title docs.guides.bluebeam\_private\_app.create\_step1\_desc 2. 2 docs.guides.bluebeam\_private\_app.create\_step2\_title docs.guides.bluebeam\_private\_app.create\_step2\_desc 3. 3 docs.guides.bluebeam\_private\_app.create\_step3\_title docs.guides.bluebeam\_private\_app.create\_step3\_desc docs.guides.bluebeam\_private\_app.callout\_secret\_title docs.guides.bluebeam\_private\_app.callout\_secret\_body ## docs.guides.bluebeam\_private\_app.section\_connect 1. 1 docs.guides.bluebeam\_private\_app.connect\_step1\_title docs.guides.bluebeam\_private\_app.connect\_step1\_desc 2. 2 docs.guides.bluebeam\_private\_app.connect\_step2\_title docs.guides.bluebeam\_private\_app.connect\_step2\_desc 3. 3 docs.guides.bluebeam\_private\_app.connect\_step3\_title docs.guides.bluebeam\_private\_app.connect\_step3\_desc ## docs.guides.bluebeam\_private\_app.section\_troubleshooting ### docs.guides.bluebeam\_private\_app.trouble\_approval\_title docs.guides.bluebeam\_private\_app.trouble\_approval\_desc ### docs.guides.bluebeam\_private\_app.trouble\_no\_studio\_title docs.guides.bluebeam\_private\_app.trouble\_no\_studio\_desc ### docs.guides.bluebeam\_private\_app.trouble\_auth\_title docs.guides.bluebeam\_private\_app.trouble\_auth\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/bluebeam/private-app --- ## Box Source: https://www.alloovium.com/en/docs/guides/integrations/box docs.guides.box.eyebrow # docs.guides.box.title docs.guides.box.intro ## docs.guides.box.section\_overview docs.guides.box.overview\_p1 Read-only access Alloovium requests read-only access to Box. We never create, modify, or delete files or folders in your Box account. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.box.section\_what\_we\_access docs.guides.box.what\_we\_access\_body ## docs.guides.box.section\_setup docs.guides.box.setup\_p1 1. 1 docs.guides.box.setup\_step1\_title docs.guides.box.setup\_step1\_desc 2. 2 docs.guides.box.setup\_step2\_title docs.guides.box.setup\_step2\_desc 3. 3 docs.guides.box.setup\_step3\_title docs.guides.box.setup\_step3\_desc 4. 4 docs.guides.box.setup\_step4\_title docs.guides.box.setup\_step4\_desc ## docs.guides.box.section\_permissions docs.guides.box.permissions\_body ## docs.guides.box.section\_troubleshooting ### docs.guides.box.trouble\_jwt\_title docs.guides.box.trouble\_jwt\_desc ### docs.guides.box.trouble\_missing\_folders\_title docs.guides.box.trouble\_missing\_folders\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/box --- ## Confluence Source: https://www.alloovium.com/en/docs/guides/integrations/confluence docs.guides.confluence.eyebrow # docs.guides.confluence.title docs.guides.confluence.intro ## docs.guides.confluence.section\_overview docs.guides.confluence.overview\_p1 Read-only access Alloovium requests read-only access to Confluence. We never create, edit, or delete pages, spaces, or comments in your Confluence site. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.confluence.section\_what\_we\_access docs.guides.confluence.what\_we\_access\_body ## docs.guides.confluence.section\_setup docs.guides.confluence.setup\_p1 1. 1 docs.guides.confluence.setup\_step1\_title docs.guides.confluence.setup\_step1\_desc 2. 2 docs.guides.confluence.setup\_step2\_title docs.guides.confluence.setup\_step2\_desc 3. 3 docs.guides.confluence.setup\_step3\_title docs.guides.confluence.setup\_step3\_desc 4. 4 docs.guides.confluence.setup\_step4\_title docs.guides.confluence.setup\_step4\_desc ## docs.guides.confluence.section\_permissions docs.guides.confluence.permissions\_body ## docs.guides.confluence.section\_troubleshooting ### docs.guides.confluence.trouble\_site\_title docs.guides.confluence.trouble\_site\_desc ### docs.guides.confluence.trouble\_missing\_pages\_title docs.guides.confluence.trouble\_missing\_pages\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/confluence --- ## DocuSign Source: https://www.alloovium.com/en/docs/guides/integrations/docusign docs.guides.docusign.eyebrow # docs.guides.docusign.title docs.guides.docusign.intro ## docs.guides.docusign.section\_overview docs.guides.docusign.overview\_p1 Read-only access Alloovium requests read-only access to DocuSign. We never send envelopes, modify recipients, or alter any signing workflows. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.docusign.section\_what\_we\_access docs.guides.docusign.what\_we\_access\_body ## docs.guides.docusign.section\_setup docs.guides.docusign.setup\_p1 1. 1 docs.guides.docusign.setup\_step1\_title docs.guides.docusign.setup\_step1\_desc 2. 2 docs.guides.docusign.setup\_step2\_title docs.guides.docusign.setup\_step2\_desc 3. 3 docs.guides.docusign.setup\_step3\_title docs.guides.docusign.setup\_step3\_desc 4. 4 docs.guides.docusign.setup\_step4\_title docs.guides.docusign.setup\_step4\_desc ## docs.guides.docusign.section\_permissions docs.guides.docusign.permissions\_body ## docs.guides.docusign.section\_troubleshooting ### docs.guides.docusign.trouble\_consent\_title docs.guides.docusign.trouble\_consent\_desc ### docs.guides.docusign.trouble\_sandbox\_title docs.guides.docusign.trouble\_sandbox\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/docusign --- ## Dropbox Source: https://www.alloovium.com/en/docs/guides/integrations/dropbox docs.guides.dropbox.eyebrow # docs.guides.dropbox.title docs.guides.dropbox.intro ## docs.guides.dropbox.section\_overview docs.guides.dropbox.overview\_p1 Read-only access Alloovium requests read-only access to Dropbox. We never create, modify, or delete files or folders in your Dropbox account. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.dropbox.section\_what\_we\_access docs.guides.dropbox.what\_we\_access\_body ## docs.guides.dropbox.section\_setup docs.guides.dropbox.setup\_p1 1. 1 docs.guides.dropbox.setup\_step1\_title docs.guides.dropbox.setup\_step1\_desc 2. 2 docs.guides.dropbox.setup\_step2\_title docs.guides.dropbox.setup\_step2\_desc 3. 3 docs.guides.dropbox.setup\_step3\_title docs.guides.dropbox.setup\_step3\_desc 4. 4 docs.guides.dropbox.setup\_step4\_title docs.guides.dropbox.setup\_step4\_desc ## docs.guides.dropbox.section\_permissions docs.guides.dropbox.permissions\_body ## docs.guides.dropbox.section\_troubleshooting ### docs.guides.dropbox.trouble\_team\_title docs.guides.dropbox.trouble\_team\_desc ### docs.guides.dropbox.trouble\_missing\_folders\_title docs.guides.dropbox.trouble\_missing\_folders\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/dropbox --- ## Trimble e-Builder Source: https://www.alloovium.com/en/docs/guides/integrations/ebuilder docs.guides.ebuilder.eyebrow # docs.guides.ebuilder.title docs.guides.ebuilder.intro ## docs.guides.ebuilder.section\_overview docs.guides.ebuilder.overview\_p1 Read-only access Alloovium requests read-only access to e-Builder. We do not modify project data. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.ebuilder.section\_what\_we\_access docs.guides.ebuilder.what\_we\_access\_body ## docs.guides.ebuilder.section\_setup docs.guides.ebuilder.setup\_p1 1. 1 docs.guides.ebuilder.setup\_step1\_title docs.guides.ebuilder.setup\_step1\_desc 2. 2 docs.guides.ebuilder.setup\_step2\_title docs.guides.ebuilder.setup\_step2\_desc 3. 3 docs.guides.ebuilder.setup\_step3\_title docs.guides.ebuilder.setup\_step3\_desc 4. 4 docs.guides.ebuilder.setup\_step4\_title docs.guides.ebuilder.setup\_step4\_desc Trimble support required e-Builder API access must be provisioned by Trimble support. Contact your account manager to request credentials before attempting to connect. ## docs.guides.ebuilder.section\_permissions docs.guides.ebuilder.permissions\_body ## docs.guides.ebuilder.section\_troubleshooting ### docs.guides.ebuilder.trouble\_title docs.guides.ebuilder.trouble\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/ebuilder --- ## Egnyte Source: https://www.alloovium.com/en/docs/guides/integrations/egnyte Integrations # Egnyte Connect Egnyte to sync files and folders from your content platform into Alloovium, with folder permissions carried through. ## Overview The Egnyte integration syncs the folders you select — from Shared and Private spaces — into your tenant's file library, where documents become searchable, citable project records. A rolling poll picks up new and updated files automatically, and Egnyte folder permissions are mapped through to document access in Alloovium, so people only find what they could already open at the source. Read-only access Alloovium reads and syncs the folders you select — it never modifies or deletes anything in Egnyte. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: the files and folders you select from Shared and Private spaces, file metadata (names, dates, sizes), and folder permission assignments — used only to map access. We do not modify or delete anything in Egnyte. ## Setup You need an Egnyte account with access to the folders you want to sync. Connecting takes a couple of minutes: 1. 1 Open Integrations Go to Settings → Integrations → Egnyte and click Connect. 2. 2 Sign in to Egnyte You are redirected to your Egnyte domain to sign in and authorise read-only access. 3. 3 Choose folders Pick the Shared or Private folders to sync. Only folders your Egnyte user can access appear in the picker. 4. 4 Start the sync Files ingest into your library in the background, and the rolling poll keeps them fresh as documents change in Egnyte. Attach synced files to projects whenever you like. ## Permissions Access is read-only OAuth — Alloovium never writes to Egnyte. Folder permissions carry through: a document synced from a restricted Egnyte folder is only findable in Alloovium by people with access at the source (or when it is explicitly shared inside Alloovium). ## Troubleshooting 1. 1 Folders not appearing in the picker The picker shows only folders the connected Egnyte user can access. Reconnect with an account that has access to the folders you need, or ask your Egnyte admin to grant it. 2. 2 A teammate can't find a synced document Permission mapping is working as intended — they don't have access to that folder in Egnyte. Grant it at the source, or share the document with them inside Alloovium. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/egnyte --- ## Fieldwire Source: https://www.alloovium.com/en/docs/guides/integrations/fieldwire docs.guides.fieldwire.eyebrow # docs.guides.fieldwire.title docs.guides.fieldwire.intro ## docs.guides.fieldwire.section\_overview docs.guides.fieldwire.overview\_p1 Read-only access Alloovium uses a read-only API key. We never create, modify, or delete tasks, plans, or forms in Fieldwire. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.fieldwire.section\_what\_we\_access docs.guides.fieldwire.what\_we\_access\_body ## docs.guides.fieldwire.section\_setup docs.guides.fieldwire.setup\_p1 1. 1 docs.guides.fieldwire.setup\_step1\_title docs.guides.fieldwire.setup\_step1\_desc 2. 2 docs.guides.fieldwire.setup\_step2\_title docs.guides.fieldwire.setup\_step2\_desc 3. 3 docs.guides.fieldwire.setup\_step3\_title docs.guides.fieldwire.setup\_step3\_desc 4. 4 docs.guides.fieldwire.setup\_step4\_title docs.guides.fieldwire.setup\_step4\_desc Fieldwire Account Settings showing the API & Integrations tab with a new API key Save your API key immediately Fieldwire shows the API key only once at creation time. Copy it before closing the dialog. If you lose it, generate a new key and revoke the old one from the same page. ## docs.guides.fieldwire.section\_permissions docs.guides.fieldwire.permissions\_body ## docs.guides.fieldwire.section\_troubleshooting ### docs.guides.fieldwire.trouble\_invalid\_key\_title docs.guides.fieldwire.trouble\_invalid\_key\_desc ### docs.guides.fieldwire.trouble\_missing\_projects\_title docs.guides.fieldwire.trouble\_missing\_projects\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/fieldwire --- ## Fieldwire — Generate an API Key Source: https://www.alloovium.com/en/docs/guides/integrations/fieldwire/api-key docs.guides.fieldwire\_api\_key.eyebrow # docs.guides.fieldwire\_api\_key.title docs.guides.fieldwire\_api\_key.intro ## docs.guides.fieldwire\_api\_key.section\_why docs.guides.fieldwire\_api\_key.why\_p1 docs.guides.fieldwire\_api\_key.why\_p2 ## docs.guides.fieldwire\_api\_key.section\_requirements * docs.guides.fieldwire\_api\_key.req\_item1 * docs.guides.fieldwire\_api\_key.req\_item2 * docs.guides.fieldwire\_api\_key.req\_item3 docs.guides.fieldwire\_api\_key.callout\_plan\_title docs.guides.fieldwire\_api\_key.callout\_plan\_body ## docs.guides.fieldwire\_api\_key.section\_generate 1. 1 docs.guides.fieldwire\_api\_key.generate\_step1\_title docs.guides.fieldwire\_api\_key.generate\_step1\_desc 2. 2 docs.guides.fieldwire\_api\_key.generate\_step2\_title docs.guides.fieldwire\_api\_key.generate\_step2\_desc 3. 3 docs.guides.fieldwire\_api\_key.generate\_step3\_title docs.guides.fieldwire\_api\_key.generate\_step3\_desc 4. 4 docs.guides.fieldwire\_api\_key.generate\_step4\_title docs.guides.fieldwire\_api\_key.generate\_step4\_desc 5. 5 docs.guides.fieldwire\_api\_key.generate\_step5\_title docs.guides.fieldwire\_api\_key.generate\_step5\_desc docs.guides.fieldwire\_api\_key.callout\_token\_once\_title docs.guides.fieldwire\_api\_key.callout\_token\_once\_body ## docs.guides.fieldwire\_api\_key.section\_connect 1. 1 docs.guides.fieldwire\_api\_key.connect\_step1\_title docs.guides.fieldwire\_api\_key.connect\_step1\_desc 2. 2 docs.guides.fieldwire\_api\_key.connect\_step2\_title docs.guides.fieldwire\_api\_key.connect\_step2\_desc 3. 3 docs.guides.fieldwire\_api\_key.connect\_step3\_title docs.guides.fieldwire\_api\_key.connect\_step3\_desc ## docs.guides.fieldwire\_api\_key.section\_troubleshooting ### docs.guides.fieldwire\_api\_key.trouble\_no\_api\_tab\_title docs.guides.fieldwire\_api\_key.trouble\_no\_api\_tab\_desc ### docs.guides.fieldwire\_api\_key.trouble\_invalid\_title docs.guides.fieldwire\_api\_key.trouble\_invalid\_desc ### docs.guides.fieldwire\_api\_key.trouble\_no\_projects\_title docs.guides.fieldwire\_api\_key.trouble\_no\_projects\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/fieldwire/api-key --- ## Gmail Source: https://www.alloovium.com/en/docs/guides/integrations/gmail Integrations # Gmail Connect Gmail to bring project-related emails and attachments into Alloovium for AI-powered search. ## Overview The Gmail integration connects Alloovium to your Gmail inbox. Emails and attachments from project-related conversations can be synced into Alloovium, giving your AI assistant access to project correspondence alongside drawings, specs, and RFIs. Read-only access Alloovium requests read-only access to Gmail. We never send emails, modify labels, or delete messages. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: email subjects, body text, and attachments from the labels or threads you select. We do not read emails you have not included in the sync scope. We do not have access to account settings, contacts, or other Google services. ## Setup Gmail uses Google OAuth. You will sign in with your Google account and grant Alloovium read access to selected emails. 1. 1 Click Connect Go to Integrations → Gmail in Alloovium and click Connect. You will be redirected to Google's authorisation screen. 2. 2 Sign in with Google Sign in with the Gmail account you want to connect. Review the permissions and click Allow. 3. 3 Select labels to sync Choose which Gmail labels or project-specific folders to sync. Only emails in those labels will be imported. 4. 4 Initial import Alloovium will import emails from your selected labels. New emails are picked up automatically. ## Permissions We request: gmail.readonly (read emails from labels you select). We do not request send, modify, or delete permissions. ## Troubleshooting ### Google consent screen warns about unverified app During initial setup, Google may show a warning about an unverified app. Click Advanced → Go to Alloovium (unsafe) to proceed. This warning appears until the app completes Google's verification process. ### Sync stopped mid-import Gmail API has rate limits. If your inbox is very large, the initial import may be throttled. Alloovium will resume automatically — check the integrations panel for progress. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/gmail --- ## Google Calendar Source: https://www.alloovium.com/en/docs/guides/integrations/google-calendar Integrations # Google Calendar Connect Google Calendar to sync project meeting notes and events into Alloovium. ## Overview The Google Calendar integration imports calendar events from your selected calendars into Alloovium. Meeting details, attached notes, and event descriptions become part of your project's AI knowledge base. Read-only access Alloovium requests read-only access to Google Calendar. We never create, modify, or delete calendar events. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: event titles, descriptions, attendees, dates, times, and attached files from the calendars you select. We do not read draft events or events you have declined without accepting. ## Setup Google Calendar uses the same Google OAuth client as Gmail. If you have already connected Gmail, you can connect Calendar with one additional step. 1. 1 Click Connect Go to Integrations → Google Calendar in Alloovium and click Connect. 2. 2 Sign in with Google Sign in with your Google account. Review the Calendar-specific permissions and click Allow. 3. 3 Select calendars to sync Choose which calendars to import. You can include personal, team, or project-specific calendars. 4. 4 Events start syncing Alloovium imports upcoming and recent past events. New events are picked up as they are created. ## Permissions We request: calendar.readonly (read event data from selected calendars). We do not create, modify, or delete calendar events. ## Troubleshooting ### Calendars not appearing Only calendars your Google account owns or has been shared are visible. Resource calendars (e.g. meeting rooms) are not included unless shared with your account. ### Shared team calendar missing Ensure the shared calendar has been added to your Google Calendar account (not just subscribed to). It must appear in your calendar list to be synced. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/google-calendar --- ## Google Drive Integration Source: https://www.alloovium.com/en/docs/guides/integrations/google-drive Integrations # Google Drive Connect Google Drive once at the tenant level. Synced files land in your library automatically and can be attached to any project instantly. ## Overview Connect Google Drive to sync files from shared drives and folders into your Alloovium tenant library. The integration supports Google Docs, Sheets, Slides (converted to DOCX/XLSX/PPTX), and standard files (PDF, Word, etc.). Connection is tenant-level — files are not tied to a specific project at sync time. Google Workspace files (Docs, Sheets, Slides) are automatically exported to plain text or PDF for ingestion. Native files (PDFs, Word documents, images) are processed directly. Folder-level permissions from Google Drive are respected — team members only see documents their account has access to. Read-only access Alloovium requests read-only access to Google Drive. We never create, modify, or delete any files in your Drive. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read files and folder structure from the specific folders you select during setup. This includes PDFs, Word documents, images, and Google Workspace files converted to text. We do not access files outside your selected folders, your Google account settings, contacts, or any other Google services. ## Setup Connecting Google Drive takes about two minutes. The connection is **tenant-level** — it does not bind to any single project. Synced files land in your tenant's file library, where any project can pull them in instantly. 1. 1 Connect Google Drive Go to Integrations → Google Drive. Click "Connect" and sign in with a Google account that has access to the drives you want to sync. 2. 2 Select folders to sync Choose which folders or shared drives to include. Files from these locations will be ingested into your tenant library — extracted, chunked, and embedded — once. Attaching them to projects is a separate step. 3. 3 Configure sync frequency Set how often Alloovium should check for new or updated files (every 15 minutes, hourly, or daily). Updates flow through to every project the file is attached to. 4. 4 Run the initial sync Alloovium begins ingesting files from your selected folders. Progress is shown in the integrations panel. Large folders may take several minutes. Alloovium Google Drive folder picker mid-setup: the Drive folder tree with two project folders ticked, a Shared Drive visible in the tree, and the Start sync button enabled below. Shared drives If your team stores project documents in a Google Workspace Shared Drive, connect using an account that has at least Viewer access to that shared drive. Personal My Drive and Shared Drives are both supported. ## Attaching files to projects Once a file is in your tenant library, attaching it to a project is instant — no re-upload, no re-extraction. You have two options: * **Attach from inside a project.** Open any project, click _Add from library_, search or filter by source (Google Drive), multi-select, and attach. Useful when you decide later which files belong where. * **Pre-attach during sync (optional).** When you start a sync, you can optionally pick one or more projects to auto-attach files to as they finish ingesting. Skippable — leave it empty to just sync into the library. The same Drive file can be attached to as many projects as you like. There is one copy of the file and one set of embeddings; project attachments are pointers. If the file is updated in Drive, every project sees the new version automatically. ## Permissions Alloovium requests the following Google OAuth scopes: * **drive.readonly** — read files and folders from drives you own or have access to * **drive.metadata.readonly** — read file metadata (names, sizes, dates) We do not request `drive.file`, `drive.appdata`, or any write scope. ## Troubleshooting ### Google consent screen warns about unverified app During initial setup, Google may display a warning about an unverified app. Click Advanced, then "Go to Alloovium (unsafe)" to proceed. This warning appears until the app completes Google's verification process. The integration itself is fully secure. ### Folders not showing in Alloovium Only folders your connected Google account has at least Viewer access to will appear. If a shared drive is missing, ensure your account has been added as a member of that shared drive — a subscription link is not sufficient. ### Google Docs not appearing as expected Google Docs, Sheets, and Slides are exported to plain text or PDF at import time. If a document shows as blank, it may have been created with drawing-only content that cannot be extracted as text. Contact support if you need custom export handling. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/google-drive --- ## MCP Server Source: https://www.alloovium.com/en/docs/guides/integrations/mcp Integrations # MCP Server Connect Claude Desktop, Cursor, and other MCP-compatible AI tools directly to your Alloovium projects. ## MCP Server Alloovium exposes a Model Context Protocol (MCP) server that allows AI tools like Claude Desktop, Cursor, and other MCP-compatible clients to access your Alloovium projects directly. This lets you query your project documents from within your AI coding or productivity tool. In plain terms: once connected, you can sit in Claude Desktop, Cursor, or any other MCP-compatible AI tool and ask questions that get answered from your Alloovium project documents — "what’s the defects liability period on this contract?" — without opening Alloovium at all. The AI tool calls the Alloovium MCP server behind the scenes; the server searches your project vault and hands back the relevant passages with citations, so the agent’s answers are grounded in your actual documents. Connections are **read-only by default** — a connected agent can search and read your documents but cannot change anything in Alloovium. Authentication uses a personal API key, which you create under **Settings → Developer**. The key carries your own permissions, so an agent connected with it can only see the projects you can see. See the [MCP Server setup guide](/en/developers/mcp) for full configuration instructions including authentication, available tools, and example configurations for Claude Desktop and Cursor. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/mcp --- ## Microsoft Teams Source: https://www.alloovium.com/en/docs/guides/integrations/microsoft-teams Integrations # Microsoft Teams Connect Microsoft Teams to sync channel messages and meeting recordings into Alloovium. ## Overview The Teams integration pulls channel messages, meeting recordings, and chat transcripts from Microsoft Teams into Alloovium. Project discussions, decisions, and action items captured in Teams become searchable and queryable alongside other project documents. Read-only access Alloovium requests read-only access to Teams channels. We never post messages, create channels, or modify team settings. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: channel messages, meeting chat transcripts, and recording transcripts from the channels you select. We do not read private 1:1 chats. We do not read files stored in Teams channels unless the SharePoint integration is also connected. ## Setup Microsoft Teams uses the same Microsoft OAuth application as SharePoint and Outlook. Admin consent from your Microsoft 365 tenant administrator is typically required. 1. 1 Click Connect Go to Integrations → Microsoft Teams in Alloovium and click Connect. 2. 2 Sign in with Microsoft Sign in with your Microsoft 365 account. Your IT administrator may need to pre-consent the Teams.ReadBasic.All permission. 3. 3 Select teams and channels Choose which Teams and channels to sync. Alloovium will import message history and monitor for new messages. 4. 4 Meeting recordings If meeting recordings are stored in SharePoint, connect the SharePoint integration to also index recording transcripts. Admin consent required The **ChannelMessage.Read.All** permission requires Microsoft 365 tenant administrator consent. Your IT administrator must approve this in Azure Active Directory before users can connect. ## Permissions We request: Team.ReadBasic.All (list teams), ChannelMessage.Read.All (read channel messages). Admin consent is required for ChannelMessage.Read.All in most tenants. ## Troubleshooting ### ChannelMessage.Read.All requires admin consent This permission cannot be granted by regular users. Your Microsoft 365 administrator must consent in Azure AD → Enterprise Applications → Alloovium → Permissions. ### Teams not appearing You must be a member of the team to see it in the list. Check with your Teams administrator if an expected team is missing. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/microsoft-teams --- ## Newforma Konekt Source: https://www.alloovium.com/en/docs/guides/integrations/newforma docs.guides.newforma.eyebrow # docs.guides.newforma.title docs.guides.newforma.intro ## docs.guides.newforma.section\_overview docs.guides.newforma.overview\_p1 Read-only access Alloovium requests read-only access to Newforma Konekt. We never modify project data. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.newforma.section\_what\_we\_access docs.guides.newforma.what\_we\_access\_body ## docs.guides.newforma.section\_setup docs.guides.newforma.setup\_p1 1. 1 docs.guides.newforma.setup\_step1\_title docs.guides.newforma.setup\_step1\_desc 2. 2 docs.guides.newforma.setup\_step2\_title docs.guides.newforma.setup\_step2\_desc 3. 3 docs.guides.newforma.setup\_step3\_title docs.guides.newforma.setup\_step3\_desc 4. 4 docs.guides.newforma.setup\_step4\_title docs.guides.newforma.setup\_step4\_desc Administrator setup required Newforma Konekt API access cannot be self-provisioned. Your Newforma administrator or Newforma support must enable API access before you can connect. ## docs.guides.newforma.section\_permissions docs.guides.newforma.permissions\_body ## docs.guides.newforma.section\_troubleshooting ### docs.guides.newforma.trouble\_credentials\_title docs.guides.newforma.trouble\_credentials\_desc ### docs.guides.newforma.trouble\_contact\_title docs.guides.newforma.trouble\_contact\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/newforma --- ## Microsoft OneDrive Source: https://www.alloovium.com/en/docs/guides/integrations/one-drive Integrations # Microsoft OneDrive Connect OneDrive to sync personal and shared drives into Alloovium via Microsoft OAuth. ## Overview The OneDrive integration connects Alloovium to your Microsoft 365 OneDrive. Files, folders, and shared documents are synced so your AI assistant can query them alongside SharePoint libraries and other project documents. Read-only access Alloovium requests read-only access to OneDrive. We never create, modify, or delete files or folders in your drives. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: files and folder structure from the drives and folders you select. We access both your personal OneDrive and shared drives you have access to. We do not read personal files outside the folders you select. ## Setup OneDrive uses the same Microsoft OAuth application as SharePoint. If SharePoint is already connected, you can add OneDrive with one click. 1. 1 Click Connect Go to Integrations → OneDrive in Alloovium and click Connect. You will be redirected to Microsoft to authorise. 2. 2 Sign in with Microsoft Sign in with your Microsoft 365 account. Review the permissions (Files.Read) and click Accept. 3. 3 Select folders to sync Browse your OneDrive and select the folders to sync. Subfolders are included automatically. 4. 4 Initial sync Alloovium imports all files in your selected folders. Changes are picked up automatically on subsequent syncs. Shared OAuth app with SharePoint OneDrive and SharePoint use the same Microsoft OAuth application in Alloovium. If SharePoint is already connected, adding OneDrive requires only the additional Files.Read scope. ## Permissions We request: Files.Read (read files from OneDrive), Sites.Read.All (read shared drives). We do not request write access. ## Troubleshooting ### Personal OneDrive vs OneDrive for Business Alloovium connects to OneDrive for Business (Microsoft 365). Personal OneDrive accounts (outlook.com, hotmail.com) are not supported. ### Large OneDrive taking long to sync Initial syncs for large drives may take several hours. Alloovium processes files in batches — check the integrations panel for current progress. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/one-drive --- ## Microsoft Outlook Source: https://www.alloovium.com/en/docs/guides/integrations/outlook Integrations # Microsoft Outlook Connect Outlook to sync project emails and meeting notes into Alloovium via Microsoft OAuth. ## Overview The Outlook integration connects Alloovium to your Microsoft 365 mailbox. Project-related emails, meeting summaries, and attachments are imported into Alloovium so your AI assistant can answer questions across email correspondence and project documents together. Read-only access Alloovium requests read-only access to your Outlook mailbox. We never send emails, move messages, or modify folders. [Read our data handling commitments →](/en/docs/legal/data-handling) ## What data we access We read: email subjects, body text, attachments, and sender/recipient information from the folders you select. We do not read your calendar, contacts, or tasks unless those integrations are separately connected. ## Setup Outlook uses the same Microsoft OAuth application as SharePoint. If you have already connected SharePoint, you may be able to skip the authentication step. 1. 1 Click Connect Go to Integrations → Outlook in Alloovium and click Connect. You will be redirected to Microsoft to authorise access. 2. 2 Sign in with Microsoft Sign in with your Microsoft 365 account. Review the permissions (Mail.Read) and click Accept. 3. 3 Select folders to sync Choose which Outlook folders or labels to import. Alloovium monitors those folders for new emails. 4. 4 Initial import Alloovium imports existing emails from your selected folders and then monitors for new messages. Shared OAuth app with SharePoint Outlook and SharePoint use the same Microsoft OAuth application in Alloovium. If SharePoint is already connected, you may only need to grant the additional Mail.Read scope. ## Permissions We request: Mail.Read (read email messages). We do not request Mail.Send, Mail.ReadWrite, or access to calendar or contacts. ## Troubleshooting ### Admin consent required Some Microsoft 365 tenants require administrator consent before users can connect third-party apps. Your IT administrator can grant consent in Azure AD → Enterprise Applications → Alloovium → Permissions. ### Conditional Access blocking sign-in If your organisation has Conditional Access policies restricting third-party app access, your IT administrator will need to create an exception or grant Alloovium access explicitly. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/outlook --- ## Primavera P6 Source: https://www.alloovium.com/en/docs/guides/integrations/primavera-p6 Integrations # Primavera P6 Schedule intelligence from the XER your scheduler already exports — no P6 account connection needed. ## Overview The Primavera P6 integration is file-based: Alloovium reads the schedule exports (.xer or P6 .xml) your scheduler already produces and turns them into live schedule intelligence for the project — a snapshot on every update, a clear account of what changed between programs, milestone trends across successive updates, and checks of the schedule against what the site record says is actually happening. There is no P6 account to connect and nothing to install. File-based — no sign-in Nothing signs in to your P6 account. Alloovium only reads the schedule exports you choose to share, and your program in P6 is never touched. [Read our data handling commitments →](/en/docs/legal/data-handling) ## Getting a schedule in Each new export becomes a snapshot on the project's Schedule page. A schedule can land in Alloovium three ways: 1. 1 Upload it Drop the .xer or P6 .xml export on the project's Schedule page. 2. 2 Let it arrive on its own New .xer files landing in connected Procore or SharePoint folders are picked up automatically. 3. 3 Email it Forward the scheduler's export email — the attached .xer is captured automatically. ## What you get * Schedule snapshots from every XER or P6 XML export, kept in order on the Schedule page. * What changed between updates — dates, logic, and the critical path. * Milestone trends across successive programs, so slippage shows up as a pattern, not a surprise. * Site-signal checks against the live schedule — diaries, meetings, and reports that contradict the program get flagged. ## Permissions Nothing signs in to your P6 account. Alloovium only reads the schedule exports you choose to share, and your program in P6 is never touched. Schedule intelligence is project-scoped: it is visible to people with access to that project. ## Troubleshooting 1. 1 An export isn't appearing Check the file is a .xer or P6 .xml export, and — for automatic capture — that it landed in a folder that is part of a connected Procore or SharePoint sync. 2. 2 The comparison looks wrong Snapshots compare in sequence on the Schedule page. Make sure successive exports come from the same program (same project and WBS) — an export of a different program starts its own history. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/primavera-p6 --- ## Procore Integration Source: https://www.alloovium.com/en/docs/guides/integrations/procore Integrations # Procore Connect your own Procore Data Connection app to Alloovium and import drawings, specifications, submittals, and RFIs. ## Overview The Procore integration imports documents from your Procore projects into Alloovium. Supported document types include Drawings, Specifications, Submittals, and RFIs. You can sync one or more document types, and select specific folders or the entire project. Procore connections in Alloovium use a Bring Your Own App model. You create a Procore Data Connection app in your own Procore Developer Portal, install it on your company, then paste its Client ID and Client Secret into Alloovium. Bring Your Own App only Alloovium does not maintain a shared Procore Marketplace listing or shared OAuth client. Every customer connects through their own Procore Data Connection app. This keeps tokens, scopes, and audit trails fully under your control. ## Connect Procore Connecting Procore is a two-step process: create the app in Procore, then plug its credentials into Alloovium. ### 1\. Create your Procore Data Connection app Set up a Data Connection app inside your Procore Developer Portal, install it on your company, and grant the scopes Alloovium needs (Drawings, Specifications, Submittals, RFIs, Documents). [Open the full Procore — Bring Your Own App setup guide →](/en/docs/guides/integrations/procore/byo-app) ### 2\. Connect it to Alloovium Once your Data Connection app is installed in Procore, paste its credentials into Alloovium. 1. 1 Open Integrations → Procore in Alloovium From the left navigation choose Integrations, then click the Procore tile. 2. 2 Click Connect with my Procore app You will be prompted for the Client ID and Client Secret of your Procore Data Connection app. 3. 3 Authorise the connection Procore opens its consent screen. Approve the requested scopes and you will be returned to Alloovium with an active connection. Where Alloovium uses your credentials Your Client ID and Client Secret are used only to start the authorised connection flow. Alloovium stores the resulting access tokens per tenant and never shares them with other customers. ## Select documents to sync After connecting, choose which Procore projects and document types to sync into Alloovium. 1. 1 Open the Procore browser From the integration page, click Browse Procore to load your accessible projects. 2. 2 Pick a project Select the Procore project you want to mirror into Alloovium. You can repeat this for multiple projects. 3. 3 Choose document types Tick Drawings, Specifications, Submittals, RFIs, or any combination. You can also drill into specific folders if you only want a subset. 4. 4 Click Sync Alloovium queues the selected items for ingestion. You can leave the page — sync continues in the background. ## How sync works Each sync run goes through three phases: 1. Fetch — Alloovium calls Procore's API using your installed app's tokens to list available items in the selected scope. 2. Check — Alloovium compares Procore item versions against what is already in Alloovium and skips anything unchanged. 3. Download — only new or updated files are pulled, ingested, and indexed for search and chat. Syncs are incremental After the first full sync, subsequent runs only transfer files that changed in Procore. This keeps re-syncs fast and within Procore's API quotas. Deletes are not propagated If a document is deleted in Procore, Alloovium does not automatically remove the previously ingested copy. Remove it from Alloovium manually if you want it gone. ## Troubleshooting ### I don't see any Procore projects Your Procore Data Connection app must be installed on the company that owns the projects, and the installing user must have access to those projects. Re-install the app or ask a Procore admin to grant access, then refresh the integration page. ### Some documents are missing after sync Procore returns documents based on the scopes granted to your Data Connection app. If a document type is missing, confirm that the matching scope (e.g. Drawings, Specifications, Submittals, RFIs) is enabled in your app and that the connecting user can see those documents in Procore directly. ### Procore rejected my credentials Double-check that the Client ID and Client Secret come from a Data Connection app (not a Marketplace app), that the app is installed on your company, and that the secret has not been rotated. Regenerate the secret in the Procore Developer Portal and reconnect if in doubt. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/procore --- ## Procore — Bring Your Own App Source: https://www.alloovium.com/en/docs/guides/integrations/procore/byo-app Integrations · Procore # Bring your own Procore app Create a Data Connection app in your own Procore Developer Portal and connect it to Alloovium. No Marketplace approval required — the app lives in your Procore account and only you use it. ## Why this route exists Procore is tightening its public Marketplace with a partner certification queue that can take months. This route skips the queue entirely: **you** create the app in your own Procore account, **you** install it on your own company, and you hand Alloovium the credentials so it can talk to Procore on your behalf. Nothing is shared with other Alloovium customers. Your credentials are encrypted at rest and only ever used by your tenant. ## Requirements * A Procore account with **Company-level admin** permissions. * Access to the **Procore Developer Portal**. Most Enterprise accounts have this enabled by default. If you don't see it, ask your Procore account manager to turn it on for your company. * About 10–15 minutes. Developer Portal access If your Procore plan does not include Developer Portal access, contact Alloovium support — we can discuss a concierge setup where we create the app on our side and install it on your company. ## 1\. Create the Procore app 1. 1 Open the Procore Developer Portal Log in to Procore, then go to developers.procore.com and click "My Apps". 2. 2 Click "Create New App" Give it a name like "Alloovium — Internal". You can keep it sandboxed while you test, then promote to production later. 3. 3 Choose app type Select "Data Connection App" (DCA). This is the service-account style that uses client credentials — no per-user login required. 4. 4 Verify developer account if prompted Procore may email the company admin for one-time verification. This is usually instant for Enterprise accounts and takes up to a couple of business days for others. ## 2\. Configure scopes & manifest Alloovium needs read access to the document and decision data it indexes. In the app's **Configuration Builder** tab, enable the following permissions: * **Company Directory** — read * **Projects** — read * **Documents** — read * **Drawings** — read * **Specifications** — read * **Submittals** — read * **RFIs** — read * **Change Orders** — read You do not need to grant write permissions unless you plan to enable Alloovium's writeback features (CR → Procore comment/status updates). Save the configuration. A new **version** of the app is created — this is the version you'll install. ## 3\. Install on your company 1. 1 Open Company Admin → App Management Back in Procore (not the Developer Portal), go to Company Admin and pick App Management. 2. 2 Install Custom App Click "Install App" → "Install Custom App". Paste the App Version ID shown in your Developer Portal app and confirm. 3. 3 Copy the Client ID and Client Secret In the Developer Portal, open your app → OAuth Credentials tab. You will see a Client ID and Client Secret specific to this install. Copy both — the secret is shown only once. Protect your credentials Treat the Client Secret like a password. Anyone with it can read your Procore data on behalf of this app. Rotate it from the OAuth Credentials tab if you suspect it was exposed. ## 4\. Connect in Alloovium 1. 1 Open Alloovium → Integrations → Procore Switch to the "Bring your own app" tab at the top of the connect panel. 2. 2 Paste the credentials Enter a friendly connection name, the Client ID, and the Client Secret. Leave the Company ID blank to auto-detect, or paste one if your app has access to multiple companies. 3. 3 Click Connect Alloovium validates the credentials by fetching a token, saves them encrypted, and returns to the main Procore page where you can start selecting projects to sync. ## Troubleshooting ### Invalid DMSA credentials — token fetch failed The Client ID and Secret don't match, or the app hasn't been installed on a company yet. Double-check you copied them from the OAuth Credentials tab of the right app version, and confirm the Custom App install completed in Company Admin. ### No companies found for this service account The app is installed but scoped to zero companies. In the Developer Portal, open the app → App Settings → Install History, and confirm at least one company is marked as installed. You may need to re-run the Install Custom App flow on the intended company. ### The Developer Portal tab isn't visible in my Procore Your Procore plan may not include Developer Portal access. Contact your Procore account manager to enable it, or use one of the other connection modes (Personal OAuth or Company-wide) while you sort it out. ### Verification email never arrives Check your spam folder and confirm the admin email on your Procore company is correct. If still stuck, Procore support (_support@procore.com_) can expedite the developer-account verification. --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/procore/byo-app --- ## Bentley ProjectWise Source: https://www.alloovium.com/en/docs/guides/integrations/projectwise docs.guides.projectwise.eyebrow # docs.guides.projectwise.title docs.guides.projectwise.intro ## docs.guides.projectwise.section\_overview docs.guides.projectwise.overview\_p1 Read-only access Alloovium requests read-only access to ProjectWise. We do not modify or delete any engineering documents. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.projectwise.section\_what\_we\_access docs.guides.projectwise.what\_we\_access\_body ## docs.guides.projectwise.section\_setup docs.guides.projectwise.setup\_p1 1. 1 docs.guides.projectwise.setup\_step1\_title docs.guides.projectwise.setup\_step1\_desc 2. 2 docs.guides.projectwise.setup\_step2\_title docs.guides.projectwise.setup\_step2\_desc 3. 3 docs.guides.projectwise.setup\_step3\_title docs.guides.projectwise.setup\_step3\_desc 4. 4 docs.guides.projectwise.setup\_step4\_title docs.guides.projectwise.setup\_step4\_desc Bentley iTwin Developer Portal application settings showing redirect URI and scopes ## docs.guides.projectwise.section\_permissions docs.guides.projectwise.permissions\_body * **projectwise:read** — read documents and folder structure * **itwins:read** — list iTwin projects ## docs.guides.projectwise.section\_troubleshooting ### docs.guides.projectwise.trouble\_scopes\_title docs.guides.projectwise.trouble\_scopes\_desc ### docs.guides.projectwise.trouble\_connection\_title docs.guides.projectwise.trouble\_connection\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/projectwise --- ## Bentley ProjectWise — Bring Your Own App Source: https://www.alloovium.com/en/docs/guides/integrations/projectwise/byo-app docs.guides.projectwise\_byo.eyebrow # docs.guides.projectwise\_byo.title docs.guides.projectwise\_byo.intro ## docs.guides.projectwise\_byo.section\_why docs.guides.projectwise\_byo.why\_p1 docs.guides.projectwise\_byo.why\_p2 ## docs.guides.projectwise\_byo.section\_requirements * docs.guides.projectwise\_byo.req\_item1 * docs.guides.projectwise\_byo.req\_item2 * docs.guides.projectwise\_byo.req\_item3 docs.guides.projectwise\_byo.callout\_cloud\_only\_title docs.guides.projectwise\_byo.callout\_cloud\_only\_body ## docs.guides.projectwise\_byo.section\_create\_app 1. 1 docs.guides.projectwise\_byo.create\_step1\_title docs.guides.projectwise\_byo.create\_step1\_desc 2. 2 docs.guides.projectwise\_byo.create\_step2\_title docs.guides.projectwise\_byo.create\_step2\_desc 3. 3 docs.guides.projectwise\_byo.create\_step3\_title docs.guides.projectwise\_byo.create\_step3\_desc 4. 4 docs.guides.projectwise\_byo.create\_step4\_title docs.guides.projectwise\_byo.create\_step4\_desc 5. 5 docs.guides.projectwise\_byo.create\_step5\_title docs.guides.projectwise\_byo.create\_step5\_desc ## docs.guides.projectwise\_byo.section\_configure docs.guides.projectwise\_byo.configure\_p1 1. 1 docs.guides.projectwise\_byo.configure\_step1\_title docs.guides.projectwise\_byo.configure\_step1\_desc 2. 2 docs.guides.projectwise\_byo.configure\_step2\_title docs.guides.projectwise\_byo.configure\_step2\_desc docs.guides.projectwise\_byo.callout\_credentials\_title docs.guides.projectwise\_byo.callout\_credentials\_body ## docs.guides.projectwise\_byo.section\_connect 1. 1 docs.guides.projectwise\_byo.connect\_step1\_title docs.guides.projectwise\_byo.connect\_step1\_desc 2. 2 docs.guides.projectwise\_byo.connect\_step2\_title docs.guides.projectwise\_byo.connect\_step2\_desc 3. 3 docs.guides.projectwise\_byo.connect\_step3\_title docs.guides.projectwise\_byo.connect\_step3\_desc ## docs.guides.projectwise\_byo.section\_troubleshooting ### docs.guides.projectwise\_byo.trouble\_scopes\_title docs.guides.projectwise\_byo.trouble\_scopes\_desc ### docs.guides.projectwise\_byo.trouble\_onprem\_title docs.guides.projectwise\_byo.trouble\_onprem\_desc ### docs.guides.projectwise\_byo.trouble\_account\_title docs.guides.projectwise\_byo.trouble\_account\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/projectwise/byo-app --- ## SharePoint Integration Source: https://www.alloovium.com/en/docs/guides/integrations/sharepoint Integrations # SharePoint Sync documents from your SharePoint sites and document libraries into Alloovium. New, updated, and deleted files are picked up automatically. ## Overview The SharePoint integration connects an Alloovium project to one or more SharePoint document libraries. Once connected, Alloovium ingests files from the selected library and keeps them in sync — new files are added, changed files are re-processed, and files already up to date are skipped to save time. You can connect SharePoint using your own Microsoft account (OAuth) or with a dedicated Azure app registration set up by your IT team. The app registration path is recommended for company-wide or automated deployments. ## Connecting SharePoint Go to **Integrations → SharePoint** in your Alloovium project. Click **Add connection**. You will be prompted to give the connection a name and choose an authentication method. Alloovium Settings → Integrations: the SharePoint card in the catalog of available sources with its Connect button, alongside the other Microsoft integrations — clicking Connect launches the Microsoft sign-in window directly, no dialog to fill in first. ### Option A: Sign in with Microsoft The simplest way to connect. Alloovium redirects you to Microsoft to sign in with your work account. No Azure configuration required — permissions are granted as yourself. 1. 1 Click "Connect Microsoft" In the Add connection dialog, select the "Sign in with Microsoft" tab and click the button. A Microsoft login window will open. 2. 2 Sign in and grant consent Log in with your Microsoft 365 work account. Review the requested permissions and click Accept. 3. 3 Name the connection Back in Alloovium, enter a name for this connection (e.g. "Acme SharePoint") and save. When to use OAuth OAuth is ideal for individual users connecting their own SharePoint access. If you need to sync files that require service account permissions, use an app registration instead. ### Option B: App registration (recommended for teams) For company-wide or unattended sync, create an Azure AD app registration and provide the credentials to Alloovium. Your IT administrator will need to complete the Azure steps. #### Step 1 — Create an app registration in Azure 1. 1 Open Azure Active Directory In the Azure portal, go to Azure Active Directory → App registrations → New registration. 2. 2 Name and register the app Give the app a name (e.g. "Alloovium SharePoint"), leave the redirect URI blank for now, and click Register. 3. 3 Note the Tenant ID and Client ID On the app overview page, copy the Directory (tenant) ID and the Application (client) ID — you will paste these into Alloovium. 4. 4 Grant SharePoint API permissions Under API permissions → Add a permission → SharePoint → Application permissions, add Sites.Read.All (or Sites.FullControl.All if write access is needed). Grant admin consent. Azure portal — Overview blade of an app registration named for Alloovium, with the Essentials panel showing the Directory (tenant) ID and Application (client) ID rows and their copy buttons. #### Step 2 — Create a credential (choose one) Client secret 1. 1 Add a client secret In your app registration, go to Certificates & secrets → New client secret. Set an expiry and click Add. 2. 2 Copy the secret value immediately Azure only shows the secret value once. Copy it now — you will not be able to retrieve it again. Certificate (more secure) 1. 1 Generate or obtain a certificate Create a self-signed certificate or use one from your organisation's PKI. You need the .pfx (or .pem) private key and the certificate thumbprint. 2. 2 Upload the certificate to Azure In your app registration, go to Certificates & secrets → Certificates → Upload certificate. Upload the public certificate (.cer or .pem). 3. 3 Note the thumbprint After upload, copy the certificate thumbprint shown in the list. #### Step 3 — Enter credentials in Alloovium In the Add connection dialog, enter the **Connection Name**, **Azure Tenant ID**, and **Client ID**. Then choose your credential type: * **Client Secret** — paste the secret value from Azure. * **Certificate** — paste the certificate thumbprint and private key (PEM format). The SharePoint connect flow on the app-registration path in Settings → Integrations: Tenant ID and Client ID filled with real-looking GUIDs, the Client Secret value masked, and the connection ready to submit. #### Step 4 — Set site permissions Choose how much of your SharePoint tenant Alloovium can access: * **Full tenant access** — the app can read any site in your tenant. Convenient but requires broad admin consent. * **Selected site only** — restrict access to a single SharePoint site by entering its URL. Recommended for security-conscious deployments. `https://contoso.sharepoint.com/sites/construction`. Least-privilege recommendation Use selected site access with a client secret that has a short expiry. Rotate the secret before it expires to avoid sync interruptions. ## Selecting a library After the connection is saved, the SharePoint browser appears. Navigate to the site and document library you want to sync, then click **Sync this library**. 1. 1 Browse to your site The browser lists all SharePoint sites your connection can access. Click a site to expand it. 2. 2 Choose a document library Select the document library (e.g. "Documents", "Drawings"). You can drill into sub-folders and sync just a folder rather than the whole library. 3. 3 Start the sync Click "Sync". Alloovium queues the files for ingestion. A progress indicator shows how many files have been processed. Alloovium SharePoint browser after connecting: the site list on the left, one site expanded with a document library ticked, and the Sync button enabled bottom-right of the panel. ## How sync works Alloovium uses eTag-based change detection to avoid re-processing files that have not changed. When you trigger a sync (or re-sync), Alloovium: 1. **Fetches the file list** from SharePoint for the selected library or folder. 2. **Compares eTags** — files whose SharePoint eTag matches the eTag recorded at last sync are skipped entirely, so a library that is 90% synced will only process the remaining 10%. 3. **Queues changed or new files** for download and ingestion. Files are processed in batches to avoid overwhelming the pipeline. 4. **Updates the record** — on success, the new eTag is stored so the next sync can skip this file again. Re-syncing a partially synced library If a previous sync was interrupted, simply click Sync again. Alloovium will skip all files that were already successfully processed and only ingest what remains. File deletions Alloovium does not currently delete documents from your project when they are removed from SharePoint. Deleted files remain in Alloovium until manually removed. ## Troubleshooting ### Sync stuck at 0 files If the progress counter stays at 0, check that your app registration has the correct API permissions and that admin consent has been granted. For client secret auth, verify the secret has not expired. ### 'Insufficient privileges' error Ensure the app registration has **Sites.Read.All** (application permission, not delegated) and that an Azure Global Administrator has granted admin consent. Delegated permissions alone are not sufficient for unattended sync. ### Certificate authentication failing Verify the thumbprint matches the certificate uploaded to Azure exactly (no spaces or colons). The private key must be in PEM format (starts with `-----BEGIN RSA PRIVATE KEY-----`). --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/sharepoint --- ## Trimble Connect Source: https://www.alloovium.com/en/docs/guides/integrations/trimble-connect docs.guides.trimble\_connect.eyebrow # docs.guides.trimble\_connect.title docs.guides.trimble\_connect.intro ## docs.guides.trimble\_connect.section\_overview docs.guides.trimble\_connect.overview\_p1 Read-only access Alloovium requests read-only access to Trimble Connect. We never modify or delete project files. [Read our data handling commitments →](/en/docs/legal/data-handling) ## docs.guides.trimble\_connect.section\_what\_we\_access docs.guides.trimble\_connect.what\_we\_access\_body ## docs.guides.trimble\_connect.section\_setup docs.guides.trimble\_connect.setup\_p1 1. 1 docs.guides.trimble\_connect.setup\_step1\_title docs.guides.trimble\_connect.setup\_step1\_desc 2. 2 docs.guides.trimble\_connect.setup\_step2\_title docs.guides.trimble\_connect.setup\_step2\_desc 3. 3 docs.guides.trimble\_connect.setup\_step3\_title docs.guides.trimble\_connect.setup\_step3\_desc 4. 4 docs.guides.trimble\_connect.setup\_step4\_title docs.guides.trimble\_connect.setup\_step4\_desc 5. 5 docs.guides.trimble\_connect.setup\_step5\_title docs.guides.trimble\_connect.setup\_step5\_desc ## docs.guides.trimble\_connect.section\_permissions docs.guides.trimble\_connect.permissions\_body ## docs.guides.trimble\_connect.section\_troubleshooting ### docs.guides.trimble\_connect.trouble\_login\_title docs.guides.trimble\_connect.trouble\_login\_desc ### docs.guides.trimble\_connect.trouble\_missing\_projects\_title docs.guides.trimble\_connect.trouble\_missing\_projects\_desc ### docs.guides.trimble\_connect.trouble\_redirect\_title docs.guides.trimble\_connect.trouble\_redirect\_desc --- Canonical page: https://www.alloovium.com/en/docs/guides/integrations/trimble-connect --- ## Lessons Source: https://www.alloovium.com/en/docs/guides/lessons DOCUMENTS & PROJECTS # Lessons Every builder has a job that bit them — the fixing spec that corroded, the clause that cost a claim, the sequence that stole three weeks. Lessons turn those scars into a working memory: captured once, and resurfaced automatically the next time the same trap appears. ## Overview Lessons live in your company workspace, under [Organization](/en/docs/guides/organization). Each one records what happened, why it mattered, and what to do differently — tied to the documents and the project where it was learned. Unlike a lessons-learned PowerPoint that nobody reopens, these are indexed alongside your documents: the assistant knows them, reviews check against them, and new work that rhymes with an old mistake gets flagged. The point is the resurface, not the archive A lesson’s value is realised the day it interrupts a repeat — a new spec that names the product that failed last time, a subcontract missing the clause that saved you before. Capture is cheap; the resurfacing is the feature. ## Capturing a lesson From the Lessons page, from chat, or as a byproduct of the work. Add lessons directly on the Lessons page, or tell the assistant — “save a lesson: never accept carbon-steel fixings against stainless balustrade, Northern Depot corroded in a year and the warranty was refused.” A lesson can cite the documents that prove it, so the evidence travels with the rule of thumb. Lessons also emerge from the work itself: a dispute post-mortem, a diary entry that records how a problem was solved, a review that caught something expensive. When Alloovium spots a moment worth keeping, it proposes the lesson — you approve the wording before anything joins company memory. ## How lessons come back Matched against new documents and decisions, with the receipts attached. | Where | How a lesson surfaces | | ------------------ | --------------------------------------------------------------------------------------------------------------- | | Arriving documents | A new spec, drawing, or subcontract that matches a lesson gets flagged at review — before the order goes out. | | The assistant | Answers that touch a lesson cite it alongside the documents, so the past job is part of the answer. | | Watch tags | Phrases cut from hard-won lessons (a failed product, a risky term) are checked against every arriving revision. | Video · \~2 min A lesson that came back Capture a lesson from a past defect — carbon-steel fixings against stainless balustrade — then watch it resurface when a new spec names the same product: the flag at review, the lesson cited in the assistant's answer, and the RFI that heads off the repeat. ## What a good lesson looks like Specific enough to match; short enough to read. The lessons that pay for themselves name the trigger, the consequence, and the rule. A strong pattern: _situation → what went wrong → what we do now_. “Galvanic corrosion: carbon-steel fixings against stainless balustrade at Northern Depot corroded inside a year, warranty refused — always spec 316 fixings with stainless balustrade.” Vague morals (“communicate earlier”) do not match anything; named products, clauses, and conditions do. ## Availability Capture is live; automatic resurfacing is rolling out. The Lessons library — capturing, browsing, and asking the assistant about lessons — is available in the company workspace. Automatic resurfacing against arriving documents is newer and enabled per workspace as it rolls out; where it is not yet on, lessons still answer through the assistant and reviews. --- Canonical page: https://www.alloovium.com/en/docs/guides/lessons --- ## Lookahead Source: https://www.alloovium.com/en/docs/guides/lookahead AUTOMATION # Lookahead A rolling view of the next few weeks that surfaces schedule risk while there is still time to act — built from the documents, meetings, and reports your routines already read. ## Overview Lookahead is the near-term slice of the programme: the activities due over the next three to six weeks, when they are planned, how they are tracking, and what threatens to slip. Its purpose is a single outcome — no surprises on schedule. Rather than wait for a monthly report to reveal that a milestone has moved, lookahead brings the movement forward while you can still do something about it. Lookahead is deliberately not a scheduling tool. It does not replace your programme in MS Project or Primavera, and it does not attempt to recompute a critical path from scratch. It is a thin, current home for near-term activities so the [Today](/en/docs/guides/today) surface can detect and raise schedule risk in plain language. A byproduct of work you already do The value of lookahead is that it is largely produced for you. As routines read meeting notes, daily reports, and connected data, activities and their dates flow into the lookahead — so keeping a live schedule view costs little extra effort. ## What it surfaces Near-term activities, their progress, and the risk of a slip. Each activity in the lookahead carries a small set of fields kept intentionally simple, so the view stays fast to read and easy to trust. | Field | Meaning | | ------------- | ----------------------------------------------------------------------------------------- | | Activity | The name and a short description of the piece of work. | | Planned dates | The start and end originally scheduled for the activity. | | Actual dates | The real start and end as work progresses. | | Status | Planned, in progress, completed, or blocked. | | Progress | Roughly how far along the activity is, as a percentage. | | Source | Where the activity came from — a meeting, a daily report, a connected tool, or an import. | From these, lookahead highlights what matters to a project manager: activities that are behind, activities that are blocked, and the knock-on effect when one late activity holds up another that was due next. That downstream ripple — where a slip in one place puts a near-term milestone at risk — is the part hardest to see by eye and the part lookahead is built to catch. A lookahead view of the next few weeks — activities in rows with planned bars, actual progress, and a highlighted item flagged as at risk of slipping. ## Where the data comes from Lookahead is fed, not typed. It draws on the sources you are already working in. Activities arrive in the lookahead from several places, each tagged with where it came from so you can trace any date back to its source: * → **Meetings** — programme discussion captured in meeting notes is read for activities and dates. * → **Daily and site reports** — on-site diary entries update what actually started, finished, or stalled. * → **Connected tools and imports** — schedule data brought in from a connected system or a spreadsheet import. Because these are the same sources your routines read, a routine run is often what quietly keeps the lookahead current — extracting the programme change out of a report and applying it without a separate step from you. ## Generating a lookahead workbook Produce a shareable Gantt workbook on demand from a routine. A routine can produce a lookahead as a deliverable using the `GENERATE_LOOKAHEAD` block. It reads schedule information out of the documents you point it at and outputs a workbook with a Gantt view you can download and share — useful for a weekly lookahead meeting or a pack sent to the team. ### Options The block can be tuned to how your project runs: * → **Window** — how many weeks ahead to cover, typically three to six. * → **Grouping** — organise activities by area, by trade, or by work breakdown. * → **Detail** — whether to show completed items, the critical path, and constraints, and whether to count working days or calendar days. To learn how to place this block in a routine and set its trigger, see [Routines](/en/docs/guides/routines). The same routine engine that fills a [template](/en/docs/guides/template-filling) produces the lookahead workbook. ## Keeping it current An optional assist that applies confident changes and flags the rest. Where it is enabled, an automatic assist can keep the lookahead in step with recent site updates and project documents. It reads for changes — dates that moved, activities that became blocked, milestones that were reached — applies the high-confidence ones, and surfaces the downstream risk they create. You can preview the changes before they are applied: a dry run shows what would move, what risk it cascades into, and why, so nothing shifts in the schedule without your sight of it. The automatic assist is gated The auto-update capability is off by default and enabled per workspace. Reading and using the lookahead does not require it — without it, activities still flow in from your routines, meetings, and imports, and you update the schedule yourself. ## Availability Lookahead is maturing. Some parts are further along than others. The near-term lookahead view and the ability to generate a workbook are available as part of the routine engine. The automatic keeping-current assist is newer and rolled out selectively. As with routines generally, if a capability described here is not visible in your account, it has not yet been turned on for your workspace. Lookahead works hand in hand with [Routines](/en/docs/guides/routines) and reports its risk signals into [Today](/en/docs/guides/today), where you see what needs attention each morning. --- Canonical page: https://www.alloovium.com/en/docs/guides/lookahead --- ## Meetings Source: https://www.alloovium.com/en/docs/guides/meetings DAILY WORKFLOW # Meetings Bring a meeting into Alloovium, get a clean transcript and a project-manager summary, and review the actions and decisions it contains before they turn into work you have to track. ## Overview A great deal of what governs a construction job is decided in meetings — the weekly lookahead, the client catch-up, the subcontractor coordination call. Alloovium captures the meeting, transcribes it, and reads it the way a project manager would: pulling out the actions, the decisions, and the things that need to become work. Nothing is dispatched automatically. Alloovium prepares a reviewable list, and you decide what becomes a task, a variation or a recorded decision. Items you approve then flow into your day on [Today](/en/docs/guides/today). Video · \~2 min One meeting, start to finish A recording is uploaded, moves through transcription to Complete, and its extracted items are reviewed and dispatched — an action, a decision, and a variation each landing where they belong. ## Capturing a meeting There are a few ways to get a meeting into Alloovium, depending on how the meeting happens. In every case the result is the same: a meeting record that will carry the transcript, summary and extracted items. 1. 1 Upload a recording Upload a pre-recorded audio or video file. Alloovium extracts the audio where needed and transcribes it. 2. 2 Send a notetaker to a call Where the Recall.ai notetaker is configured, Alloovium can send a bot to a Zoom, Teams or Meet link to record the call. 3. 3 Join from your calendar Where calendar auto-join is configured, you can connect a Google or Microsoft calendar and have the notetaker attend selected events. Attendees are recorded against the meeting, and recurring meetings can be linked as a series so Alloovium can note what has changed since the previous session. Recording and notetaker availability The notetaker and calendar auto-join depend on integrations that are configured for your workspace, and some capture methods may not be enabled in every environment. See [Availability](#availability) below. ## Transcription and summary Once a meeting is captured, transcription runs automatically. Alloovium produces a full transcript, then cleans up the kind of errors speech recognition makes with construction terminology so the text reads correctly. From the transcript, Alloovium writes a concise project-manager summary — grounded in what was actually said, not invented — and identifies the attendees. The transcript, the summary and the recording (where one exists) are all available on the meeting. Grounded, not embellished The summary is instructed to stay faithful to the transcript. If a decision was not reached in the meeting, it will not appear as one. ## Meeting status A meeting moves through a sequence of states as it is captured and processed. | Status | What it means | | ------------ | ----------------------------------------------------------------- | | Uploaded | The meeting record exists and is awaiting its recording or file. | | Scheduled | A notetaker is set to join an upcoming call. | | Recording | The notetaker is in the call, or a live recording is in progress. | | Transcribing | The audio is being converted to text. | | Processing | The summary and extracted items are being prepared. | | Complete | The meeting is ready for review. | | Failed | Transcription or processing did not complete. | | Canceled | The notetaker was stopped before it recorded. | ## What Alloovium extracts Alloovium does not stop at a summary. It reads the transcript for the things a project manager would act on and files each one as an extracted item, tagged by kind, with a verbatim quote from the transcript so you can see exactly where it came from. | Kind | What it captures | | ----------- | -------------------------------------------- | | Action item | Something you need to do. | | Team task | Something to assign to a member of the team. | | Variation | A change to scope, cost or programme. | | Programme | An update to the schedule or lookahead. | | Decision | A decision the meeting reached. | | Risk | A threat or blocker that was raised. | | Lesson | Knowledge worth keeping for future work. | Each extracted item can carry a suggested assignee named in the discussion and a best-guess project, so you can triage the list quickly. A variation surfaced here connects to the same drafting flow described in [Variations](/en/docs/guides/variations). ## Review and dispatch Extracted items sit in a review queue on the meeting until you decide what to do with them. Nothing is routed on until you approve it. 1. 1 Review Open the meeting and read each extracted item alongside the quote it came from. 2. 2 Adjust or dismiss Correct the kind or assignee where needed, or dismiss an item that is not actionable. 3. 3 Confirm the project For items that belong to a job — actions, team tasks, variations and programme updates — confirm which project they apply to. 4. 4 Dispatch Approve the items you want to keep. Alloovium routes each one to its destination. Dispatched items land where they belong. Actions become your own tasks and appear on [Today](/en/docs/guides/today); team tasks are assigned; variations open as drafts; decisions, risks and lessons are recorded at the workspace level. Because decisions, risks and lessons are workspace-wide, they do not require a project to be chosen. ## Availability Meetings capabilities depend on how your workspace is configured. Automatic transcription and summarisation run whenever a meeting has a recording. The notetaker for live calls requires the Recall.ai integration to be configured, and calendar auto-join requires a connected Google or Microsoft calendar. Some capture methods may be gated Where an integration is not configured — for example, if the notetaker is not enabled for your workspace — the related option will be unavailable. If a capture method you expect is missing, check with your administrator on how it has been set up for your environment. Whatever the capture method, the review-before-dispatch step is the same, so items only enter your workflow once you have confirmed them. --- Canonical page: https://www.alloovium.com/en/docs/guides/meetings --- ## Mobile — Voice-first capture Source: https://www.alloovium.com/en/docs/guides/mobile Mobile # Mobile capture The Alloovium mobile app is a voice-first way to capture what happens on site. You pick up your phone, say what happened, and the recording becomes a draft in Alloovium through the same backend as the web app — the phone is for capture, the desktop is for review and export. Mobile capture screen with the record button, project selector and live transcript ## Overview The most reliable moment to record a site update is the moment it happens, standing in front of the work. The mobile app is built for that: a large record button, a project to file the note against, and an optional photo. It does not try to reproduce the whole product on a phone. Its job is to get an accurate account of the day off the site and into a draft you can finish later. When you record, the audio is transcribed on the server and the transcript is turned into a document draft in the background. That draft lands against the project and appears in the web app, where you edit, review and export it. In preview The mobile app is an early, voice-first companion (v0.1) and is available on request rather than as a general release. Some capabilities are still being finished. If you would like to try it, ask your Alloovium contact. ## Signing in You sign in with the same Alloovium account you use on the web — the app uses the same identity, so your projects and documents are the ones you already have. After signing in once, the session is held securely on the device, and you can optionally require Face ID, Touch ID or a passcode to reopen the app. ## Capturing a note Capture is the home of the app. You choose the project the note belongs to, optionally pick a template so the draft takes a familiar shape, and record. 1. 1 Pick the project Select the project this note is about so the resulting draft is filed in the right place. 2. 2 Record what happened Tap the microphone and speak naturally — for example, a site diary entry describing the pour, a delay and the inspection result. 3. 3 Review the transcript The recording is transcribed and shown on screen so you can confirm it captured what you said. 4. 4 Attach photos and draft Optionally add site photos, then draft. The transcript, and any photos as evidence pages, become a document in the background. If you only want a record of what you said, a log-only mode saves the transcript without drafting a document. The same spoken-note idea drives the web experience described in [Voice Input](/en/docs/guides/ai-assistant/voice-input). ## Working offline Site connectivity is rarely dependable, so the app is built to tolerate it. If a recording cannot be transcribed or drafted because the network is unavailable, the app queues it on the device and retries automatically when connectivity returns. You do not have to remember to resend anything. The Activity screen shows the state of each capture so you always know where a note stands. | State | Meaning | | ------------ | ----------------------------------------------------------- | | queued | Waiting to send; will retry when the device is back online. | | transcribing | Audio has been sent and is being turned into text. | | drafting | The transcript is being turned into a document. | | drafted | The draft is ready and will appear in the web app. | | failed | Something went wrong; the capture can be retried. | ## Where the draft goes A finished capture becomes a document draft on the project, created through the same document-generation path the web app uses. Once it is ready you receive a notification on the phone, and the draft is waiting in the project’s documents on the web, where you can edit and export it as usual. The mobile app draws on the same backend intelligence as everything else in Alloovium. It is deliberately narrow — capture on the phone, finish on the desktop — so that recording an update on site is a matter of seconds rather than a task you defer until you are back at a computer. Because it is in preview, expect the surface to keep changing as it is completed. Preview limitations As an early build, the app focuses on capture and drafting. Editing, review and export happen on the web, and some convenience features are still being added. Availability and behaviour may change as the app matures. --- Canonical page: https://www.alloovium.com/en/docs/guides/mobile --- ## Monthly Report Source: https://www.alloovium.com/en/docs/guides/monthly-report AUTOMATION # Monthly Report The month-end report is the most predictable document on a project — and the one that eats a weekend. Alloovium drafts it from the records the month already produced, in your template, with every claim cited, and hands it to you for review instead of a blank page. ## Overview Everything a monthly report says already exists somewhere: diaries, meeting minutes, programme updates, variation registers, cost records, photos. The monthly report engine reads that month’s slice of the project and assembles it into your report format — the same sections, the same voice, the numbers reconciled against the records they came from. Cited like everything else Each statement in the draft traces to its source — the diary entry behind a delay paragraph, the register row behind a variation total. Reviewing the report is checking citations, not reconstructing the month from memory. ## What it assembles The month’s records, organised into the sections your report already uses. | Section | Drawn from | | -------------------------- | -------------------------------------------------------------------------- | | Progress & works completed | Site diaries, meeting minutes, and activity records for the period. | | Programme | Schedule updates and the lookahead — what moved, and what it puts at risk. | | Variations & commercial | The variation register: raised, approved, and pending this month. | | Risks & issues | Open findings, delays, and flagged conditions from the month’s records. | | Photos & attachments | Site photos and referenced documents from the period. | A drafted monthly report open for review in the project's own template — a progress paragraph highlighted, with the citation panel beside it showing the site diary entries the paragraph was built from. ## How it runs On demand, or as a routine that lands the draft before month-end. Generate a report on demand for any period, or set it up as a [routine](/en/docs/guides/routines) so the draft is waiting in the last week of each month. The engine uses your own report template — the one your client already expects — via the same machinery as [template filling](/en/docs/guides/template-filling), so the output needs formatting from nobody. Video · \~2 min Month-end without the weekend Open the draft a monthly-report routine left waiting, check a cited figure against the variation register, adjust a progress paragraph, and approve the report at the review gate. ## Review before it ships The draft proposes; you approve. Like every generated document, the monthly report stops at a human review gate. You read the draft with its citations, adjust the judgement calls — tone, emphasis, what to say about the tricky bits — and nothing leaves the building until you say so. The aim is not to remove you from the report; it is to remove the weekend from it. --- Canonical page: https://www.alloovium.com/en/docs/guides/monthly-report --- ## Company Workspace Source: https://www.alloovium.com/en/docs/guides/organization DOCUMENTS & PROJECTS # Company Workspace Projects come and go; the company accumulates. The Organization workspace holds what outlives any single job — your shared library, the companies and people you work with, your templates and procedures, and the lessons the jobs taught you. ## Overview Where a project workspace answers “what is true on this job?”, the Organization workspace answers “what is true for us?”. It is the home for documents that belong to the whole company, the master records other surfaces draw on, and the shared knowledge every new project should start with. Everything here participates in the assistant’s answers, permission-filtered like all retrieval. | Section | What lives there | | ------------- | ------------------------------------------------------------------------------------------------------ | | Company files | The shared document library — templates, references, and every project record in one searchable place. | | Contacts | The companies and people you work with — clients, subs, suppliers, consultants. | | Playbook | Your reusable templates and standard procedures. | | Lessons | Company memory — what past jobs taught you, resurfaced when it matters. | | People | Your team — members, roles, and groups. | | Data quality | A health check on your master data — gaps, duplicates, and stale records. | ## Company files Every document your company works from, in one place. The library shows your whole corpus — company-level files like policies, standard details and reference specs, alongside every project’s records. Organise with folders, filter by source (uploads, SharePoint, Procore, and other [integrations](/en/docs/guides/integrations)) or by classification, and open any file in the preview with its extraction and version history. Documents not attached to a project live here as company files; attach them to projects at any time without re-uploading. The company library mid-filter: folder tree on the left, document rows with classification badges on the right, and the source filter open showing Uploads, SharePoint, and Procore options. ## Contacts One directory for the companies and people around your projects. Contacts is the master record of who you work with — client organisations, subcontractors, suppliers, consultants, and the individuals inside them. Projects reference these records (a project’s client is a company here, not a typed-in name), so the same party looks the same on every job, and questions like “which projects has this sub worked on?” have a real answer. ## Playbook The documents you reuse: templates and procedures. The Playbook holds your standard forms and procedures — the templates that [template filling](/en/docs/guides/template-filling) and [document generation](/en/docs/guides/reports) draw on, and the written procedures your routines and reviews enforce. Keep the canonical version here; every project starts from it rather than from whatever was on someone’s desktop. ## Lessons Company memory, on the record. Lessons capture what jobs taught you and bring it back when a new document or decision walks into the same trap. They are important enough to have [their own guide](/en/docs/guides/lessons). ## People Members, roles, and groups. People is where your team is managed — who has access, what role they hold, and which groups they belong to. Roles and permissions are covered in [Permissions & Roles](/en/docs/guides/collaboration/permissions), and team setup in [Admin & Teams](/en/docs/guides/admin). ## Data quality A health dashboard for the records everything else relies on. Master data decays: contacts duplicate, projects miss numbers, records go stale. The data quality dashboard measures it — surfacing gaps and duplicates with a path to fix each one — so the answers built on this data stay trustworthy. Why it matters to answers The assistant is only as sharp as the records underneath it. Ten minutes on the data quality queue pays back across every question anyone asks afterwards. --- Canonical page: https://www.alloovium.com/en/docs/guides/organization --- ## Working with Trades Source: https://www.alloovium.com/en/docs/guides/portal Document Control # Working with your trades The document loop with subcontractors, trades and consultants runs through Exchange: the packages you issue for review, the documents you request back, and the outcome of every round-trip, tracked in one place instead of scattered across email threads. ## Overview Most document exchange with subcontractors happens over email and shared drives, where nothing is tracked and the status of any given file is a matter of memory. [Exchange](/en/docs/guides/exchange) replaces that: each package you send or request is a tracked loop with a state, a version history, and an activity ledger recording who did what and when. The other party doesn't need to be an Alloovium customer. You address a package to an email; the recipient gets access to the exchanged copy only — never your project — and everything they do to it is recorded on your side of the loop. ## Sending and requesting Exchange runs in both directions from the same workspace. **Send a document** issues a package to a trade — an amendment for execution, a drawing set for pricing. **Request documents**asks a trade to supply something — their shop drawings, mix designs, or insurances — and gives them a place to put it that isn't an email attachment. You can also work conversationally: ask the assistant to `send the amendment to Lumino for signature` and it stages the Exchange package as a draft for your confirmation — the same review gate as everything else the assistant does. This is how project mobilisation runs without the spreadsheet: from every subcontractor, the same rule-based list — the signed contract back, licences, insurances — raised as requests with a state each, chased automatically, and visible on one board instead of forty follow-up emails. Video · \~2.5 min Mobilisation without the spreadsheet A just-won job: per-subcontractor document requests going out through Exchange, a sub uploading against them from a clean guest link, an expired certificate sent back revise-and-resubmit, the board's one red row, and the chase routine nagging so you don't. ## The review loop When a trade submits their document back, the package lands in your queue marked `Ready to review`. You accept it or send it back with a note, and the outcome is recorded with the submittal vocabulary the industry already uses: | Review outcome | Meaning | | ----------------- | ----------------------------------------------------------- | | approved | Accepted as submitted — becomes the current version. | | approved as noted | Accepted with comments; no resubmission required. | | revise & resubmit | Sent back with a note; the trade revises and submits again. | | rejected | Not accepted; the round-trip restarts. | Every round-trip accrues to the package's revision history — who issued each version, who reviewed it, the notes exchanged, and the read and acknowledgement events along the way. The full cycle is shown on the [Exchange overview](/en/docs/guides/exchange), and when a returned document needs line-by-line scrutiny, [paragraph-level diffs](/en/docs/guides/exchange/clause-diffs) show exactly what changed between revisions. ## Transmittals A transmittal is a tracked package of project documents issued to another party — a consultant, subcontractor or client — with acknowledgement tracking, so you can see who has received a package and who is still outstanding rather than assuming a document landed because you attached it to an email. Issue one conversationally — `issue a transmittal to the structural engineer for the latest slab drawings` — and the assistant assembles the package for your confirmation. [Routines](/en/docs/guides/routines) can also issue transmittals automatically, for example sending the updated drawing set whenever a new revision arrives. Ask the assistant “did the architect acknowledge the drawings?” to check status without leaving chat. ## The board The Exchange workspace under `Between companies` is the status view for the whole loop: the `Documents` and `Requests` tabs split outbound from inbound, each row carries its state — shared, ready to review, changes requested, approved — with the counterparty organisation and dates alongside, and the `Needs you`count keeps score of what's waiting on your review. Video · \~2 min A submission you can see moving One trade package end to end through Exchange: send the package for review, the subcontractor submits their version, a revise-and-resubmit with a note, the corrected resubmission, and the approval — then the whole history read back off the revision timeline. Availability Exchange is enabled per workspace. If you don't see Exchange under Between companies in the sidebar, ask your administrator or your Alloovium contact. --- Canonical page: https://www.alloovium.com/en/docs/guides/portal --- ## Document Hub Source: https://www.alloovium.com/en/docs/guides/projects Document Hub # Document Hub Upload, organise, and manage your documents securely in projects. Projects are the foundation of everything in Alloovium. Video · \~3 min From upload to answers Follow a batch of project documents from drag-and-drop upload, through automatic OCR, classification, and indexing, into Auto-sort's per-file folder proposals — and finish by asking the assistant a question the freshly indexed documents answer. ## Overview The Document Hub is the central repository for all files in a project. Every document you upload is automatically processed — text is extracted via OCR if needed, chunked into semantically meaningful segments, and embedded for fast AI retrieval. Once a document is processed, it is immediately available for querying in the AI Assistant, template filling, document generation, and use in analysis tools like Change Impact. ## Uploading Documents You can upload documents to a project in several ways: by dragging files directly onto the upload area, by clicking the "Upload" button and selecting files from your computer, or by syncing from a connected cloud storage provider (SharePoint, Google Drive, Procore). 1. 1 Open the project Navigate to Projects in the sidebar and open the project you want to upload into. 2. 2 Upload your files Drag and drop files onto the upload area, or click "Upload" and select files from your computer. You can upload multiple files at once. 3. 3 Wait for processing Alloovium processes each document automatically. You will see a progress indicator next to each file. Processing typically takes 30–120 seconds depending on file size. 4. 4 Start querying Once processing is complete, the document status changes to "Ready". You can now query it via the AI Assistant. Batch uploads You can upload up to 50 files at once. Large batches are queued and processed in parallel — you do not need to wait for one batch to finish before uploading another. ### Supported Formats | Format | Extension | Notes | | ---------- | ----------------- | ---------------------------------------------------- | | PDF | .pdf | Includes scanned PDFs — OCR is applied automatically | | Word | .docx | Tracked changes and comments are included | | Excel | .xlsx, .xls | All sheets are processed; tables extracted | | PowerPoint | .pptx | Slide text and speaker notes extracted | | Images | .png, .jpg, .tiff | OCR applied; useful for scanned drawings | | Plain text | .txt, .csv | Processed directly; no OCR required | ## Managing Projects Documents are always organised inside a project. A project is a workspace that holds a set of related documents — for example, all documents related to a specific contract or tender. You can create as many projects as you need and invite different team members to each one. Within a project, you can create **folders** to further organise your documents. Folders are optional — all documents in a project are available for querying regardless of which folder they are in. Storage limits Storage limits depend on your subscription plan. Contact support if you need to increase your storage quota. ## AI Organizer & Auto-Sort Got a loose batch of mixed files? Upload them into the project, then let Auto-sort file them: it reads each document and proposes a destination folder, with a confidence rating and an editable picker, before anything moves. 1. 1 Upload the loose files Drag the whole batch into the project's Files tab. Upload the full intake together — more context means better filing decisions. 2. 2 Run Auto sort Click "Auto sort" in the Files toolbar. Alloovium reads filenames and extracted content to classify each document and pick its folder. 3. 3 Review, adjust, apply Every file shows its suggested folder and a confidence rating. Change any destination from the picker, then apply — nothing moves until you do. | Loose file type | Suggested folder | Typical examples | | ---------------------------- | ------------------------------------------------------------ | --------------------------------------------------------------------- | | Specifications and standards | 03 Specifications & Technical / 03.01 Specifications | Specifications, scopes, standards, technical requirements, manuals | | Drawings and issued sheets | 02 Drawings & Models / 02.01 Sheets & Issued Drawings | Drawings, plans, elevations, details, issued sketches | | Registers and logs | 06 Controls & Schedules / 06.02 Registers & Logs | Registers, trackers, logs, schedules, look-aheads | | Memos and correspondence | 07 Meetings & Correspondence / 07.02 Correspondence & Emails | Memos, letters, general correspondence, email exports | | Instructions and notices | 08 Changes & Instructions / 08.02 Instructions & Notices | Site instructions, architect instructions, notices, formal directions | Use the whole batch Auto-sort works best when you upload the full unsorted intake together. That gives it enough context to classify each file and choose the right folder paths. Review safeguards Low-confidence files are marked for review instead of being silently moved into a folder you did not expect. ## Document Types Alloovium automatically classifies uploaded documents into categories to help with organisation and to optimise analysis. Common types include contracts, specifications, drawings, reports, and correspondence. You can also assign a custom type to any document. | Type | Description | Best used for | | -------------- | -------------------------------------- | ---------------------------------------------- | | Contract | Legal agreements and subcontracts | Payment terms, obligations, liquidated damages | | Specification | Technical and performance specs | Scope verification, compliance checking | | Drawing | Engineering and architectural drawings | Clash detection, revision tracking | | Report | Site reports, inspection reports | Issue tracking, defect management | | Correspondence | Emails, RFIs, instructions | Notice tracking, variation history | ## Processing Pipeline Understanding how Alloovium processes your documents helps you get the most out of the platform. Each document goes through the following stages: 1. **Text extraction** — Alloovium reads the raw text from each page. For scanned PDFs and images, GPU-accelerated OCR is applied automatically. 2. **Layout analysis** — The document structure is analysed: headings, tables, figures, and paragraph blocks are identified and tagged. 3. **Chunking** — The document is split into semantically meaningful segments (typically 200–500 tokens each) that are small enough for precise retrieval. 4. **Embedding** — Each chunk is converted into a vector embedding using a high-dimensional model trained on technical and legal language. 5. **Indexing** — Embeddings are stored in a vector index for fast similarity search, enabling sub-second retrieval during AI queries. GPU extraction Alloovium uses GPU-accelerated document extraction for complex PDFs and engineering drawings. This means higher accuracy on scanned documents, tables with merged cells, and multi-column layouts. --- Canonical page: https://www.alloovium.com/en/docs/guides/projects --- ## Document Types Source: https://www.alloovium.com/en/docs/guides/projects/document-types Projects # Document Types Alloovium automatically classifies documents to optimise analysis. Learn what types are supported and how the processing pipeline works. ## Document Types Alloovium automatically classifies uploaded documents into categories to help with organisation and to optimise analysis. Common types include contracts, specifications, drawings, reports, and correspondence. You can also assign a custom type to any document. | Type | Description | Best used for | | -------------------------- | ------------------------------------------------ | ---------------------------------------------- | | Contract | Legal agreements and subcontracts | Payment terms, obligations, liquidated damages | | Specification | Technical and performance specs | Scope verification, compliance checking | | Drawing | Engineering and architectural drawings | Discipline and revision tracking | | RFI / RFI Response | Requests for information and their answers | Open-item tracking, notice history | | Submittal | Shop drawings, samples, product data | Approval status, review turnaround | | ITP | Inspection & test plans | Hold points, quality verification | | SWMS / JSA | Safe work method statements, job safety analyses | Safety compliance, high-risk work | | Register / Log | Schedules, registers, and logs | Variation, defect, and notice tracking | | Report | Site and inspection reports | Issue tracking, defect management | | Correspondence | Emails, letters, instructions | Notice tracking, variation history | | Meeting Minutes | Records of site and project meetings | Actions, decisions, attendance | | Permit / Transmittal | Permits and document transmittals | Approvals, issue records | | Environmental (EMP / CEMP) | Environmental management plans | Compliance obligations, conditions | Around twenty categories are recognised in total, plus a general fallback. You can override the type on any document. ## Processing Pipeline Understanding how Alloovium processes your documents helps you get the most out of the platform. Each document goes through the following stages: 1. **Text extraction** — Alloovium reads the raw text from each page. For scanned PDFs and images, GPU-accelerated OCR is applied automatically. 2. **Layout analysis** — The document structure is analysed: headings, tables, figures, and paragraph blocks are identified and tagged. 3. **Chunking** — The document is split into semantically meaningful segments (typically 200–500 tokens each) that are small enough for precise retrieval. 4. **Embedding** — Each chunk is converted into a vector embedding using a high-dimensional model trained on technical and legal language. 5. **Indexing** — Embeddings are stored in a vector index for fast similarity search, enabling sub-second retrieval during AI queries. GPU extraction Alloovium uses GPU-accelerated document extraction for complex PDFs and engineering drawings. This means higher accuracy on scanned documents, tables with merged cells, and multi-column layouts. --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/document-types --- ## Drawing Intelligence Source: https://www.alloovium.com/en/docs/guides/projects/drawing-intelligence Projects # Drawing Intelligence Alloovium automatically extracts structured metadata from construction drawings on upload — no configuration required. Sheet numbers, revisions, disciplines, and title block data are indexed for search and AI queries. ## Overview When a drawing PDF is uploaded to a project, Alloovium reads the title block and extracts structured metadata — discipline, sheet number, revision, title, date, and scale. This happens automatically as part of the standard processing pipeline. Once indexed, drawings are queryable by any of these fields in both the project search and the AI assistant. The assistant can answer questions like "what is the latest revision of A-101?" or "list all structural drawings issued after June" without any manual tagging. No routine required Drawing intelligence is not a routine step. It runs automatically on every drawing uploaded to a project — there is nothing to configure. A drawing open in the preview with its extracted title-block metadata in the details panel: sheet number A-101, revision C, discipline Architectural, and date shown as structured fields beside the title block they were read from. Video · \~2 min Drawings that know what they are Upload a small drawing set and watch title-block extraction fill in sheet numbers, revisions, and disciplines automatically — then ask the assistant for the latest revision of a sheet and jump straight to it. ## What Gets Extracted Alloovium reads the title block from each drawing page and extracts the following fields where present: | Field | Example | Notes | | ----------------------------- | -------------------------------------------- | ------------------------------------------------------------------------- | | Sheet number | A-101, S-03, M-204 | Indexed for search; used to group revisions of the same sheet | | Discipline | Architectural, Structural, Mechanical, Civil | Indexed for search; inferred from the sheet number prefix if not explicit | | Drawing title | Ground Floor Plan, Column Schedule | Indexed for search; from the title block title field | | Revision | Rev C, Rev 3, P4 | Read from the title block; used to link successive revisions of a sheet | | Date, scale, and other fields | 14 Apr 2025, 1:100 | Read where present, but not currently exposed as separate search filters | Scanned drawings Scanned drawing PDFs are processed using GPU-accelerated OCR before title block extraction. Accuracy is high for standard title blocks, but may be lower for handwritten annotations or non-standard layouts. ## Searching Drawings Drawing metadata is searchable directly in the project and through the AI assistant. Both support natural language queries using any extracted field. 1. 1 Project search Type a sheet number (e.g. "A-101"), discipline ("structural"), or keyword from a drawing title. Results are filtered to matching drawings with revision context shown. 2. 2 AI assistant Ask in plain language — "What does drawing S-04 show?", "Are there any civil drawings for the eastern retaining wall?", or "Show me the latest revision of the floor plans." 3. 3 Filter by discipline In the project drawing view, use the discipline filter to narrow results to Architectural, Structural, Mechanical, Electrical, or Civil drawings. Best results For best AI results, upload full drawing sets rather than individual sheets. The assistant uses cross-sheet context — for example, it can correlate a structural column grid from one sheet with a detail on another. ## Limitations | Limitation | Detail | | ------------------------- | ------------------------------------------------------------------------------------------ | | Non-standard title blocks | Extraction accuracy drops on highly customised or handwritten title blocks | | Raster-only PDFs | Very low-resolution scans (below \~150 DPI) may produce lower-accuracy OCR results | | Multi-page sheets | Each page is processed independently; multi-page assemblies are treated as separate sheets | --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/drawing-intelligence --- ## Managing Projects Source: https://www.alloovium.com/en/docs/guides/projects/projects Projects # Managing Projects Projects are the primary unit of organisation in Alloovium. Each project holds a set of related documents with its own team and settings. ## Managing Projects Documents are always organised inside a project. A project is a workspace that holds a set of related documents — for example, all documents related to a specific contract or tender. You can create as many projects as you need and invite different team members to each one. Within a project, you can create **folders** to further organise your documents. Folders are optional — all documents in a project are available for querying regardless of which folder they are in. Storage limits Storage limits depend on your subscription plan. Contact support if you need to increase your storage quota. --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/projects --- ## Registers Source: https://www.alloovium.com/en/docs/guides/projects/registers DOCUMENTS & PROJECTS # Registers Every project carries a Registers band in the sidebar — the records you check, kept separate from the feed of things that need you and the tools you run. Between them the registers cover the state of the job: what changed and what it is worth, who is holding which drawing revision, what the contract obliges you to do, where the time went, what has been claimed and paid, and what security someone is still holding against you. ## Overview A register answers one question: _what is the state of X, right now?_ It is a record you consult, sort, export and hand to other people — a subcontractor, a certifier, an adjudicator. It is not a to-do list, and it does not chase you. Registers are not summaries generated on demand. Each one is a live view over the underlying records — document lineages, extracted obligations, delay events, commercial changes, claims and security instruments — so a row in a register is always backed by the documents and events that put it there. Open the row and you land on the thing itself. Two rules that hold for every register **Rows are backed, never invented.** A register never composes a figure or a status out of thin air; it reads the typed records underneath and shows you what they say. **Retrieval is permission-filtered to the person asking.** Two people can open the same register on the same project and correctly see different rows — the register shows each of them the records they are entitled to. See [Permissions](/en/docs/guides/collaboration/permissions). A project's left sidebar with the REGISTERS band open — Commercial, Distribution, Obligations, Delays & EOT, Claims and Security stacked between the Work and Knowledge bands, with the Distribution row active. ## The registers band The band sits in the project sidebar and holds the following registers. Each one is its own page, with its own filters and its own row actions. | Register | What it records | | ------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Commercial | The change record for the project and the variations minted from it — what changed, who directed it, and what it is worth. Splits into two tabs: Changes and Variations. | | Distribution | Which revision each company is holding right now — every document you have sent out, one row per document lineage, one column per counterparty. | | Obligations | Every deliverable, submittal and obligation in your contracts — one row each, cited to its clause. | | Delays & EOT | Each delay event with its cited evidence pack, its critical-path impact and the served-notice trail behind it. | | Claims | Progress claims in both directions and where the money actually is — claimed, certified, paid. | | Security | Every retention and bank guarantee a counterparty holds against you, when it must come back, and the chase that gets it back. | Two further registers — insurances and RFIs — are rolling out and appear only where they have been enabled for your workspace. The rest of this page covers what the registers have in common, then the Distribution register in depth because it carries two capabilities the others build on: lineage rows and as-at time travel. The commercial registers have pages of their own — see [the claims register](/en/docs/guides/commercial-control/claims-register), [delay substantiation](/en/docs/guides/commercial-control/delay-substantiation), [variations](/en/docs/guides/variations) and [the obligations register](/en/docs/guides/compliance/obligations-register). ## Registers, the Watcher and tools Three different things live in a project sidebar and they are deliberately not mixed. Getting the distinction straight is the fastest way to know where to look. | Surface | Question it answers | What you do with it | | ----------- | ----------------------------------- | ---------------------------------------------------------------------------------- | | A register | What is the state of X, right now? | Consult it, filter it, export it, hand it to someone else. | | The Watcher | What needs me right now? | Work the feed down. Everything that does not need you today stays in its register. | | A tool | None — a tool is something you run. | Kick it off, like a compliance chase campaign or an evidence pack. | The [Watcher](/en/docs/guides/projects/watcher) describes itself exactly that way: _what needs you right now — missing register entries, clocks running, decisions waiting and documents that disagree. Everything else lives in its register._ The two surfaces are complements. The register is the whole population; the Watcher is the slice of it that is on fire. ### Where recovery findings went Recovery findings — _events shaped like money with no register row_ — used to sit on a register page of their own. They now surface in the Watcher under the group heading **Missing from registers**, which is where they belong: each one is a prompt to create a row, not a row in its own right. Four kinds are detected: * **Directed work with no variation** — someone was told to do something and no variation was ever raised for it. * **Delay evidence with no notice** — the diaries and weather support a delay, but nothing was served. * **A revision change with no variation link** — a drawing moved and no commercial change followed it. * **A docket never claimed** — signed work that appears on no progress claim. A missing row is the expensive kind of gap The rows you are looking at are the easy ones. The money leaks through work that was done, directed and documented but never made it onto any register at all — which is precisely the population the Watcher’s missing-from-registers group exists to surface. Clear it, and the registers become a complete record rather than a partial one. ## Registers that share a row Some registers are the same object seen from another side. Rather than give every one of them its own sidebar row and let the band sprawl, siblings share a single row and split into tabs across the top of the page. | Sidebar row | Tabs | Why they are together | | ----------- | -------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Commercial | Changes · Variations | A variation is minted from a commercial change. It is the same story at two stages of itself — the change that happened, and the claim for it you can actually price. | Each tab is a real route with its own URL — the tabs change how you _find_ a register, not what it is. That matters because deep links keep working: an alert, a digest line, an assistant citation or a bookmark that points at the variations register still resolves straight to it, and simply opens with the right tab selected. If you live in one tab all day, bookmark that tab rather than the parent row. ## The Distribution register The Distribution register answers the document controller’s oldest question: _who has what?_ Rows are document lineages and columns are counterparties, so one glance shows which revision each company was sent, when, and whether they have opened it. Amber means that company is holding a revision older than yours. ### Rows are lineages, not files A lineage is the whole life of a document — the original issue and every revision after it — treated as one thing. When a supplier returns a revised version of something you issued, the returned document joins the same lineage rather than appearing as an unrelated file. That is what lets a single row truthfully say “Rev C is current, Rev B is superseded, and this counterparty still holds Rev A.” ### What each cell tells you Every row-by-counterparty cell carries the distribution facts for that pairing: * **Status** — whether the counterparty holds the current revision, a superseded one, was never sent the document, or has been sent it and not yet opened it. * **Version and timestamps** — which revision went out, when it was sent, and when it was first opened. * **Channel** — how the document reached them: an Exchange share, email, the assistant, a notice, or a transmittal. * **Evidence grade** — whether the link is _proof_ (the file demonstrably travelled) or _probable_ (a fuzzy association). The register never dresses a guess up as a receipt. No fake read receipts Open tracking is only shown for channels that actually support it. A document sent by plain email has no read receipt, and the register says so rather than implying tracking that does not exist. From any cell you can re-issue the current revision to a counterparty who is behind — the register is where you notice the gap and close it in the same motion. The Distribution register grid: document lineage rows down the left with their current revision labels, counterparty columns across the top, and cells showing mixed statuses — current, amber superseded, awaiting open — with sent and opened timestamps visible in an expanded cell popover. ## As-at time travel The register has a time machine. Set the **As at** date and the whole grid re-renders as it stood on that day: which revisions existed, who had been sent what, and what each counterparty had opened — as of that moment, not as of now. This is the question disputes actually turn on. Not “what is the current drawing?” but “what did the subcontractor hold on the day they poured the slab?” The as-at view answers it from the distribution record rather than from anyone’s memory. You can also download the as-at register as an evidence pack — a machine-readable snapshot of the register as it stood at your chosen date, suitable for attaching to a claim, a notice, or a dispute file. Why the record matters before the argument starts Of 990 adjudication applications lodged in NSW in FY24-25, 212 were withdrawn and 44 timed out before a determination was released (Building Commission NSW, Q4 FY24-25 report) — outcomes decided by records and deadlines before any adjudicator read the merits. A register that can prove who held what, when, is the record half of that equation. ## Registers the AI can read Registers are not just screens. The assistant reads them directly — the same rows, through the same permission filter as the person asking — rather than re-reading the source documents and re-deriving an answer. So when you ask a register question in plain language, the answer matches what the register shows you, down to the row. | Ask this | It reads | | ------------------------------------------------------------------------- | ----------------------------------------------- | | “Who is still on a superseded revision of the level 3 slab drawing?” | Distribution | | “What variations are still unapproved, and what are they worth?” | Commercial — Changes and Variations | | “Which contract obligations are overdue this month?” | Obligations | | “How many delay days have we substantiated, and which ones were noticed?” | Delays & EOT, plus the issued-notice trail | | “How much have we certified but not been paid?” | Claims, plus the cost position | | “What retention is still held against us and when is it due back?” | Security | | “What is our cash position over the next 90 days?” | A rollup across claims, variations and security | | “What approved variations have we never billed?” | Claims reconciled against Commercial | The assistant can also run the cross-register rollups — the cost position, the cash-flow outlook, unbilled variations, back-to-back gaps between what you owe upstream and what you have secured downstream, and the clock position against a single counterparty. All of it is read-only: the assistant reports the register, it does not quietly edit it. ### What else reads the registers * **Routines** — a [routine](/en/docs/guides/routines) can be triggered by a register change, so “when a variation is approved, do this” is a rule rather than a habit. * **The daily digest** — register movement becomes digest lines on [Today](/en/docs/guides/today), so the day opens with what shifted overnight. * **Evidence packs** — a pack snapshots register state at the moment it is assembled, so the state it cites cannot drift out from under it after you send it. ## Related pages * [The Watcher](/en/docs/guides/projects/watcher) — the attention feed beside the registers, including missing register entries. * [Document control overview](/en/docs/guides/exchange) — the send-and-receipt machinery that feeds the Distribution register. * [Drawing intelligence](/en/docs/guides/projects/drawing-intelligence) — how sheet numbers, revisions and disciplines are extracted from drawings. * [The commercial loop](/en/docs/guides/commercial-control) — how change signals become the typed records the Commercial register holds. * [Claims register & cost position](/en/docs/guides/commercial-control/claims-register) — claims in both directions and where the money actually is. * [Obligations register](/en/docs/guides/compliance/obligations-register) — the contract’s obligations as a live register with chase campaigns. --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/registers --- ## Project Timeline & Activity Source: https://www.alloovium.com/en/docs/guides/projects/timeline PROJECTS # Project Timeline & Activity Every project keeps a live record of what happened on it — uploads, meetings, routines, claims, schedule changes and more, newest first. The timeline is where that record lives, and it is the destination the assistant's \[T#\] citations link back to. ## Overview The `Timeline` tab on the project dashboard shows everything that’s happened on the project as one append-only feed. Events are written as things happen — by people, by routines, and by Alloovium itself — and they are never edited afterwards, which is what makes the timeline usable as a record: what it says happened on the 14th is what happened on the 14th. The timeline is permission-aware. Each person sees the events they are entitled to see, the same access rules that govern documents and retrieval. Video · \~2 min Read the project record A tour of one project's timeline: scanning today's events, filtering to Risks, searching for a variation code, expanding a collapsed routine run — then clicking a \[T#\] citation in an assistant answer and landing on the exact event it cited. ## What lands on the timeline Events come from across the platform. The major families: | Family | Examples | | --------------------- | ---------------------------------------------------------------------------------------------------------- | | Documents | Uploads, ingestion, status changes, supersessions, shares. | | Emails & updates | Updates and correspondence received and sorted into the project. | | Variations & RFIs | Variations and RFIs raised, and action items created and completed. | | Routines & AI actions | Routine runs starting, waiting and completing, drafts becoming ready, lessons matched. | | Schedule | Activities created, blocked and completed, milestones hit or slipped, lookahead changes, schedule imports. | | Compliance & risk | Compliance gaps found, requirements satisfied, risks raised and escalated. | | Meetings & site | Meetings recorded and daily site conditions. | | People & syncs | Members added or removed, integration syncs completing or failing. | Because variations, documents and routine outputs all write here, the timeline doubles as the project’s audit trail. ## \[T#\] citations from the assistant When the assistant answers from the project’s live records rather than from a document, it cites them with `[T#]` badges — `[T1]`, `[T2]` and so on — the live-record counterpart to document citations. Ask “what happened with the facade this week?” and each claim in the answer carries a badge pointing at the record behind it. Hovering a badge previews the record — its title, date and where the link leads. Clicking it goes to the source: a timeline event opens the project’s Timeline tab scrolled to that exact event, briefly highlighted; other record types — action items, meetings, routine runs — open on their own surfaces. Register rows cited in an answer (a variation or RFI, say) open the timeline with its search pre-filled to that record’s code, so every event mentioning it is in view. Two kinds of citation Document citations point into files and open the document at the cited passage. `[T#]` citations point at live records — timeline events, action items, meetings — and deep-link to where those records live. An answer can carry both. ## Filtering and searching Across the top of the feed sit filter chips — `All`, `Routines`, `Schedule`, `Risks`, `Docs`, `Syncs` and `People` — each showing a count of matching events. Chips with nothing in them stay hidden, so the row reflects what the project actually has. The search box filters events by their text as you type. Searching a code — `VO-118`, `PC-005` — is the quickest way to pull one thread out of the feed: every event that mentions the record lines up in date order, which is exactly the view a dispute or a handover needs. ## Day grouping and rollups Events are grouped by day — `Today`, `Yesterday`, then dated headings — and within a day the feed keeps itself readable by rolling up noise: * **Routine runs collapse into one entry.** A run that produced several steps shows as a single row named for the routine, expandable to the full cascade. * **Compliance gaps roll up per day.** A sweep that found many gaps reads as one line — “Found 5 compliance gaps” — with the worst severity shown, expandable to the individual findings. * **Repeated identical events compress.** A burst of the same event shows once with a count instead of flooding the day. A deep link into a collapsed group — from a `[T#]` citation, for instance — expands the group automatically so the target event is visible. ## Cited sources on events Events can carry their own citations back into the document set. An event with sources shows a `cite` chip; expanding it lists the documents behind the event, each with its page and the quoted passage. Clicking a source opens the document preview located at that passage, highlighted — so the chain runs unbroken from “something happened” to the exact line of the file that evidences it. --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/timeline --- ## Uploading Documents Source: https://www.alloovium.com/en/docs/guides/projects/uploading Document Hub # Uploading Documents Upload files to a project by drag-and-drop, file picker, or connected cloud storage. Alloovium handles the rest automatically. ## Uploading Documents You can upload documents to a project in several ways: by dragging files directly onto the upload area, by clicking the "Upload" button and selecting files from your computer, or by syncing from a connected cloud storage provider (SharePoint, Google Drive, Procore). 1. 1 Open the project Navigate to Projects in the sidebar and open the project you want to upload into. 2. 2 Upload your files Drag and drop files onto the upload area, or click "Upload" and select files from your computer. You can upload multiple files at once. 3. 3 Wait for processing Alloovium processes each document automatically. You will see a progress indicator next to each file. Processing typically takes 30–120 seconds depending on file size. 4. 4 Start querying Once processing is complete, the document status changes to "Ready". You can now query it via the AI Assistant. Batch uploads You can upload up to 50 files at once. Large batches are queued and processed in parallel — you do not need to wait for one batch to finish before uploading another. ### Supported Formats | Format | Extension | Notes | | ---------- | ----------------- | ---------------------------------------------------- | | PDF | .pdf | Includes scanned PDFs — OCR is applied automatically | | Word | .docx | Tracked changes and comments are included | | Excel | .xlsx, .xls | All sheets are processed; tables extracted | | PowerPoint | .pptx | Slide text and speaker notes extracted | | Images | .png, .jpg, .tiff | OCR applied; useful for scanned drawings | | Plain text | .txt, .csv | Processed directly; no OCR required | --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/uploading --- ## The Watcher Source: https://www.alloovium.com/en/docs/guides/projects/watcher DOCUMENTS & PROJECTS # The Watcher The Watcher is the one place on a project that answers a single question: what needs me right now? Missing register entries, clocks running, decisions waiting and documents that disagree. Everything else lives in its register. ## Overview The Watcher is an _attention feed_, not a register. A register answers _“what is the state of X?”_ — every claim, every variation, every obligation, whatever its status. The Watcher answers _“what needs me right now?”_ A row is in the feed because it needs a decision from you, and it leaves when the decision is made. That gives the feed one hard contract: **the badge count always equals the number of rows**. There is no separate unread state to clear, nothing to mark as seen, and no inbox to keep tidy. If the badge says four, four things need a decision. Clear them and the badge goes; there is no way to make the number lie. You reach it from the **Knowledge** band in the project sidebar — not the **Registers** band, because it is not one. The full registers keep their own routes and the Watcher never duplicates them; it links into them. See [Registers](/en/docs/guides/projects/registers) for the state-of-the-project side of the same records. The Watcher on a project — the unread count on the sidebar row matching the number of rows in the feed, with grouped headings running down the page: Missing from registers with a Recovery finding, On the clock showing a notice clock with days left, Needs a decision on an incoming claim, and Conflicts tagged Detected conflict. Quiet is the goal An empty Watcher reads “Nothing needs you right now” — Alloovium is still checking missing register entries, notice clocks, claims and contract conflicts in the background. They appear here only when they need a decision, so an empty feed is a result, not a gap. ## What lands in the feed Rows arrive grouped by the kind of attention they want, so a scan reads as a shape rather than a list. | Group | What it holds | | -------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Missing from registers | Findings from the Recovery radar — events shaped like money with no register row behind them. Covered in depth below. | | On the clock | Notice clocks currently running. Each row says "notice clock running" with the time remaining — "N days left", or "due today" on the last day. | | Needs a decision | Claims waiting on your response, claims in dispute, and claims Alloovium registered from a document that need confirming — those read "auto-registered — confirm it". | | Conflicts | Documents that disagree with each other. Each row is tagged either Contract set or Detected conflict so you can see whether it is a known set or something Alloovium found. | | Lapsed — review now | Windows that have already closed. The row states "lapsed N days ago" and the date the window closed, because a lapsed window is still a decision — pursue it or record that you did not. | | Their clock, their silence | Notices you served through Exchange that nobody has acknowledged. Grouped alongside Served, not acknowledged. | | Today's record | Today’s site diary hasn’t been filed — the scheduled call went unanswered, and the day stays off the record unless you call the diary line back or open the diary yourself. | Groups render only when they have rows in them. The feed never shows an empty heading, so the page length is a fair proxy for how much is outstanding. ## The Recovery radar Events shaped like money with no register row — directed work without a variation, delay days without a notice, revisions and dockets nobody claimed. The Recovery radar is the part of the Watcher that goes looking. Everything else in the feed reacts to a record that exists; the radar hunts for the record that _should_ exist and does not. Money is rarely lost on a job by being refused — it is lost by never being claimed, and the radar is the sweep for exactly that. It used to be a standalone **Recovery** register with its own sidebar row. It is now a group inside the Watcher, under the heading **Missing from registers** — because a finding is not a state to browse, it is a decision waiting on you. ### The four kinds of finding | Finding | What it means | | ----------------------------------- | ------------------------------------------------------------------------------------- | | Directed work — no variation | Someone directed work that reads as a change, and no variation was raised against it. | | Delay evidence — no notice | Evidence of delay days on the project record with no delay notice served. | | Revision change — no variation link | A drawing or document revision changed scope, and nothing links it to a variation. | | Docket — not in a claim | A docket sitting on the project that has never appeared on a progress claim. | ### Notice window and confidence A finding is only useful if you know how long you have. Every finding therefore carries a **notice window** state and a **confidence**, and the row says which in plain language. | Notice window | Reading | | ------------------------ | --------------------------------------------------------------------------------------------------------------------------- | | Window open | There is still time to act, with no computed end date on the row. | | Window open — due {date} | Time remains and the row shows the date it runs out — "notice due {date}". | | Window lapsed | The row reads "notice opportunity may have lapsed". Review it anyway — see Lapsed below. | | No time bar computed | Alloovium could not compute a deadline from the contract, so the row says "no verified deadline" rather than inventing one. | Confidence is **High** or **Medium**, and it is stated on the row rather than hidden in a score. A Medium finding is a prompt to look, not an assertion that money is owed — the radar surfaces candidates and you make the call. ### Acting on a finding Each finding offers the next real step rather than a generic open: **Review candidate** to see the evidence behind it, **Log delay** for a delay finding, and **Open variations** or **Open claims** to take it into the register where it belongs. A Recovery radar finding inside the Watcher's Missing from registers group — a Directed work, no variation row with High confidence, a Window open, due date chip reading notice due, the Review candidate and Open variations actions, and the Not recoverable dismissal beside them. ### Dismissal is a recorded decision The dismissal action is **Not recoverable**, and it is deliberately not a delete. It asks `Why not?` and will not proceed without a reason. On confirm the toast says it plainly: _“Dismissed — the reason remains on the audit trail.”_ That matters six months later. A finding you dismissed in March with “covered by the lump-sum inclusion” is a defensible position; a finding that silently vanished is a hole in the story. Dismissing removes the row from your attention, not from the record. ### The scan is read-only The radar runs automatically on a nightly sweep, so findings appear without anyone pressing anything. You can also re-check on demand with **Check again**, which re-reads the last stretch of project records — the hint on the control states the guarantee: _“This only reads project records; nothing is created or sent.”_ 1. 1 The strip tells you how current the picture is A line above the group reads "Recovery records checked {date}", so you always know how stale the findings are before you trust them. 2. 2 Press Check again to re-read The strip switches to "Checking records…" while the sweep runs. It reads existing project records only — no notice is drafted, no message is sent, nothing is registered. 3. 3 Findings update in place A rescan refreshes the findings it already raised rather than raising them again, so running it twice never doubles the list. Resolved items drop out; new ones appear. Safe to run in front of the client A scan changes nothing on the project. It creates no variation, serves no notice and sends no email — it reads what is already there and reports what is missing. If it is worth checking, it is worth checking now. The assistant can read the findings too Recovery findings are exposed to the AI assistant through a read-only, permission-filtered lookup, so “what money have we left on the table this month?” is a question you can ask in chat and get a grounded answer to. The assistant can read the findings; it cannot dismiss them. See [AI Assistant](/en/docs/guides/ai-assistant). ## Clocks and silence Two failure modes the feed exists to catch: a clock you are running out of, and a clock running against someone who has gone quiet. ### Served, not acknowledged When you serve a notice through [Exchange](/en/docs/guides/exchange) and the other side says nothing, the Watcher raises it under **Their clock, their silence** and **Served, not acknowledged**. The row states the position without softening it: _“served via Exchange — no acknowledgment in N days — chase it; the statutory clock doesn’t wait.”_ It covers the notices where silence costs you something — an **EOT notice**, a **Payment schedule**, or a **Variation notice**. Delivery is not receipt, and a served notice nobody answered is not a closed loop. ### Lapsed — review now A window that has closed does not disappear from the feed. It moves to **Lapsed — review now**, showing how long ago it lapsed and the date the window closed. The point is not to make you feel bad about a missed date — it is that a lapsed window still needs a decision recorded against it, either a pursuit on another footing or a deliberate note that it was reviewed and not pursued. A clock chip is not a reminder “N days left” and “due today” are contractual, not administrative. Missing a notice time-bar does not cost a late fee, it costs the entitlement — which is why these rows sit at the top of the feed and stay there until they are answered. ## Acting on a row Every row carries the actions that actually close it, and nothing else. Three verbs cover most of the feed. | Action | What it does | | ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Open in register | Jumps straight to the underlying row in the register that owns it — the claim, the variation, the obligation. The Watcher hands you off; it does not hold a second copy. | | Review | Opens the item and its evidence so you can make the call with the context in front of you. | | Not a variation | Records that the flagged event is not a change. Requires a second tap to confirm — the row asks "Sure? Tap again". | | Dismiss | Sets the row aside as reviewed. Also a two-tap confirm, and also recorded rather than deleted. | Both dismissive verbs are _waivers with a reason_, and the toast says so: _“Waived — the register keeps the decision on record.”_ The reason is stored against the register row as `Waived from the attention feed — reviewed, not a variation` or `…window lapsed, reviewed and not pursued`, so the register can always show who decided what and on what basis. Nothing leaves the feed silently There is no hide, no snooze and no clear-all. Every exit from the Watcher is either a decision taken in the register or a waiver with a reason attached to it — because a “we looked at it and let it go” you cannot evidence is the same as never having looked. ## Watcher, registers and Today Three surfaces, three different questions. Getting them straight is most of learning the product. | Surface | Scope | Question it answers | | ----------- | ----------------------------------------- | --------------------------------------------------------------------------------------------------------------------- | | Today | Every project you touch, personal to you. | "What is my day?" — the calendar, the queue of items waiting on you, and what changed overnight across all your work. | | The Watcher | One project, everyone on it. | "What needs a decision on this job right now?" — a live feed that empties as decisions are made. | | A register | One record type on one project. | "What is the state of X?" — every claim, variation or obligation, whatever its status, including the settled ones. | So a claim awaiting your response is a row in the Watcher _and_ a row in the claims register — but they are doing different jobs. Answer it and it leaves the Watcher; it stays in the register forever, now marked answered. See [Today](/en/docs/guides/today) for the cross-project view. ## Related pages [Registers](/en/docs/guides/projects/registers) — the state-of-the-project side: every record, settled or not. [Today](/en/docs/guides/today) — the cross-project, personal counterpart to the Watcher’s single-project feed. [Claims register & cost position](/en/docs/guides/commercial-control/claims-register) — where the claim rows in the feed land. [Variations](/en/docs/guides/variations) — raising the variation a Recovery finding says is missing. [Exchange](/en/docs/guides/exchange) — serving notices, and the acknowledgment trail the feed watches. [Site Calls](/en/docs/guides/site-calls) — the diary line behind the “Today’s record” row. --- Canonical page: https://www.alloovium.com/en/docs/guides/projects/watcher --- ## Reports & Document Generation Source: https://www.alloovium.com/en/docs/guides/reports Reports # Reports & document generation Turn the documents already in your project into finished long-form outputs — a monthly report, an infrastructure report, a tender review or a management plan — with the source material cited and a review step before anything leaves your hands. ## Overview Reports are Alloovium's long-form generated documents. Rather than writing them from a blank page, you choose the report type, point it at the right period and sources, and Alloovium drafts the sections from the documents in your project. You then review, edit and export. These outputs draw on your project corpus — uploaded documents, connected sources and prior conversations — so the content reflects your project rather than generic text. Everything remains a draft until you approve it. A generated report open in the editor — the section list down the left, the drafted body in the middle, and the sources view alongside grouping citations by document. ## Report types Each type is shaped for a different purpose and audience. Alloovium offers several distinct report and document types. Which one you reach for depends on what you are producing and who will read it. | Report type | What it produces | Notes | | --------------------- | ----------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------- | | Monthly report | A recurring project report covering progress, schedule, financials, risks, safety and quality. | Template-based; pick a reporting period and an audience (client, internal or executive). | | Infrastructure report | A branded project report generated without a template — overview, progress, safety, quality, risks and recommendations. | Sections are written from scratch and assembled into a branded document, following a fixed section set. | | Management plan | A plan built section by section against a template you provide, with a review gate on each section. | Best when you have an existing plan template to fill and want to sign off section by section. | | Tender review | A side-by-side review comparing a tender document against a response. | Produced as an in-page review rather than a downloadable file. | Audiences change the emphasis For the monthly report you can choose an audience — client, internal or executive — which adjusts which sections are included and how they are framed. The infrastructure report uses a fixed section set and does not take an audience. ## How generation works From template and sources to a finished draft. Generation follows the same broad shape across report types. You configure the report, Alloovium gathers the relevant material from your sources, drafts each section, and assembles the result into a document you can open and edit. 1. 1 Configure the report Choose the type, the reporting period where relevant, and which data sources to draw on. 2. 2 Aggregate the sources Alloovium pulls the relevant content from your project documents, connected sources and conversation history. 3. 3 Draft the sections Each section is written from the gathered material — filled into a template, or generated and assembled for template-free reports. 4. 4 Assemble the document Sections are combined into the finished output, ready to open in the editor. You can select the sources a report considers — for example your project documents, connected storage and integrations, and past conversations — so the draft reflects the material you trust. ### Templates vs. template-free A monthly report and a management plan are `template-based`: they fill a document you supply. An infrastructure report is `template-free`: it writes and assembles the sections without a pre-built layout. If you mainly want to fill a fixed form or template with project data, see [Template Filling](/en/docs/guides/template-filling), which is a distinct, more structured feature. ## Grounding & citations Generated content points back to the documents it came from. Reports are grounded in your project material. As Alloovium drafts, it retrieves the relevant passages from your documents and uses them to write each section, so the output is anchored to real source content rather than invented. Where citations are surfaced, each cited claim links back to the source document — and the specific passage within it — so a reviewer can open the original and check it side by side. Citation support is most complete in the document-generation editor, where a sources view groups citations by document and lets you jump straight to the underlying text. Check the sources, not just the prose Grounding reduces error, but it does not remove your judgement. Open the cited passages for anything material — figures, dates and commitments especially — before the report goes out. ## Review before you send Nothing is final until you approve it. Every report is a draft when it finishes generating. You open it in the editor, read through it, make edits directly, and export only when you are satisfied. This human review step is deliberate: the report is a first draft to refine, not a document to send unread. The management plan makes this explicit with a per-section review gate. Each section carries a status — pending, generating, generated, then approved, edited or rejected — and the plan cannot be finalized until you have signed off every section. You can edit a section's text, or regenerate it with your own instructions, before approving it. You are accountable for the output Generated reports can be exported to editable documents. Treat them as drafts prepared by an assistant: review, correct where needed, and confirm the facts before they represent your project externally. ## Generating a report The path from a new report to an exported document. The flow is consistent across report types. You start a new report, work through a short configuration, watch it generate, then review and export. 1. 1 Start a new report Choose the report type you want to produce and begin the setup. 2. 2 Set the period and audience Pick the reporting window where relevant; for the monthly report, also choose the audience it is written for. 3. 3 Choose sources and sections Select which sources to draw on and confirm the sections to include; audience presets pre-select sensible sections. 4. 4 Generate and watch progress Alloovium works through analysis, aggregation, drafting and assembly, showing progress as it goes. 5. 5 Review, edit and export Open the finished draft in the editor, refine it, check the cited sources, then export or save it to your project. Longer report types show their progress as they run — a management plan streams sections in as each one is drafted, so you can begin reviewing early sections while later ones are still being written. --- Canonical page: https://www.alloovium.com/en/docs/guides/reports --- ## Routines Source: https://www.alloovium.com/en/docs/guides/routines ROUTINES # Routines A routine is a repeatable piece of work you set up once and let Alloovium run for you — starting on a trigger, then working through an ordered chain of blocks that check the inbox, read the documents, ask the team, and produce a draft, before handing it back to you for review. ## Overview A routine is a repeatable job Alloovium runs for you. If you have found yourself running the same sequence week after week — check the drawings for new revisions, ask the site team for an update, draft the report — a routine captures that sequence so it happens on its own. Every routine follows the same shape: a **trigger** starts it, then an **ordered chain** of blocks runs one after another. Each step does its one job and passes what it produced to the next — a check of the documents feeds the draft that follows it. When a step would send something outward, the routine stops at a **human review gate** before it acts, and the result surfaces in the [Today](/en/docs/guides/today) feed. The review gate is not optional. Naming Routines were previously called Workflows, and you may still see that word in some screens and technical references. They are the same thing. ## The block model A routine is an ordered chain of blocks. It starts with one trigger, then runs its action and logic blocks in the order you placed them. Rather than a single opaque prompt, a routine is assembled from typed blocks. A routine begins with exactly one **trigger** that decides when it runs, followed by a sequence of **action** and **logic** blocks. The chain runs top to bottom, in order: "ask the team, then draft the minutes, then text the client" runs those three steps one after another, and the context each step gathers is available to the steps below it. It is a straight chain, not a branching or merging graph. A routine runs up to 20 action steps A single run executes at most 20 action steps. This is a blast-radius guard — a routine that would fire more than 20 side-effecting steps (texts, calls, drafts) in one run is almost certainly misconfigured. If a chain is longer, the remaining steps are left for you as a reminder rather than fired in a storm. ### Triggers A trigger decides when a routine runs. You choose exactly one when you set the routine up. | Trigger | When it fires | | ------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Run manually | You open the routine and press Run. Good for anything you do on demand rather than on a clock. | | On a schedule | Runs on a cadence you set — daily, weekdays, weekly, fortnightly, or monthly, at a chosen time. | | Around a calendar event | Fires before or after a matching calendar event, so prep or follow-up happens around a meeting. | | When an email arrives | Fires when a message lands in a connected Gmail or Outlook inbox, optionally filtered by sender or subject. | | When a document arrives | Fires when a document lands from a watched source — SharePoint, Procore, Google Drive, or a direct upload. | | When a drawing is revised | Fires when a new revision of a drawing or spec is detected. | | When someone joins the project | Fires once per new team member — an accepted invitation, a direct add, or an invite link. Steps can address the person who joined. | | When a claim moves | Fires when a progress claim changes state — submitted, certified, scheduled, paid, or disputed — optionally narrowed to one state or direction. | | When a variation moves | Fires when a variation changes state — submitted, approved, or rejected. | | Before a contract deadline | The notice clock — fires ahead of a contract obligation’s due date (notice time-bars, reports, insurances), with a lead window you choose. Fires once per obligation. | Some triggers are gated Scheduled and on-arrival triggers are rolled out per workspace and may be off by default. If you only see manual runs, those triggers have not been enabled for your account yet. ### Action blocks Action blocks are the steps that do the work. They fall into a few families. Availability can vary by workspace and some blocks depend on a connected inbox or calendar. | Block | Family | What it does | | -------------------------- | -------------- | ------------------------------------------------------------------------------------------------------------------------ | | Ask the team for an update | Talk to people | Opens an update cycle — asks named people by text, email, in-app, or phone and captures their replies back into the run. | | Send a text message | Talk to people | Sends a one-off text to the crew. No reply is captured. | | Send an email | Talk to people | Sends an email to chosen recipients from your mailbox or the Alloovium service. | | Call someone | Talk to people | Places a phone call that speaks a short script you provide. | | Check the inbox | Inbox | Reads recent mail from a connected Gmail or Outlook inbox into the run, optionally filtered by search or sender. | | Reply to the email | Inbox | Replies to the email that fired the routine (with the “when an email arrives” trigger). | | Forward the email | Inbox | Forwards the email that fired the routine to chosen recipients, with an optional covering note. | | Check the documents | Documents | Searches your project documents by query, type, or source and reads the matches into the run. | | Fill a template | Produce | Fills a saved template from matching documents. See Template Filling. | | Generate a document | Produce | Drafts a new document — a letter, notice, report, or similar — as Word or PDF. | | Add to the calendar | Track | Creates a calendar event from a description or a set date and time. | | Move a calendar event | Track | Reschedules or edits an existing calendar event. | | Cancel a calendar event | Track | Cancels an existing calendar event. | | Add to your to-do list | Track | Posts an item to your to-do list in Today. | Start from a template, not a blank canvas Most people never assemble blocks by hand. The **starter gallery** on the routines page offers ready-made templates — the notice clock, the claim response window, the RFI chase, the Friday free-work audit — that open in the builder with the trigger and steps already laid out for you to review and publish. Building from scratch is there when you need it. ### Logic blocks Logic blocks shape how the chain flows without acting on the outside world themselves. | Block | What it does | | -------------------- | --------------------------------------------------------------------------------------------- | | Split on a condition | Evaluates a condition against the run so far and continues only when it holds. | | For each | Repeats the steps that follow once per item the run gathered. | | Wait | Pauses the run for a set time or until a moment before continuing. | | Await replies | Pauses the run until people respond to an earlier ask; when a reply arrives, the run resumes. | ## The builder Where you assemble the ordered chain and see the routine end to end. The builder is where you assemble the chain. You add a trigger, then add the action and logic blocks in the order they should run, and configure each block in a side panel. Because a routine is a straight ordered chain, the steps run top to bottom — there are no branches to wire up or separate paths to merge back together. 1. 1 Open the builder From Routines, choose a catalogue card to fork or start a blank routine. The builder opens ready for your trigger and first steps. 2. 2 Choose a trigger Pick how the routine starts — manually, on a schedule, around a calendar event, or when an email, document, or drawing revision arrives. 3. 3 Add blocks in order Add the action and logic blocks the routine should run, in the sequence you want them to fire. Each step can use the context the steps above it gathered. 4. 4 Configure each block Select a block to set its options in the side panel — who to ask, the query for a document check, the template for a fill, what a drafted document should include. 5. 5 Save and test Save the routine, then run it once to confirm each step behaves as expected before you rely on it. Video · \~4 min Build your first routine Fork a catalogue card into a project, choose the trigger, add and configure each block in the side panel, run the routine once, then approve its drafted output at the review gate. ## Running a routine Watch each step as it happens, with its output in view. When a routine runs, it does so a block at a time, in order. The run sidebar shows each step with its live status — running, waiting, done, or failed — and the output it produced. You can follow a document check pulling matches, an ask-the-team step going out, and a draft composing a letter, then open any step to inspect what it returned. If a routine reaches a logic block, it can pause and wait. An _await replies_ step holds the run until people respond to an earlier ask; when the replies arrive, the run resumes from where it stopped and their answers feed the steps below. Completed runs stay in the routine's history so you can revisit the output and see how it was produced. Everything is cited When a routine reads your documents, that retrieval is grounded and carries citations through to the output, so a drafted notice or a figure it used can be traced back to the source document it came from. ## Project and cross-project scope A routine belongs to a project. Retrieval respects your permissions. A routine instance lives inside a single project. The catalogue card it was forked from is reusable across your company, but the running routine — its configuration, its history, and the documents it touches — is bound to one project and is removed with that project. Retrieval is permission-filtered. A routine only reads documents the person who owns the run is allowed to see, so a routine never widens access to material a user could not open themselves. Steps that reach beyond a single project's folders are constrained the same way. ## The human review gate Nothing leaves the workspace or acts on your behalf until you approve it. The review gate is the core promise of routines. A routine drafts; it does not send. When a run produces something that would go outward — a filled document, an email, a notice — it stops and places the result in a queue for you to review. You read it, edit it if needed, and approve or discard it. Approval is what triggers the downstream action. Workspaces that have enabled it can raise the level of autonomy per routine, from _suggest only_ to _run and hold for review_. Even at higher levels the reviewed output waits in your queue; the gate remains. This is deliberate — the routine does the tedious assembly, and a person makes the call. Available capability varies Routines and their triggers are enabled per workspace and several capabilities are still rolling out. If a block, trigger, or autonomy option described here is not visible in your account, it has not yet been turned on for you. Schedule-risk flags raised by routine runs land on [Today](/en/docs/guides/today), in the Needs attention queue. For the template mechanics behind the fill block, see [Template Filling](/en/docs/guides/template-filling). --- Canonical page: https://www.alloovium.com/en/docs/guides/routines --- ## Settings Source: https://www.alloovium.com/en/docs/guides/settings Settings # Settings Configure your account preferences, notification settings, security options, and API access. ## Account Open the account menu by clicking your avatar at the bottom of the left sidebar, then choose **Settings**. The **Profile** tab is where you set your display name and company, and pick a role so Alloovium can tailor its AI suggestions to the way you work. Your profile photo and sign-in email live in your account profile — click your avatar and open your profile to change them. On the Settings page the email field is read-only and labelled "Managed by your authentication provider", because it comes from how you sign in; if your organisation uses single sign-on, update it through your identity provider. ## Notifications Alloovium notifies you in-app and, if you choose, by email when important events occur. Tune email delivery and the priority threshold in **Account Settings → Notifications**. * **Document generation complete** — notified when a document generation job finishes. * **Background AI queries** — notified when a background mode query completes. * **Shared with me** — notified when a colleague shares a conversation or document with you. * **Document processing** — notified when a large document batch finishes processing. * **SharePoint sync errors** — notified if a sync job encounters an error. Weekly digest Alongside per-event email, a weekly digest summarises activity across your projects — a safety net even when the email threshold is set high. ## Security Alloovium uses **Clerk** for authentication, which supports single sign-on (SSO) via your organisation's identity provider, multi-factor authentication (MFA), and session management. * **Multi-factor authentication (MFA)** — enable MFA in Account Settings → Security. Alloovium uses authenticator apps (TOTP) — any standard app works. * **Active sessions** — view all active sessions for your account and remotely terminate any session you do not recognise. * **Organisation SSO** — contact your organisation admin to configure SSO via Microsoft Entra ID or Google Workspace. BYOK encryption Enterprise plans support Bring Your Own Key (BYOK) encryption for data at rest. With BYOK, Alloovium uses your organisation's AWS KMS key to encrypt all stored data. Contact sales to enable BYOK. ## API Keys If you need to access the Alloovium API programmatically (for example, to build a custom integration), you can generate API keys from **Account Settings → API Keys**. API keys are scoped to your user account and inherit your project permissions. Keys can be given a name and an optional expiry date. You can revoke a key at any time. Keep your API keys secret API keys provide full access to your Alloovium account. Never commit them to source control or share them in chat. If a key is compromised, revoke it immediately and generate a new one. --- Canonical page: https://www.alloovium.com/en/docs/guides/settings --- ## Notifications Source: https://www.alloovium.com/en/docs/guides/settings/notifications Settings # Notifications How the in-app bell and email delivery work, and how to tune the noise with the priority threshold and weekly digest. ## What sends a notification Alloovium notifies you in-app and, if you choose, by email when important events occur. Tune email delivery in **Account Settings → Notifications**. Notifications come from across the platform — document batches finishing (or failing), background answers completing, content shared with you, tasks and approvals assigned to you, routine runs that need review, and integrations that hit a sync problem. Every notification carries a priority, which is what your settings filter on. ## Your controls In-app notifications always arrive in the bell tray — they are quiet, and they give you a reviewable history. Email is where you tune the noise, with two controls: | Control | What it does | | ------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------- | | Email notifications | One switch. On, and notifications that clear your priority threshold are also emailed to your account address. Off, and everything stays in-app only. | | Minimum priority | The threshold for email. Four levels: Everything · Medium and above · High and urgent · Urgent only. | A sensible low-noise setup is email on with the threshold at **High and urgent** — you hear about failures and approvals when you are away from the app, without a message per processed document. Changes apply immediately; there is no save step. ## The weekly digest Alongside per-event email, Alloovium sends a **weekly digest** — a single summary of what happened across your projects since last week, so even a quiet email threshold never means losing the thread. One email, once a week The weekly digest is the safety net: even with email notifications off or the threshold set high, the digest keeps a once-a-week summary arriving so nothing important slips past unseen. --- Canonical page: https://www.alloovium.com/en/docs/guides/settings/notifications --- ## Security Source: https://www.alloovium.com/en/docs/guides/settings/security Settings # Security Configure multi-factor authentication, manage active sessions, and set up single sign-on for your organisation. ## Security Alloovium uses **Clerk** for authentication, which supports single sign-on (SSO) via your organisation's identity provider, multi-factor authentication (MFA), and session management. ### Multi-factor authentication (MFA) MFA adds a second verification step at sign-in, so a stolen password alone is not enough to reach your projects. Alloovium uses authenticator apps (TOTP) — any standard app works. 1. 1 Open Account Settings → Security Sign in to Alloovium, open your account menu, and go to the Security tab. 2. 2 Set up your authenticator app Scan the QR code with any TOTP app — Google Authenticator, Microsoft Authenticator, 1Password. 3. 3 Confirm with a code Enter the six-digit code from your app to verify the pairing and switch MFA on. 4. 4 Save your recovery codes Store the one-time recovery codes somewhere safe — they are the only way back into your account if you lose your device. ### Single sign-on (SSO) SSO lets your whole organisation sign in with company credentials through Microsoft Entra ID or Google Workspace, so access follows your identity provider — join, leave, and password policy included. 1. 1 Talk to your organisation admin SSO is configured for the whole organisation, not per user. If you are not the admin, this is a request, not a setting. 2. 2 Connect your identity provider Your admin arranges the Entra ID or Google Workspace connection with Alloovium — usually during onboarding, before inviting people at scale. 3. 3 Sign in through your IdP Once enabled, members pick "Continue with Microsoft/Google" at sign-in. Passwords and MFA policy are then governed by your identity provider. ### Session management Every device and browser signed in to your account holds a session. Review them periodically — especially after using a shared or site-office machine. 1. 1 Review active sessions Account Settings → Security lists every active session with device, browser, approximate location, and last-active time. Your current session is marked. 2. 2 Sign out anything you don’t recognise Terminating a session revokes it immediately — the device is signed out on its next request. 3. 3 Change your password if a session looked suspicious A revoked session cannot come back, but a compromised password can create a new one. Rotate it and keep MFA on. ### Bring Your Own Key (BYOK) On Enterprise plans, your organisation can hold the encryption key for data at rest. Alloovium encrypts stored data with your AWS KMS key, so revoking the key revokes Alloovium’s ability to read your data. 1. 1 Confirm your plan supports BYOK BYOK is an Enterprise-plan feature. Contact sales if you are unsure what your organisation is on. 2. 2 Nominate an AWS KMS key Create or choose a KMS key in your own AWS account. Your security team keeps full control of the key policy. 3. 3 Complete setup with Alloovium Alloovium walks your admin through the key-policy grant and re-encrypts your stored data under your key. From then on, all data at rest uses it. BYOK encryption Enterprise plans support Bring Your Own Key (BYOK) encryption for data at rest. With BYOK, Alloovium uses your organisation's AWS KMS key to encrypt all stored data. Contact sales to enable BYOK. --- Canonical page: https://www.alloovium.com/en/docs/guides/settings/security --- ## Site Calls & SMS Source: https://www.alloovium.com/en/docs/guides/site-calls DAILY WORKFLOW # Site Calls & SMS The best site record is the one that costs nothing to make. Alloovium can ring you at knock-off or on your site walk, hold a short spoken conversation about the day, and turn it into the diary, the flags, and the follow-ups — while you drive. ## Overview Most site knowledge dies in the ute on the way home. The site-call features exist to catch it: a scheduled call from Alloovium at a time you choose — end of day is the classic — where a voice agent asks the questions a diligent PM would ask. Anything change against the programme? Any delays, incidents, deliveries? Anything you need chased? You talk; it listens. The conversation is transcribed, read for the things that matter, and filed — a diary entry written, a latent condition flagged, an action assigned — with the call itself kept as the source, so every line of the record traces back to your own words. The commercial point is the gap this closes. Projects lose entitlement in the days between what the site knew at half past four and what the commercial record found out on Friday — notices go stale, and undocumented changes quietly become free work. A latent condition mentioned in the ute becomes a drafted variation with its notice clock visible before you're home, while the evidence is freshest and the clock has barely started. Two minutes, hands free A typical knock-off call runs under two minutes, and a mid-day site-walk call can be shorter still — thirty seconds of “hit rock, tell Dave” is enough to file a record and put two tracked actions on the board. There is nothing to type, nothing to remember, and nothing to write up later — the paperwork is the byproduct of the conversation. Video · \~2 min The knock-off call, start to finish Alloovium rings at knock-off, holds a two-minute spoken check-in about the day, and the call becomes a diary entry, a flagged variation, and an assigned action — reviewed on Today afterwards. ## The knock-off call A daily check-in that writes the site diary for you. The agent opens with the day’s essentials and follows up on what you say — if you mention hitting rock in a trench, it asks enough to write the entry properly, and checks what it hears against your project documents where it can. The output is a structured diary entry, not a raw transcript: | You say | It files | | ----------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- | | “Hit rock in the SW trench about two metres down.” | A diary entry, checked against the geotech report — and a latent-condition variation drafted if the documents disagree. | | “Storm stood two crews down from 1pm.” | A delay entry with the time window, ready for an EOT record. | | “Tell Dave to chase the switchboard before Thursday.” | An action assigned to Dave with a due date, on his list by morning. | ## What happens after you hang up The call becomes records — each one traceable to the words that created it. After the call ends, the diary entry files to the project, flagged items land where they belong — a variation draft for review, an action on someone’s list, a risk on the register — and anything that needs your sign-off appears on [Today](/en/docs/guides/today). Nothing publishes without the same human review gate that applies to everything Alloovium drafts: the call proposes, you approve. The call itself stays attached to the record: the filed entry keeps the spoken transcript alongside the structured records it produced — so if a diary line is ever challenged, the words that created it are one click away. And because the diary is a structured record rather than a note, it keeps working after it is filed: diary entries feed the project’s registers, and when a delay turns into a claim, the [delay evidence pack](/en/docs/guides/commercial-control/delay-substantiation) cites the day’s diary alongside the weather and the programme. The record you made in the ute is the record that carries the claim. ## If the diary never gets written A same-day nudge, not a Friday audit. The failure mode of every site diary is silence — a busy day ends, no entry gets made, and nobody notices until the record is needed. Alloovium watches for that: when a project day passes with no diary activity, it can send the responsible person an SMS nudge while the day is still in their head. Reply to the text or take a call, and the missing entry gets written the same evening instead of reconstructed weeks later. ## Texting with Alloovium Action items and confirmations, by SMS. Alongside calls, Alloovium can text you: the action items from your calls, reminders from your routines, and confirmations that a filed record went through. Replies work too — a quick “done” closes the loop without opening the app. Message frequency depends on what you have set up; reply STOP at any time to opt out. ## Setting it up A verified number and a time of day. Add and verify your mobile number when prompted — during onboarding, or later from Settings. Then choose when the calls should come: a daily knock-off time is the usual pattern, set up as a routine so the schedule is yours to change. The same verified number is used for SMS. ## Availability Rolled out per workspace. Voice calls and SMS are enabled per workspace and per region as calling capacity is rolled out. If the call features described here are not visible in your account, they have not yet been switched on for your workspace — talk to your Alloovium contact to get on the list. Captured calls, transcripts, and the records they produce follow the same [data handling commitments](/en/docs/legal/data-handling) as the rest of your project data. --- Canonical page: https://www.alloovium.com/en/docs/guides/site-calls --- ## Tasks & Action Items Source: https://www.alloovium.com/en/docs/guides/tasks DAILY WORKFLOW # Tasks & Action Items Most of the work Alloovium raises isn't something it does for you — it's something it puts in front of you. Tasks and action items are that queue: everything that needs a person, gathered in one place with its source and its deadline attached. ## Overview The queue lives in the **Needs attention** panel on [Today](/en/docs/guides/today). It mixes two kinds of entries: **action items** that Alloovium raised on your behalf — extracted from a meeting, a site call, an email, or produced by a routine — and **your own to-dos**, added by hand or by asking the assistant. Both kinds behave the same way: open the item, see where it came from, act on it. The point of the queue is that nothing arrives naked. An action item knows which meeting raised it, which project it belongs to, and when it is due — so deciding what to do about it rarely means going looking for context. ## Where items come from Items flow into the queue from the places work actually happens: * **Meetings** — when a [recorded meeting](/en/docs/guides/meetings) is processed, the commitments made in it become action items, each linked back to the meeting record. * **Site calls** — actions you raise on a [site call](/en/docs/guides/site-calls) land in the queue with the call, its transcript and its summary one click away. * **Routines** — a [routine](/en/docs/guides/routines) can add an item to Needs you as one of its steps, so a scheduled check that finds something becomes a task for a person, not just a notification. * **Email** — action items extracted from your [project inbox](/en/docs/guides/inbox) surface here alongside everything else. * **You** — add a to-do directly on Today, or ask the assistant to add one for you. ## What an item carries Opening an item slides out its detail panel. The due date sits at the top — and where a deadline is contractual, like the notice period on a latent-condition variation, the panel shows the clause it comes from and a live business-day countdown. Below that: the description Alloovium pre-filled from the source, the **Where it came from** link that opens the meeting, call or email that raised the item, and the project it belongs to. ## Working the queue Every item offers the same verbs: **Mark done** when it's handled, **Assign** to hand it to a teammate, **Ask assistant** to take the item into a chat with its context already loaded, and **Dismiss**for things that don't need action (with an undo, in case you dismiss the wrong one). Completed to-dos can be reopened. ### Tasks, alerts and "On it" Not everything in the queue is a task. Alerts — a flagged schedule risk, for instance — aren't "done" the way a to-do is; they're _acknowledged_. An alert row carries a quiet **On it** control instead of a completion tick: press it and the row settles back, marked as acknowledged, so the rest of the team can see someone has picked it up. When the queue gets busy, filter chips above it split the view by kind — alerts, tasks, email — so you can work one lane at a time. ### Handing work to a teammate **Assign**opens a teammate picker scoped to the item's project. The item moves to their queue with all of its context — source link, deadline, pre-filled description — intact, so a hand-off never means re-explaining. Why the source link matters The Where-it-came-from link is the quiet workhorse of the queue. A task that says "chase the revised drainage layout" is an errand; the same task one click from the meeting where it was promised is a conversation you can have immediately. --- Canonical page: https://www.alloovium.com/en/docs/guides/tasks --- ## Template Filling Source: https://www.alloovium.com/en/docs/guides/template-filling Routines # Template Filling Template Filling takes a DOCX template you provide and fills its placeholders using information extracted from the documents in your project. Useful for RFIs, variation notices, transmittals, submittal cover sheets, ITPs, monthly reports, or any recurring document your team produces from a standard format. ## Overview Template Filling takes a DOCX template you provide and fills its placeholders using information extracted from the documents in your project. Useful for RFIs, variation notices, transmittals, submittal cover sheets, ITPs, monthly reports, or any recurring document your team produces from a standard format. When to use which If you have a blank DOCX template with placeholders, use **Template Filling**. If you need a brand-new document and don't have a template, use **Document Generation**. ## Template Filling 1. 1 Click "Template Filling" in the sidebar The page loads with upload zones for your source documents and blank template. 2. 2 Upload your files and click "Fill Template" Drop in your source documents and blank DOCX template, then click "Fill Template". The filled document appears with every value cited back to its source — export as DOCX when ready. Video · \~2 min Fill a template from your documents Pick the source documents, drop in a blank Word template, run the fill, and open the completed file — every inserted value arrives as a tracked change citing the document it came from. ## History Every filled template and generated document is saved under the **History** tab on the Routines page. From there you can re-open, re-export, or regenerate any past document. Running the same template again on an updated project pulls the latest information without losing the edits you made to earlier versions. A completed fill opened in Word — inserted values shown as tracked changes through the document, with the comment on one fill citing the source document it was taken from. Citations stay with the document When you re-open a past document from History, the original citations remain intact. You can click any cited value to jump straight to the source document and page. --- Canonical page: https://www.alloovium.com/en/docs/guides/template-filling --- ## Today Source: https://www.alloovium.com/en/docs/guides/today DAILY WORKFLOW # Today Today is the home you open first each morning. It brings the calendar, the items waiting on you, and everything that changed across the project into one place — deadlines, drafts and flags surfaced while there is still time to act on them — so the day starts ahead of the project rather than chasing it. ## Overview Today is designed as a cockpit for the project manager. Rather than asking you to check email, then the schedule, then site reports, then approvals, it assembles those streams into a single surface. The intent is simple — a “no surprises” view of the project you can scan in a minute and act on without leaving the page. Two things sit side by side. On the left is the calendar, the spine that anchors the day in time. On the right is `Needs you`, the queue of items waiting on a decision from you. Below the queue is a recent activity feed that answers “what happened since I last looked?” across documents, meetings, site updates, routines, and more. A morning on Today · 1 min 17 sec ## How Today is laid out The left column toggles between two views. `Calendar` is the default and shows your schedule as a timeline. `Inbox` switches the same column to your connected mailbox when you want to read and triage email in context. The right column holds the `Needs you` queue with a set of filter tabs across the top, followed by the recent activity feed. Every item in Today is scoped to your projects, so the view reflects your work rather than the whole organisation. ## The calendar spine The calendar is the core of Today. It offers `Day`, `Week`, and `Month` views, and defaults to Week. Day and Week lay hours down the side so meetings and scheduled work read at a glance; Month gives you the wider shape of the period ahead. The timeline is continuous — calendar events, task due dates, scheduled activities, site updates, and routine runs all appear on the same rail so the day reads as one picture rather than several disconnected lists. Your chosen view is remembered between visits. Meetings appear here Recorded meetings and their outcomes flow onto the calendar and into the activity feed. See [Meetings](/en/docs/guides/meetings) for how recording and transcription work. ## Needs you `Needs you` is the queue of items waiting on you. It gathers your to-dos and action items — raised from site calls, meetings, routines and emails, or added by hand — together with anything flagged as at risk, and lets you narrow the list with filter tabs. Tabs only appear when they have something in them, so the queue stays quiet when there is nothing to do. | Tab | What it shows | | ------ | -------------------------------------------------------------------------------------------- | | All | Every item waiting on you, in one list. This is the default. | | Tasks | To-dos and action items — raised from calls, meetings, routines and emails, or added by you. | | Alerts | Work you or your team have flagged as at risk, such as a slipping pour date. | | Email | Commitments Alloovium has picked out of your connected inbox. | Opening an item brings up its detail panel, including where it came from and any deadline attached to it. From there you can act on it, mark it done, or set it aside. The Compliance lens Compliance items live in their own console rather than in this queue. Organisations that run the full compliance engine get a dedicated view of gaps and conformance — see [Compliance](/en/docs/guides/compliance). ## Action items and drafts Some items in `Needs you` are action items — a specific piece of downstream work that Alloovium has recognised and staged for you. An action item is typically raised automatically from a source such as a site-walk call, a meeting summary, a routine run, or an email, and it carries a link back to that source so you can see exactly where it came from. For typed action items — a variation, an RFI, or a purchase order — the point is not just the reminder but the head start. Alloovium opens an intake pre-filled from the source context, you complete the remaining details, and it produces a draft you can refine. A variation raised from a note on site can therefore arrive on Monday morning already framed, rather than as a blank page. 1. 1 An item is raised A routine run, meeting summary, call, or email produces an action item that lands in Needs you with a link to its source. 2. 2 Open it and review the context The detail view shows where the item originated and any deadline. For a variation or RFI, it also offers to start a pre-filled draft. 3. 3 Complete the intake Alloovium fills what it can from the source and asks you for the rest, so you confirm rather than compose from scratch. 4. 4 Produce and refine the draft The finished intake generates a variation or RFI draft, linked back to the action item, ready for you to review. You can also mark an action item done, dismiss it, hand it to a teammate, or take it to the assistant for help. Items that have produced a draft show as drafted so you can see progress at a glance. For the document side of variations and RFIs, see [Variations](/en/docs/guides/variations). Contractual notice clocks Some action items — such as a latent-condition variation — carry a notice deadline shown directly on the item. Treat these as time-sensitive; they exist so a contractual clock does not quietly run out. ## Ahead of the clock Two cards below the hero row watch the things that go wrong quietly: contract deadlines and delegated work. **Contract deadlines** is the radar over every dated obligation Alloovium extracted from your contracts — notices, reports, insurances, milestones — counting down with a day chip per row. Notice time-bars are flagged and given extra runway, because missing one does not cost a late fee; it costs the entitlement. Each row names its clause and the contract it came from, and anything already late says so in days over. **Waiting on others** is the ball-in-court view: every open task you have handed to someone else, with who has it and how long they have had it. Anything with no movement for a week is flagged as stalled — not an accusation, a prompt — so a two-day job cannot quietly become a three-week problem. Quiet by default Both cards render only when they have something to say. No dated deadlines and no outstanding delegations means no cards — an empty radar is noise, not reassurance. ## The activity feed Below the queue, the activity feed answers “what changed since yesterday?” It pulls together events from across the project — documents added, meetings recorded, RFIs and change requests, calendar changes, routine runs, team check-ins from site, call transcripts, and assistant conversations — into one time-ordered stream. When many similar events happen together, they collapse into a single line so a bulk upload reads as “several documents added” rather than flooding the feed. Meetings, site calls, and messages stay as individual rows because each one usually matters on its own. The feed refreshes as new activity arrives, so a call that just finished shows up without a manual reload. ## Inbox Switching the left column to `Inbox` opens your connected mailbox inside Today. This keeps email beside the calendar and the queue, so you can read a message and act on the item it raised without moving between tools. The triage classification and mailbox setup are covered in depth on [Project Inbox](/en/docs/guides/inbox). Email that carries a task or decision can be turned into an action item so it is not buried in a thread — see [Action items and drafts](#action-items) below. Connecting a mailbox is covered in [Integrations](/en/docs/guides/integrations). ## What feeds Today Today is a reflection of the rest of Alloovium rather than a place you fill in by hand. The more of the platform you use, the more complete the view becomes. ### Routines Scheduled routines run on their own cadence and post their results into Today — upcoming fires and recent completions appear on the calendar and in the activity feed, and a routine can raise an action item for you to finish. See [Routines](/en/docs/guides/routines). ### Meetings and calls Recorded meetings and site-walk calls contribute their outcomes to Today. A summary can surface the commitments and decisions from a conversation and raise the follow-up work as action items. See [Meetings](/en/docs/guides/meetings). ### Email and site updates Connected email and updates from site both feed Today. Email produces triage-ready action items, and team check-ins from the field appear in the activity feed and roll up for review, so what happens on site is visible without chasing it. --- Canonical page: https://www.alloovium.com/en/docs/guides/today --- ## Variations Source: https://www.alloovium.com/en/docs/guides/variations CHANGE MANAGEMENT # Variations A variation is a change to the scope, cost or programme of your contract. Alloovium helps you catch one early, turn it into a claim while the detail is fresh, and keep track of it through to approval — so entitlement is not lost to the paperwork. ## Overview Variations in Alloovium are more than a list. Each one records the change itself, the contract clause that entitles you to claim it, the cost impact, and the time impact — and it remembers where the change was first mentioned, whether that was a site update, an email, or a phone call. Most variations do not begin life as a variation. They begin as a signal in the day-to-day flow of the job. Alloovium surfaces that signal as an action item on [Today](/en/docs/guides/today), helps you draft the variation from it, and then follows the draft as it is submitted and approved. Why this matters On a live job the risk is not that you raise a bad variation — it is that a legitimate change is never claimed because no one wrote it up in time. Surfacing the signal and pre-filling the draft is how Alloovium reduces that under-claim risk. ## From signal to claim A variation moves through a predictable path. You are always in control of each transition; Alloovium does the retrieval and the first draft, and you decide what is submitted. 1. 1 Surface A potential change is detected in a site update, email or call and appears as an action item on Today with the kind set to variation. 2. 2 Intake Alloovium pre-fills the variation questions — scope, trigger, clause, cost basis, time impact — from the surrounding context, leaving you to confirm or correct. 3. 3 Draft From your answers, Alloovium creates a variation record and links it back to the action item so the trail is preserved. 4. 4 Submit You review the draft, then submit it. Alloovium renders the document and the status moves to submitted. 5. 5 Track When the contract-side signer responds, you record the outcome and the variation moves to approved or rejected. Video · \~2 min Raise a variation from a site signal A change surfaces as an action item on Today; confirm the pre-filled intake questions, create the draft variation, submit it, and watch it appear in the register as submitted. ## Where variations start The starting point for most variations is an action item. Action items are the short list of things that need a decision from you, and they carry the provenance of where each one came from so you can trace a claim back to its source. An action item destined to become a variation has its kind set to `variation`. It records the source of the signal, a verbatim quote where one is available, and any citations to the underlying update. Working an action item through to a draft is done from [Today](/en/docs/guides/today). Meetings feed the same flow. When a change is discussed in a recorded meeting, Alloovium can surface it for your review and route it to a draft variation. See [Meetings](/en/docs/guides/meetings) for how items extracted from a transcript are dispatched. ## Intake and drafting Drafting a variation happens in two moves. First, intake gathers the facts. Second, the draft turns those facts into a variation record you can edit and submit. ### Intake When you open the action item, Alloovium pre-fills a set of variation questions from the context it already has. Each question is pre-populated where possible and clearly marked so you know what was inferred and what still needs your input. The questions cover the scope of the change, what triggered it, the entitling clause, who identified it and when, the cost impact and its basis, the time impact and the affected activities, and whether the client has been notified. ### Draft Once you confirm the answers, Alloovium creates the variation. The cost impact and time impact you entered are written to the record, the entitling clause is stored as the source clause reference, and the citations and supporting evidence are kept alongside the draft. The action item is then linked to the new variation, so the two remain connected. Grounded in your documents The clause reference and supporting citations point back to the contract and correspondence in your workspace, not to generated text. Review them before you submit. ## Status and impact fields A variation carries a small set of fields that together describe the change, its commercial effect, and where it sits in the approval process. | Field | Meaning | | ----------------------- | ------------------------------------------------------------------------------------------------ | | Variation number | Your reference for the change, for example VO-042. | | Title | A short description of the change. | | Cost impact | The value of the change in the contract currency (default AUD). | | Time impact (days) | Positive delays completion, negative recovers schedule, and empty means the change is cost-only. | | Source clause reference | The contract clause that entitles the claim, for example cl 25.3. | | Approver | The name of the contract-side signer, such as the Superintendent. | | Disputed | A flag for a change that is contested. | The status field follows a fixed sequence: | Status | What it means | | --------- | ----------------------------------------------- | | Draft | Under composition — not yet sent to the client. | | Submitted | Rendered and sent to the client for review. | | Approved | The client accepted the change. | | Rejected | The client declined the change. | Rejected is not the end of it A rejected variation is worth keeping. Understanding why a claim was declined informs how you frame the next one, and the record of the change stays available for reference. ## Notice deadlines Some changes — latent conditions in particular — carry a contractual obligation to give notice within a set window of discovering them. When a discovery date is captured, Alloovium tracks the notice deadline against it and shows how many business days remain. The countdown appears on the action item so the deadline stays visible while you work. Items with a notice deadline are given prominence on [Today](/en/docs/guides/today), with the soonest due first, so a notice period does not quietly expire. ## The variation register The variations for a project can be listed together, with their number, title, status, cost impact and time impact, and the dates they were submitted and approved. This gives you a single view of every open and resolved change on the job. The variation register for a project — rows listing variation number, title, cost impact and time impact, with a mix of draft, submitted and approved status badges and the submitted/approved dates visible. --- Canonical page: https://www.alloovium.com/en/docs/guides/variations --- ## Data Handling Commitments Source: https://www.alloovium.com/en/docs/legal/data-handling Legal # Data Handling Commitments How Alloovium accesses, stores, and protects your documents and data when you connect an integration. ## Our Principles Alloovium is built for construction professionals who trust us with sensitive project documents. We treat your data with the same care you would apply to a physical filing cabinet: we only touch what you give us access to, we use it only to provide the service you asked for, and we delete it when you ask us to. Every integration you connect is scoped to read-only access. We never write, modify, or delete files in your connected systems. ## What we store When you connect an integration (SharePoint, Google Drive, Procore, etc.), Alloovium retrieves document content and stores it in your organisation's encrypted vault on our infrastructure. We store: the full text of documents (for AI querying), document metadata (file name, size, last modified date), and where available, the original file (for full-fidelity preview). We also store the OAuth access tokens required to read your files. These tokens are encrypted at rest using AES-256 and are scoped to the minimum permissions required for each integration. ## What we do not do We do not use your documents to train AI models. We do not share your documents with other Alloovium customers. We do not sell or broker access to your data. We do not retain copies of your documents after you disconnect an integration beyond the deletion SLA described below. ## Retention and deletion When you disconnect an integration or delete a project, Alloovium begins purging the associated documents from its storage. Complete deletion from all systems (including backups) takes up to 30 days. If you close your Alloovium account, all your data is deleted within 30 days of account closure. You can request an accelerated deletion by contacting support. ## Security All data is encrypted in transit (TLS 1.2+) and at rest (AES-256). Infrastructure is hosted on AWS with SOC 2 Type II compliant controls. Access to production data is restricted to authorised Alloovium engineers on a need-to-know basis and is logged. ## Integration vendor compliance Alloovium is built to be compatible with the developer policies of every construction platform we integrate with. We do not bulk-export, mirror, or warehouse data from third-party platforms (Procore, Autodesk Construction Cloud, Aconex, SharePoint, etc.). We read documents on demand to answer a specific query, surface the result, and discard the raw retrieval. We do not train, fine-tune, or benchmark any AI model on data sourced from these platforms — including aggregate or de-identified data. Large-language-model use is strictly inference-time, scoped to the question the user is asking in the moment — the same operational pattern as Procore Assist's own Azure OpenAI usage. This stance is contractually binding under our DPA with each customer and is designed to keep Alloovium on the right side of every integration partner's terms of service. ## Questions Questions about how your data is handled? Contact us at privacy@alloovium.com or through the in-app support chat. --- Canonical page: https://www.alloovium.com/en/docs/legal/data-handling --- ## Use Cases by Role Source: https://www.alloovium.com/en/docs/use-cases Use cases # How construction teams use Alloovium Sixty real workflows, organised by the seat you sit in. Every use case has a short walkthrough video showing the real thing — pick your role and start with the workflow that hurts most. [Project ManagersRun the project and the month-end report from one place.10 use cases](/en/docs/use-cases/pm) * [01Start the day ten miles ahead](/en/docs/use-cases/pm#morning-triage) * [02Prep the weekly meeting in minutes](/en/docs/use-cases/pm#weekly-meeting) * [03Answer the client with receipts](/en/docs/use-cases/pm#client-question) * [04Catch a program landmine early](/en/docs/use-cases/pm#program-landmine) * [05Turn a site instruction into a variation](/en/docs/use-cases/pm#site-instruction-variation) * [06Chase deliverables automatically](/en/docs/use-cases/pm#chase-subcontractors) * [07The monthly report without the weekend](/en/docs/use-cases/pm#monthly-report) * [08Onboard a new engineer in a day](/en/docs/use-cases/pm#onboard-engineer) * [09Know what changed while you were away](/en/docs/use-cases/pm#back-from-leave) * [10Escalate a delay with an evidence pack](/en/docs/use-cases/pm#delay-evidence) [Site ManagersCapture what changed on site before it gets lost.10 use cases](/en/docs/use-cases/site-manager) * [01Phone the site diary in from the ute](/en/docs/use-cases/site-manager#phone-in-diary) * [02Check the detail against the current drawing](/en/docs/use-cases/site-manager#check-the-drawing) * [03What does the spec say about this pour?](/en/docs/use-cases/site-manager#spec-on-the-spot) * [04Raise the RFI before it blocks the work](/en/docs/use-cases/site-manager#raise-rfi) * [05Know what’s coming next fortnight](/en/docs/use-cases/site-manager#lookahead-week) * [06Hold points before you cover it up](/en/docs/use-cases/site-manager#hold-points) * [07The prestart that writes itself up](/en/docs/use-cases/site-manager#toolbox-record) * [08Verify the delivery against the docket](/en/docs/use-cases/site-manager#delivery-docket) * [09The latest revision, always](/en/docs/use-cases/site-manager#latest-revision) * [10Capture the lesson while it’s fresh](/en/docs/use-cases/site-manager#capture-lesson) [Safety & HSEQCatch non-conformances before they surface in an audit.10 use cases](/en/docs/use-cases/safety-manager) * [01Catch it before the audit does](/en/docs/use-cases/safety-manager#conformance-dashboard) * [02Credentials chased before they lapse](/en/docs/use-cases/safety-manager#expiring-credentials) * [03The audit pack in an afternoon](/en/docs/use-cases/safety-manager#audit-pack) * [04SWMS reviewed against the plan](/en/docs/use-cases/safety-manager#swms-review) * [05Root cause with receipts](/en/docs/use-cases/safety-manager#ncr-root-cause) * [06ITP records complete before you cover up](/en/docs/use-cases/safety-manager#itp-completeness) * [07The incident chronology from the record](/en/docs/use-cases/safety-manager#incident-chronology) * [08Who’s inducted, who isn’t](/en/docs/use-cases/safety-manager#induction-status) * [09Toolbox talks that hold up](/en/docs/use-cases/safety-manager#toolbox-records) * [10Lessons that change behaviour](/en/docs/use-cases/safety-manager#lessons-loop) [Commercial ManagersEvery entitlement claimed, every notice in before the bar.10 use cases](/en/docs/use-cases/commercial) * [01The progress claim built from records, not memory](/en/docs/use-cases/commercial#progress-claim) * [02Notice deadlines that never slip](/en/docs/use-cases/commercial#notice-deadlines) * [03Assess a subcontractor’s variation in an hour](/en/docs/use-cases/commercial#assess-subbie-variation) * [04What did they change?](/en/docs/use-cases/commercial#what-changed) * [05A correspondence record that holds up](/en/docs/use-cases/commercial#correspondence-record) * [06Contractual letters drafted from the contract](/en/docs/use-cases/commercial#contract-letters) * [07The EOT claim with a cited chronology](/en/docs/use-cases/commercial#eot-chronology) * [08Reconcile the variations account](/en/docs/use-cases/commercial#reconcile-variations) * [09Security-of-payment dates, jurisdiction-aware](/en/docs/use-cases/commercial#sopa-dates) * [10The evidence pack for the dispute that’s coming](/en/docs/use-cases/commercial#evidence-pack) [ExecutivesSee margin and schedule risk across every project.10 use cases](/en/docs/use-cases/vp-cxo) * [01The one project that needs you today](/en/docs/use-cases/vp-cxo#portfolio-exceptions) * [02The cash position without chasing PMs](/en/docs/use-cases/vp-cxo#cash-position) * [03The margin leak, surfaced](/en/docs/use-cases/vp-cxo#margin-leaks) * [04The monthly board pack from the record](/en/docs/use-cases/vp-cxo#board-pack) * [05Schedule risk weeks earlier](/en/docs/use-cases/vp-cxo#schedule-risk) * [06Know what the AI did and who approved it](/en/docs/use-cases/vp-cxo#ai-activity-ledger) * [07Learn across tenders, not just jobs](/en/docs/use-cases/vp-cxo#win-loss-learning) * [08Every team runs the same play](/en/docs/use-cases/vp-cxo#standardise-teams) * [09Show the client the receipts](/en/docs/use-cases/vp-cxo#client-transparency) * [10A new team productive in a week](/en/docs/use-cases/vp-cxo#onboard-team) [SubcontractorsYour copy of every package, your proof of every send.10 use cases](/en/docs/use-cases/subcontractor) * [01The package lands with its paper trail](/en/docs/use-cases/subcontractor#receive-package) * [02Sign it where you received it](/en/docs/use-cases/subcontractor#sign-in-place) * [03Read the amendment like a mark-up](/en/docs/use-cases/subcontractor#what-they-changed) * [04Prove you sent it, prove they read it](/en/docs/use-cases/subcontractor#provenance-trail) * [05One link and you’re on the project](/en/docs/use-cases/subcontractor#join-project) * [06Know what you signed up to](/en/docs/use-cases/subcontractor#ask-your-contract) * [07Submissions with a status, not a thread](/en/docs/use-cases/subcontractor#portal-submissions) * [08The request that can’t get lost](/en/docs/use-cases/subcontractor#document-requests) * [09Dayworks the builder can verify](/en/docs/use-cases/subcontractor#daywork-records) * [10The claim that can’t be “never received”](/en/docs/use-cases/subcontractor#claim-on-record) New to Alloovium? Start with [Getting Started](/en/docs/guides/getting-started) — account, first project, and your first cited answer in under an hour — then come back to your role's use cases. --- Canonical page: https://www.alloovium.com/en/docs/use-cases --- ## Use Cases for Commercial Managers Source: https://www.alloovium.com/en/docs/use-cases/commercial Use cases · Commercial Managers # Eleven ways commercial managers use Alloovium The commercial manager lives in notices, claims, and the contract — and entitlement is lost in the gaps between them. These are the eleven workflows commercial managers run in Alloovium every month, each with a short video showing the real thing. ## The loop the eleven workflows feed Every workflow on this page is a link in the same chain. Something happens — an instruction buried in an email, a latent condition on the evening site call, a decision in Tuesday's meeting — and from that moment your entitlement is either preserved or it quietly expires. Alloovium runs the whole chain: the event is captured where it first appears, the contract clause and its notice clock are identified while there is still time to act, the notice is drafted with citations and waits for your approval, the send leaves a receipt, the variation enters the register, and the register is checked against every claim until the money is certified and paid. Read the eleven workflows below as entry points into that loop. You can start anywhere — the record carries it the rest of the way. Video · \~4 min No missed notice, no undocumented change, no free work One event walked through the whole chain — a latent condition surfaces on the evening site call, the notice clock lands on Today with the clause behind it, the cited notice is drafted and approved, the variation enters the register, and the reconciliation confirms it was never left as free work. ## The progress claim built from records, not memory Claim day arrives and the narrative has to be reconstructed: what was actually done this month, which variations moved, what the site diaries say happened. Alloovium assembles the claim narrative from the project record — site diaries, meeting records, and the variations register — so the claim describes the month as it was captured, not as anyone remembers it. The **claims register** then carries the claim through its life — submitted, assessed, certified, paid — with a receipt for every transition, so three months later nobody argues about when it was lodged or what it said. Video · \~2.5 min Claim day without the reconstruction Drafting the monthly claim narrative from site diaries and the variations register, then moving the claim through the claims register with a receipt landing on each transition. ## The 20th of the month, without the lost week Claims run the other way too. On the 20th, every subcontractor's progress claim lands at once — percent-complete line items, up to a hundred per contractor, each one a rolling document that has to be checked against last month's claim and against what was actually done on site, inside a contracted payment window. Alloovium files each claim to its project and puts it in front of the right person, then answers the checking question directly: _“Compare this month's claim against last month's — what's moved, and is anything claimed that isn't on the record as delivered?”_The movements come back as a table with a citation into each document, and anything claimed with no supporting record — the pump that never arrived — comes back as a flag with a specific question for tomorrow's site walk. Video · \~2.5 min The 20th of the month A subcontractor's claim filing itself to the project and landing with the right engineer, the month-on-month comparison answered as a cited table, the item claimed with no delivery on the record flagged for the site walk, and the verified claim going up for approval inside the response window. ## Notice deadlines that never slip Most entitlement isn't lost on the merits — it's lost to the time bar. The contract's notice periods for EOTs, latent conditions, and variations are buried in clauses nobody re-reads under pressure. Ask the contract directly: _“What must we notify within 10 business days under this contract?”_ — and get the clauses back, cited. Then set a **routine** to watch for the events that start the clock and remind you before the bar falls — so the notice goes in with days to spare, not a scramble at 4:55pm on the last day. Video · \~2 min Beating the time bar Asking the contract for its notice obligations in the project ask bar, reading the cited clause answers, then creating a routine that flags a delay event days before the notice period expires. ## Assess a subcontractor's variation in an hour A subcontractor's variation claim lands: eleven pages, day rates, and a scope story that sounds generous. Assessing it properly means checking their claim against the subcontract scope, the schedule of rates, and the directions that were actually given — normally a day of digging through three document sets. Ask Alloovium to check the claim against the record and the discrepancies come back cited: the rate that doesn't match the schedule, the work already inside the subcontract scope, the direction they say they received that isn't on the record. You assess from the discrepancy list, not from scratch. Video · \~2.5 min The variation claim under the microscope Uploading a subcontractor's variation claim, asking the assistant to check it against the subcontract scope and schedule of rates, and clicking through each cited discrepancy to its source. ## What did they change? The amended contract comes back through **Exchange** looking like the one you sent — except it isn't. Somewhere in ninety pages, a liability cap moved, a notice period shrank, and a “reasonable” became a “sole discretion.” Reading both versions side by side is how edits get missed. Alloovium compares the returned document against the version you sent and surfaces every paragraph-level change — added, deleted, reworded — in one list. Every edit surfaced, none buried on page 47. Video · \~2 min Every edit on page 47 Receiving an amended subcontract back through Exchange, running the comparison against the sent version, and walking the paragraph-level diff list to the clause they hoped nobody would read. ## A correspondence record that holds up When a claim is contested, the first question is always the same: _“When did you send it, and can you prove they got it?”_ Email trails are a weak answer — versions drift, attachments go missing, and forwarding chains prove nothing. **Exchange** keeps the full audit trail on every send and receive: who sent what, when, which version of which document, when it was opened, and who signed. The transmittal record is built as you work, not assembled under subpoena. Video · \~90 sec The transmittal trail, end to end Sending a contract amendment through Exchange, then opening the share's audit trail — sent, delivered, read, signed and returned — each event stamped against the exact document version. ## Contractual letters drafted from the contract A delay notice or a show-cause response has to lean on the actual clauses — the right clause numbers, the right defined terms, the notice requirements met on their face. Drafting from a template means someone else's contract; drafting from memory means risk. Alloovium's document generation grounds the letter in this contract: the relevant clauses pulled and referenced, the facts drawn from the project record, the draft waiting for your review. Nothing sends without a human reading it first. Video · \~2.5 min The delay notice that cites its clauses Asking for a delay notice grounded in the head contract, stepping through the drafted letter against the cited clauses, editing one paragraph, and downloading the DOCX for issue. ## The EOT claim with a cited chronology An EOT claim lives or dies on its chronology: what happened, in what order, and what the records show at each point. Building it manually means weeks across inboxes, diaries, and minute books — and the superintendent's assessor will test every date. Ask Alloovium to build the delay chronology and it assembles the sequence across emails, site diaries, and meeting minutes — every fact linked to its source document, so the claim reads as a record, not a story. Video · \~2.5 min The chronology that assessed itself Requesting a delay chronology for the facade slip, watching the dated sequence assemble across emails, diaries and minutes, and clicking a chronology entry through to its highlighted source. ## Reconcile the variations account The variations account drifts in three directions at once: what's been approved, what's been claimed, and what's actually been paid. The gaps between those three columns are real money — and the classic leak is the approved variation nobody ever claimed. Ask for the reconciliation across the register and the gaps come back as a list: approved but never claimed, claimed but not certified, certified but short-paid — each line linked to the variation and its documents, so the recovery starts the same afternoon. Video · \~2 min Finding the variation nobody claimed Asking for an approved-versus-claimed-versus-paid reconciliation across the variations register, and opening the approved-but-never-claimed line straight into the variation record. ## Security-of-payment dates, jurisdiction-aware Security-of-payment deadlines are unforgiving and different in every jurisdiction: the reference date, the due date for the payment schedule, the adjudication window. Getting one wrong doesn't weaken your position — it extinguishes it. Ask the contract and the applicable Act together: _“Which dates bind this payment claim, and when does the schedule fall due?”_ The answer comes back cited against both, and the deadlines land on the project calendar where the whole team can see them. Video · \~2 min The dates the Act actually sets Asking the assistant which security-of-payment dates bind the current claim under the contract and the applicable Act, reading the cited answer, and watching the deadlines land on the project calendar. ## The evidence pack for the dispute that's coming You can usually see the dispute coming: the variation the superintendent keeps “assessing,” the delay the client won't acknowledge. The side that wins is the side that arrives with the record already assembled — not the side that starts reconstructing after the letter of demand. Ask for an **evidence pack** on the contested variation or delay and Alloovium assembles the chronology and the cited bundle — notices, directions, correspondence, diary entries — built from records captured as they happened, ready for the contract administrator or your lawyer before it escalates. Video · \~3 min Arriving armed Requesting an evidence pack for a contested variation, reviewing the assembled chronology with every entry citing its source document, and exporting the bundle before the dispute letter arrives. Looking for a different seat? See the use cases for [project managers](/en/docs/use-cases/pm), [site managers](/en/docs/use-cases/site-manager), [safety & HSEQ](/en/docs/use-cases/safety-manager), [executives](/en/docs/use-cases/vp-cxo), and [subcontractors](/en/docs/use-cases/subcontractor). --- Canonical page: https://www.alloovium.com/en/docs/use-cases/commercial --- ## Use Cases for Project Managers Source: https://www.alloovium.com/en/docs/use-cases/pm Use cases · Project Managers # Twelve ways project managers use Alloovium The PM runs the job from the cockpit: schedule, cost, client, and team all land on your desk. These are the twelve workflows PMs run in Alloovium every week — each with a short video showing the real thing. ## Start the day ten miles ahead The first fifteen minutes of a PM's day decide whether the day runs you or you run it. **Today** opens on the week's calendar with planned site work pinned above the hours, the **Needs attention** queue on the right, and everything that moved overnight — emails classified, documents received, routines fired — already triaged below. Instead of opening six tabs and an inbox of 80 unread, you scan one screen: what's due, what's waiting on you, and what changed. Anything that needs a decision is a click away from its source document. Video · \~2 min A PM's first fifteen minutes on Today Opening Today at 6:45am — scanning the calendar spine with the pour pinned to Saturday, working the Needs-attention queue top to bottom, and clearing two approvals before the site call. ## Prep the weekly meeting in minutes Upload last week's recording — or record the meeting live — and Alloovium produces the meeting record: a PM-voice summary at the top, the full transcript beneath it, and the attendees identified from the conversation. Action items, decisions, variations, and risks are extracted with the verbatim quote each one came from, waiting in a review queue for you to approve or dismiss. Walking into this week's meeting, you ask: _“What did we agree last week and what's still open?”_ — and get a cited answer against the record, not someone's memory of it. Video · \~2.5 min From recording to meeting record Uploading a 34-minute site coordination recording, opening the finished record — summary, transcript, attendees — then approving three extracted items from the review queue and watching them land on the project. ## Answer the client with receipts The client emails: _“Why is the facade behind and what are you doing about it?”_ You ask the assistant the same question, scoped to the project. The answer comes back with numbered citations — the crane allocation decision from the meeting record, the panel delivery dates from the supplier correspondence, the recovery sequence from the lookahead. Click any citation and the source opens with the passage highlighted. Share the conversation link with the client, or copy the cited answer into your reply — either way you're answering with the record, not a recollection. Video · \~90 sec A client question answered with citations Typing the client's exact question into the project ask bar, the answer streaming in with citation chips, clicking through to the highlighted source, then sharing the conversation. ## Catch a program landmine early The blindside is the worst kind of delay: a long-lead item nobody re-checked, a sequencing clash two weeks out. The **Lookahead** keeps the next few weeks in one live view, and routines watch the record for the signals that precede a slip — a supplier email moving a delivery date, an RFI that gates a pour, a subcontractor going quiet. When a schedule risk is flagged, it lands in Needs attention with the reason and the source — days before it would have surfaced at the weekly meeting. Video · \~2 min The landmine that didn't go off A supplier email slips a delivery by nine days; the routine flags the clash against the lookahead, the PM opens the flag from Needs attention, reads the cited reasoning, and re-sequences before it costs float. ## Turn a site instruction into a variation The superintendent directs extra work on site. That instruction is money — but only if it becomes a variation with a price and a paper trail before everyone moves on. Ask Alloovium to raise the variation from the instruction: it drafts the scope from the directive, pulls the relevant contract clauses on valuation and notice, and starts the record that every later claim will lean on. The variation lives on the project register with its status, value, and every linked document — so nothing directed on a Tuesday gets forgotten by claim day. Video · \~2.5 min Site instruction to variation in one sitting Starting from a scanned site instruction, asking the assistant to raise the variation, reviewing the drafted scope and clause references, and saving it to the register with the instruction attached. ## Chase deliverables automatically Half of project management is following up: the SWMS that hasn't arrived, the shop drawings promised last Thursday, the unsigned subcontract. A **routine** watches for the reply — and if it doesn't come, drafts the chase, waits for your approval, and sends it. Every chase and every response is on the record. You set it once in plain language: _“Every morning, check whether Calabria have returned the signed subcontract — if not, chase them and copy me.”_ Video · \~2 min The routine that waits and chases Creating a chase routine in plain language, the routine detecting no reply after two days, drafting the follow-up for approval, and the reply landing on the project record. ## The monthly report without the weekend The monthly report eats the last weekend of every month: progress, program, cost, safety, photos — assembled from a dozen sources. Alloovium drafts it from the project record: what was built (site diaries, meeting records), where the program sits (lookahead, flagged risks), where the money is (claims, variations register), grounded in the documents with citations you can check. You review and edit a draft instead of assembling from scratch — and download it as DOCX in your company template. Video · \~3 min Drafting the monthly report from the record Generating the monthly report, stepping through the drafted sections against their cited sources, editing the program narrative, and exporting the DOCX. ## Get across the project you didn't tender The job is won in pre-construction and handed to you in a one-hour meeting — and more often than not, you had nothing to do with the tender. What you inherit is a SharePoint link and a week of reading. Load the pile into Alloovium instead and ask the questions you'd otherwise dig for: what are we building, for how much, what did estimating allow, who's the client's rep, and what in this contract is unusual. Every answer is cited to the page it came from, so you're checking receipts rather than trusting a summary. The same trick onboards anyone who joins mid-job — a new engineer can question the entire record from day one instead of interrupting yours. Video · \~2 min The project you didn't tender A PM handed a won job blind — loading the tender pile, asking the first five questions a PM actually asks, checking a citation, and finishing with 'what's unusual in this contract' before the chair is warm. ## Mobilise the subs without the spreadsheet Month zero: you might be the only person on the project, with a thousand things to organise and, from every subcontractor, the same rule-based list — the signed contract back, licences, insurances — before anyone sets foot on site. Every gap is a burnt day on site or a variation from a sub who started under-informed. Raise the list as **Exchange** document requests instead of emails. Each item carries a state — sent, viewed, supplied, signed — the board shows which sub is stalling, and a routine chases the stragglers politely until the item arrives or a human needs to lean in. The checklist runs itself; you run the job. Video · \~2.5 min Mobilisation without the spreadsheet A just-won job: per-subcontractor document requests going out through Exchange, a sub uploading against them from a clean guest link, an expired certificate sent back revise-and-resubmit, the board's one red row, and the chase routine nagging so you don't. ## Survive the management-plan comment round You spend days writing the environmental or traffic management plan, and the client's reviewer always comes back with comments — sometimes a hundred of them — because that's their job. Answering them used to be weeks of Word-and-Outlook ping-pong, mostly re-finding facts that already exist in the geotech report, the approval conditions, and the spec. Hand the comment round to the record: responses draft themselves with citations, the comments that genuinely need an engineer come back flagged instead of buried, and a leftover-from-the-last-job check catches the wrong creek or the wrong council before the reviewer does — these plans are contractually required to be project-specific. Video · \~2.5 min A hundred comments on the management plan A returned management plan dense with reviewer comments — per-comment responses drafted from the project record with citations, two comments flagged as needing a real decision, the leftover references from another project caught, and round two going back through Exchange with a trail. ## Know what changed while you were away Two weeks of leave means two hundred emails and a site that moved without you. The project **Timeline** holds every event — documents received, decisions logged, variations raised, claims moved, flags fired — in one stream. Ask for the summary: _“What changed on this project in the last two weeks? What needs me first?”_ You get the cited digest in minutes: what happened, what was decided, and the three things that actually need your attention — instead of archaeology through your inbox. Video · \~2 min Back from leave in twenty minutes Opening the project timeline after two weeks away, asking for the what-changed digest, and clicking through the three flagged items that need a decision. ## Escalate a delay with an evidence pack When a delay becomes a claim, the difference between winning and wearing it is the evidence: the notices sent on time, the correspondence trail, the site records that show cause and effect. Ask for an **evidence pack** on the event and Alloovium assembles the chronology — every relevant document, dated and cited, in a structured bundle ready for the contract administrator or your lawyer. Because the record was captured as it happened — meeting records, site diaries, emails filed to the project — the pack is built from facts, not reconstruction. Video · \~2.5 min Building the evidence pack Requesting an evidence pack for a delay event, reviewing the assembled chronology with each entry citing its source document, and exporting the bundle. Looking for a different seat? See the use cases for [site managers](/en/docs/use-cases/site-manager), [safety & HSEQ](/en/docs/use-cases/safety-manager), [commercial managers](/en/docs/use-cases/commercial), [executives](/en/docs/use-cases/vp-cxo), and [subcontractors](/en/docs/use-cases/subcontractor). --- Canonical page: https://www.alloovium.com/en/docs/use-cases/pm --- ## Use Cases for Safety & HSEQ Managers Source: https://www.alloovium.com/en/docs/use-cases/safety-manager Use cases · Safety & HSEQ # Ten ways safety and HSEQ managers use Alloovium Non-conformances should surface before the audit, not in it. These are the ten workflows safety and HSEQ managers run in Alloovium every week — each with a short video showing the real thing. ## Catch it before the audit does The quarterly audit shouldn't be the first time anyone looks. The **Conformance** surface watches the record continuously: tickets and insurances approaching expiry, ITP sign-offs that never landed, findings still open past their due date — everything ranked by how close it is to becoming a finding. You open one screen each morning and see what an auditor would see three months from now — with time to fix it. Each item links straight to the record that's missing or expiring, so the gap and the fix live one click apart. Video · \~2 min The audit that found nothing Opening the Conformance surface at the start of the week, walking the expiring-credentials and open-findings lists, and clicking through to the record behind the worst item. ## Credentials chased before they lapse A lapsed white card or an expired insurance certificate is a stop-work waiting to be discovered. A **routine** watches tickets, licences, and insurances across the project roster, and when one nears expiry it drafts the chase to the holder — waiting for your approval before anything sends. You set it once in plain language: _“Every Monday, check for credentials expiring within 30 days — chase the holder and copy me.”_ Every chase and every renewal certificate that comes back lands on the record. Video · \~2 min The routine that renews the tickets Creating the credential-watch routine in plain language, the routine flagging an EWP licence 28 days out, approving the drafted chase from Needs attention, and the renewed ticket landing on the record. ## The audit pack in an afternoon External audit week normally means a fortnight of folder archaeology: pulling inductions, calibration certificates, ITP records, and training registers into a binder nobody will open twice. Ask Alloovium to assemble the **evidence bundle** against the audit scope and it builds the pack from the project record — every entry cited back to its source document. When the auditor asks _“show me the evidence for this”_, the citation is the answer: click it and the source opens with the passage highlighted. The pack is built from the record, not reconstructed for the occasion. Video · \~2.5 min The audit pack in an afternoon Requesting an evidence bundle against the audit scope, stepping through the assembled sections with each record citing its source, and exporting the pack before the auditor's flight lands. ## SWMS reviewed against the plan A subcontractor's SWMS arrives the afternoon before the work starts, and the review happens in the gap between two other jobs. Ask Alloovium to check it: _“Review Calabria's SWMS for the facade access work against our safety management plan and the actual activity.”_ The gaps come back cited — the high-risk work the SWMS doesn't address, the controls the safety management plan requires that the document skips, the mismatch between the stated methodology and what the lookahead says is actually happening. You send back a specific list instead of a vague “please revise”. Video · \~2.5 min A SWMS that doesn't match the work Asking the assistant to review an uploaded SWMS against the safety management plan, reading the cited gap list, clicking through to the plan clause each gap references, and drafting the return email. ## Root cause with receipts A non-conformance gets raised and the root-cause line on the form fills itself in with the convenient answer: operator error, one-off, closed. The actual cause is usually upstream — in a site diary from three weeks earlier, a delivery docket, an ITP record that was signed under pressure, an email that changed the spec. Ask Alloovium to trace it: _“What led to NCR-014 on the level 3 slab?”_ The answer comes back as a cited chain through the diaries, ITP records, deliveries, and correspondence — the actual cause, with the documents that prove it, so the corrective action fixes the system instead of blaming the operator. Video · \~2.5 min The root cause nobody wrote down Asking the assistant to trace an NCR through the record, following the cited chain from the defect back to a substituted product on a delivery docket, and attaching the trail to the NCR. ## ITP records complete before you cover up Once the work is buried, the missing witness-point signature is missing forever — and it resurfaces at handover, or worse, in a dispute. Before a lot closes, ask Alloovium to verify the ITP records: every hold and witness point checked against the lot register, with the sign-offs that exist cited and the ones that don't listed. The gap list arrives while the work is still open and the signatory is still on site — which is the only time it's cheap to fix. Video · \~2 min Checked before the concrete pour Asking for an ITP completeness check on a lot the day before cover-up, reviewing the two missing witness-point sign-offs against the lot register, and chasing them while the engineer is still on site. ## The incident chronology from the record After an incident, the pressure is immediate: the regulator, the client, the insurer all want the sequence of events, and memory is already drifting. Ask Alloovium to assemble the chronology from the site records — diaries, prestart records, inductions, correspondence, deliveries — and it comes back as a dated timeline with every fact linked to its source. The chronology is built from what was captured as it happened, not what people recall under stress a week later. When the investigator asks how you know, the citation answers. Video · \~2.5 min Seventy-two hours, reconstructed in twenty minutes Requesting the incident chronology for a plant contact event, stepping through the assembled timeline with each entry citing its site record, and exporting the sequence for the notifier. ## Who's inducted, who isn't The person on site who was never inducted is invisible right up until something goes wrong. Alloovium checks site access against the induction and credential records continuously — who signed in, who's inducted, whose tickets cover the work they're rostered on — and produces the gap list. Set it as a routine and the list is waiting every morning: _“Each morning, compare yesterday's site sign-ins against the induction register and flag anyone without a current induction.”_ The conversation with the subcontractor happens at 7am, not after the incident. Video · \~90 sec The morning gap list Opening the morning induction-gap flag from Needs attention, seeing two sign-ins with no matching induction record, and clicking through to the crew's credential records before the prestart. ## Toolbox talks that hold up The toolbox talk happened, the hazard was discussed — and the proof is a signature sheet that half the crew missed on their way past. Record the prestart instead, and Alloovium turns it into a meeting record: a summary of what was covered, the full transcript beneath it, and the attendees identified from the conversation itself. When the question comes — _“was this crew briefed on the exclusion zone before the lift?”_ — the answer is a record with the words that were said and the people who were there, not a sheet of scrawls. Attendance you can prove. Video · \~2 min The prestart that became evidence Recording a seven-minute prestart on a phone, opening the finished meeting record — summary, transcript, attendees identified from the conversation — and asking who was briefed on the exclusion zone. ## Lessons that change behaviour Every business has the PowerPoint: lessons learned, presented once, filed forever. The lesson that would have prevented this month's near-miss is in it — captured two jobs ago, seen by nobody who needed it. Alloovium captures lessons on the project record, and they resurface on the next job at the moment they're relevant: when the same activity appears in the lookahead, when a similar SWMS arrives, when the assistant answers a question the lesson bears on. The lesson stops being a slide and becomes part of how the next crew plans the work — cited, in context, at the point of decision. Video · \~2 min The lesson that showed up on time Capturing a lesson off a near-miss on one project, then watching it resurface on the next — flagged against the lookahead when the same scaffold sequence appears, cited in the planning answer. Looking for a different seat? See the use cases for [project managers](/en/docs/use-cases/pm), [site managers](/en/docs/use-cases/site-manager), [commercial managers](/en/docs/use-cases/commercial), [executives](/en/docs/use-cases/vp-cxo), and [subcontractors](/en/docs/use-cases/subcontractor). --- Canonical page: https://www.alloovium.com/en/docs/use-cases/safety-manager --- ## Use Cases for Site Managers Source: https://www.alloovium.com/en/docs/use-cases/site-manager Use cases · Site Managers # Ten ways site managers use Alloovium The site manager runs the job from the ute and the deck, not a desk — and what changed on site has to reach the record before knock-off, or it never does. These are the ten workflows site managers run in Alloovium every week, each with a short video showing the real thing. ## Phone the site diary in from the ute The diary is the record that wins disputes — and the task that loses to knock-off every single day. Alloovium calls _you_: a scheduled call or SMS at the end of the shift, and the AI walks the diary questions — weather, labour on site, plant, materials, deliveries, incidents. Mention the crane breakdown mid-answer and wander off onto tomorrow's pour, and it captures both, then brings you back to the question it was asking. By the time you're out of the car park, a structured site diary is on the project record — dated, complete, and searchable — instead of a promise to write it up tonight. Video · \~2.5 min The diary that calls you at knock-off Taking the 4:30pm scheduled call in the ute, answering the diary questions with one off-track tangent about a late delivery, then opening the finished structured diary on the project record. ## Check the detail against the current drawing The formworker is asking about a detail and you're holding a print that might be two revisions old. Alloovium's drawing intelligence reads the title blocks and revision tables across the drawing set, so the register knows which revision of every sheet is current and what each issue changed. From the phone: _“Which revision of S-201 is current, and what changed between Rev C and Rev D?”_ — and the answer comes back with the revision notes and the sheet itself, before anyone builds off the wrong issue. Video · \~90 sec Two revisions, one truth Asking the project ask bar which revision of a structural sheet is current, reading the cited what-changed answer, and opening the drawing to the revised detail. ## What does the spec say about this pour? The agitator is twenty minutes out and someone raises curing. The spec is a 400-page PDF on a laptop in the site office — or it's in your pocket. Ask the assistant, scoped to the project: _“What does the spec require for curing this slab — and what's the test frequency?”_ The answer streams back with numbered citations into the specification — curing method, tolerances, sampling rates — and one tap opens the clause with the passage highlighted. You're quoting the document at the deck, not recalling it. Video · \~90 sec The spec answers before the truck arrives Asking the curing and test-frequency question from a phone on the deck, the cited answer streaming in, and tapping a citation to the highlighted spec clause. ## Raise the RFI before it blocks the work You've found a clash — the penetration lands on a beam and the drawings disagree. The old workflow is a photo on your phone, a mental note, and an RFI written from memory two days later. Instead: photograph the clash, dictate the question on the spot, and ask Alloovium to raise the RFI. The RFI is drafted from your words, the photos attached, and it's logged on the project register with a number and a status — so the question is with the engineer before you've walked off the deck, and the crew isn't standing around waiting on it. Video · \~2 min From clash to logged RFI on the deck Photographing a services clash, dictating the question into the assistant, reviewing the drafted RFI with the photos attached, and saving it to the register. ## Know what's coming next fortnight The crew works today; the site manager works two weeks ahead. The **Lookahead** holds the next fortnight in one live view — the pours, the crane days, the inspections, the deliveries that gate them — with clashes and dependencies flagged while there's still time to move something. When the scaffold strip lands on the same day as the facade delivery, the clash is flagged with the reason — days before the two crews meet at the hoist arguing about who was booked first. Video · \~2 min The fortnight in one view Scrolling the lookahead across the next two weeks, opening a flagged sequencing clash between the scaffold strip and a delivery, and shifting the row before it costs a day. ## Hold points before you cover it up Nothing on a job is more expensive than pouring over a hold point. The ITP's hold and witness points are surfaced against the activity itself — before the reinforcement gets covered, you can see exactly which points apply, who has to witness them, and how much notice they're entitled to. _“What are the hold points before we pour the level 3 slab, and who needs to witness?”_ — the answer comes off the ITP with the notice periods cited, so the superintendent's representative is booked in time instead of being an excuse to stand the pour down. Video · \~2 min The hold point that didn't get poured over Asking for the hold and witness points ahead of a slab pour, reading the cited ITP answer with witnesses and notice periods, and opening the ITP to the row. ## The prestart that writes itself up The toolbox talk happens every morning; the record of it happens sometimes. Record the prestart on your phone and Alloovium produces the meeting record — a summary at the top, the transcript beneath it, and the attendees identified from the conversation itself. The actions raised — the missing edge protection, the exclusion zone for the crane lift — are extracted with the verbatim quote each came from and land in a review queue, so what was said at 6:30am becomes a record by smoko, not a memory by Friday. Video · \~2.5 min From toolbox talk to meeting record Recording a morning prestart, opening the finished record with attendees identified from the conversation, and approving two extracted actions from the review queue. ## Dockets without the data entry Trucks run all day, and the day ends with a stack of paper dockets that someone — usually the youngest engineer on the job — spends the evening re-keying into a spreadsheet: once for productivity, again for who gets paid by the truckload, again for the tip fees. Instead, photograph the stack and email it to the project's own address. A routine reads each docket — truck, load, quantity, time — into a register the routine maintains, with the docket image attached to every row. What moved this week and what's owed becomes a question the register answers, cited back to the paper — and a disputed load is a click, not an argument. Video · \~2 min Fifty dockets, zero data entry Photographing the day's docket stack at the gate, emailing it to the project's own address, the dockets routine extracting each one into the project's docket register with the source image attached, and the week's material and truckload payments answered off the register. ## The latest revision, always Every builder has a story about the wall built off the superseded drawing. Document control in Alloovium means a superseded sheet can't sneak onto the deck: when a new issue arrives, the old revision is superseded on the record, and every answer and every register points at the current issue. Anyone on the crew can ask — _“Give me the current issue of A-301”_ — and get the live sheet, not the one that was current when the print run happened. The drawing on the phone is the drawing of record. Video · \~90 sec The superseded drawing that never reached the deck A revised architectural sheet arriving on the project, the old issue marked superseded on the register, then asking for the current issue from a phone and opening the live sheet. ## Capture the lesson while it's fresh The dearest knowledge on a job is the mistake you just paid for — and it usually walks off site in someone's head. While it's fresh, dictate what went wrong and why: the blockwork sequence that trapped the plumber, the pump that couldn't reach the far corner. The lesson joins the company knowledge base, tagged to its trade and context — and on the next job, when the same situation is taking shape, it resurfaces where it's relevant instead of living in a spreadsheet nobody opens. Video · \~2 min The lesson that made the next job Dictating a lesson from the ute after a blocked pour, reviewing the captured entry in the knowledge base, then seeing it resurface against a similar activity on another project. Looking for a different seat? See the use cases for [project managers](/en/docs/use-cases/pm), [safety & HSEQ](/en/docs/use-cases/safety-manager), [commercial managers](/en/docs/use-cases/commercial), [executives](/en/docs/use-cases/vp-cxo), and [subcontractors](/en/docs/use-cases/subcontractor). --- Canonical page: https://www.alloovium.com/en/docs/use-cases/site-manager --- ## Use Cases for Subcontractors Source: https://www.alloovium.com/en/docs/use-cases/subcontractor Use cases · Subcontractors # Ten ways subcontractors use Alloovium The subcontractor builds off what the head contractor sends and gets paid off what they can prove. These are the ten workflows subbies run in Alloovium — your own copy of every package, a stamped trail of every send and signature, and answers out of the contract you actually signed — each with a short video showing the real thing. ## The package lands with its paper trail A subcontract, an amendment, a drawing set — instead of an attachment buried in a thread, the package arrives through **Exchange** as a link. Open it and the document previews in the browser with the sender's note beside it, and a saved copy lands in your own workspace — yours to keep and file to the job, regardless of what happens to the email thread it used to live in. The activity rail shows the provenance from your side too: who sent it, when, and what note came with it — so “which version did they actually send us?” is a lookup, not an argument. Video · \~90 sec Receiving a package with its trail Opening an Exchange link as the recipient, previewing the document with the send note beside it, and filing the saved copy to the job — then reading the activity rail showing who sent it and when. ## Sign it where you received it The signature copy arrives for execution; instead of print-sign-scan, you sign it in place. The signature is bound to the exact content you were sent — hash and all — and the signed counterpart goes back on the same timeline the package arrived on. Both sides now hold the same record: sent, opened, signed, returned, each stamped with who and when. No more “the version I signed didn't have that clause.” Video · \~2 min Signed and returned, stamped Opening a signature copy from an Exchange link, signing in place, and sending it back — then the timeline showing send, open, signature, and return in order, each stamped. ## Read the amendment like a mark-up The amended subcontract comes back and it's 42 pages. The old workflow is reading it side-by-side with the last issue and hoping you catch everything. In Alloovium, the paragraph-level diff shows every change between the revisions — the notice period that quietly halved, the new bar on latent conditions, the reissued rates schedule. You price the contract they actually sent, not the one you remember — and the clause they hoped nobody would read on page 47 is the first thing you see. Video · \~2 min Every edit on page 47 Opening the paragraph-level diff on an amended subcontract received through Exchange, walking the changed clauses one by one, and landing on the notice-period change buried mid-document. ## Prove you sent it, prove they read it When a claim or notice is disputed, the head contractor has a register and you have a sent folder. Exchange puts you on equal footing: everything you send carries a stamped trail — sent, delivered, opened, signed — and everything you receive carries the same. The next time someone says the notice never arrived, the answer is a timeline entry with a date and a name on it, not a screenshot of your outbox. Video · \~90 sec The trail that ends the argument Opening the Exchange trail on a contested notice — sent, opened, and responded events in order with who and when — and sharing the timeline as the answer. ## One link and you're on the project The builder invites you to the project and you get one link. It lands you on a page that says who invited you and to what, you accept, and you're in — with access scoped to exactly what your trade needs through the project's groups. No IT ticket, no license audit, no week of onboarding. From that seat you see the documents you've been given — the current drawings, your subcontract, the specs that govern your work — and nothing you haven't. Video · \~2 min From invite link to first answer Clicking a project invite link, accepting from the landing page, arriving on the project with trade-scoped access, and asking a first question against the subcontract. ## Know what you signed up to Retention, dayworks caps, notice periods, defects liability, set-off rights — the answers are in the subcontract, and the subcontract is 42 pages plus schedules. Ask in plain language: _“How many days do I have to notice a delay, and what does the notice have to include?”_ The answer comes back cited to the clause, and one tap opens the passage highlighted — so you know your entitlements before you do the work for free, and you quote the contract back with a clause number attached. Video · \~90 sec The subcontract, on tap Asking a notice-period question against the signed subcontract, reading the cited answer, and tapping through to the highlighted clause. ## Submissions with a status, not a thread Shop drawings, samples, procedures — every submission runs through **Exchange**with a status you can see, instead of disappearing into a contract administrator's inbox. When it comes back — approved, approved as noted, revise and resubmit, or rejected — the outcome and the reviewer's note are on the record, and every round-trip accrues to the package's revision history. That history means you always know what you submitted, which revision, and what came back — the question “did the builder ever respond to our shop drawings?” has a screen, not a search. Video · \~2 min The submission you can see moving Submitting a shop-drawing package through Exchange, the revise-and-resubmit landing with its note, the corrected resubmission, and the approval — with the whole exchange readable off the revision timeline. ## The request that can't get lost The builder needs your insurances, your SWMS, your mix designs — and normally that ask lives in an email you meant to come back to. Through Exchange it arrives as a **document request**: a tracked ask with its own place to put the file, so responding is an upload, not a reply-with-attachment into the void. Both sides see the same state — requested, supplied, reviewed — so “we never got it” and “we definitely sent it” stop being competing recollections and become one record. Video · \~2 min The request answered in one upload A document request landing from the builder, the file uploaded straight into it, and the request's state moving to supplied — with the record visible to both sides. ## Dayworks the builder can verify Directed work happens on the day it happens; the record of it usually happens at claim time. Capture the daywork as it's done — the direction, the crew hours, the plant — and the sheet is dated, structured, and tied to the job from day one. When the claim goes in, the dayworks item is backed by records the builder's team can open and check — and sending the sheets across for countersignature through Exchange puts the request itself on the stamped record. Video · \~2 min The daywork sheet with receipts A daywork record captured on the day with crew hours and plant, then the register of sheets shared through Exchange for countersignature with the send stamped on the trail. ## The claim that can't be “never received” Security-of-payment regimes turn on dates: when the claim was made, when the response was due. Send the progress claim through Exchange and the date it went — and the date it was opened — are stamped on a trail both sides hold. If the response window passes quietly, you're not reconstructing evidence from an email thread; the chronology is already made, in order, with names on it. That's the difference between a claim and a case. Video · \~2 min The claim with a timestamp Sending a progress claim through Exchange, the sent and opened events landing on the trail, and reading the stamped chronology when the response date approaches. Looking for a different seat? See the use cases for [project managers](/en/docs/use-cases/pm), [site managers](/en/docs/use-cases/site-manager), [safety & HSEQ](/en/docs/use-cases/safety-manager), [commercial managers](/en/docs/use-cases/commercial), and [executives](/en/docs/use-cases/vp-cxo). --- Canonical page: https://www.alloovium.com/en/docs/use-cases/subcontractor --- ## Use Cases for Executives Source: https://www.alloovium.com/en/docs/use-cases/vp-cxo Use cases · Executives # Ten ways executives use Alloovium You can't attend every meeting on every job — but the record can. These are the ten ways executives use Alloovium to see margin and schedule risk across the whole book, each with a short video showing the real thing. ## The one project that needs you today Ten projects, ten weekly reports, and every one of them says “on track” — until the one that wasn't costs you a quarter. Alloovium scans across the book and puts the real exceptions on top: schedule flags fired by routines, notices sitting unanswered past their clock, claims that stopped moving. Each exception cites the source it came from — the email, the register entry, the lookahead clash. Instead of reading ten summaries written to reassure you, you open one view and see the two things that actually need an executive today — with the evidence one click away. Video · \~2 min Scanning the book for the real exceptions Opening the portfolio view across ten projects, working the exceptions top to bottom — a schedule flag, an unanswered notice, a stalled claim — and clicking each one through to its cited source. ## The cash position without chasing PMs The question that matters at month end — _“where is our cash?”_ — is usually answered by chasing five PMs for five spreadsheets, each a week stale. In Alloovium, progress claims live on each project's register with their real status: what's been claimed, what's certified, what's paid, and what's stuck between the two. Roll that up across every project and you have the company's cash position off the record itself — _“What's claimed but not yet certified across all projects?”_ — without a single chasing email. Video · \~2 min Where the cash actually is Asking for the claimed-certified-paid position across the book, stepping through the claims register on two projects, and opening the one claim that's been sitting uncertified for nineteen days. ## The margin leak, surfaced Margin doesn't vanish in one bad decision — it leaks through the gaps: work directed on site that never became a variation, variations approved but never claimed, delay costs worn quietly because the notice window closed. The record holds all of it — the site instruction, the register, the claims — and Alloovium can cross-check them before the QS review ever gets scheduled. _“Which approved variations haven't been claimed? Is there directed work on any project without a variation against it?”_ — the answer comes back cited, project by project, while the money is still recoverable. Video · \~2.5 min Finding the money before the QS does Asking the assistant to cross-check directed work against the variations register, surfacing an approved variation absent from the last two claims, and opening the cited instruction it traces back to. ## The monthly board pack from the record The board pack is a week of somebody's month: chasing summaries, reconciling numbers, smoothing prose. Alloovium drafts each project's summary from its actual record — progress from site diaries and meeting records, program from the lookahead and flagged risks, commercial position from the claims and variations registers — every figure cited back to its source. Your job becomes what it should be: reviewing the pack, not assembling it. And when a director asks where a number came from, the citation answers before you have to. Video · \~3 min A board pack you review instead of write Generating the monthly project summaries across the book, checking the commercial figures against their cited register entries, editing one program narrative, and exporting the pack as DOCX. ## Schedule risk weeks earlier By the time a delay reaches the executive report, the float is spent and the options are expensive. On each project, the lookahead and the routines watch for the signals that precede a slip — a delivery date moving in a supplier email, an RFI gating a pour, a subcontractor going quiet — and flag them into Needs attention with the reason and the source. Rolled up across projects, those flags are an early-warning radar for the whole book. The blindside surfaces while there's still float to spend — not in next month's report, after it's already cost you. Video · \~2 min The radar across every program Reviewing the schedule flags rolled up across the portfolio, opening one routine alert to its cited supplier email, and reading the lookahead clash it would have caused three weeks out. ## Know what the AI did and who approved it The first question a client — or a board, or an auditor — asks about AI on their project is the right one: _“what did it actually do, and who signed off?”_ The admin activity ledger answers it. Every agentic action is recorded: what the AI did, which routine or ask triggered it, who approved it, and the source it acted on. That's governance you can put on the table, not a policy you can only describe. Drafts waited for approval, sends were authorised by a named person, and the ledger shows all of it — exportable, filterable, org-admin gated. Video · \~2 min The ledger you can show an auditor Opening the admin activity ledger, filtering to agentic actions on one project, tracing a sent chase email back through its approver and the routine that drafted it, and exporting the trail. ## Learn across tenders, not just jobs Every job teaches the company something, and most of it retires with the project team. When lessons and outcomes live on the record, you can query them across the whole company: which contract risks actually bit, which clients certify slowly and pay late, what the last three jobs of this type taught you about allowances and sequencing. _“What did we learn on our last three fit-out jobs about services coordination?”_ — asked at tender time, answered from the lessons register with citations, so the next bid prices the risk the last job paid for. Video · \~2 min Pricing the next job with the last job's scars Querying lessons across the company at tender review, reading the cited lessons from two comparable jobs, and pulling the client's certification history into the risk discussion. ## Every team runs the same play Growth breaks consistency: every new PM brings their own folder structure, their own report format, their own idea of what a variation record looks like. Company templates, the playbook, and user groups mean a new project starts with your standards already in place — the document templates your clients recognise, the routines your best PMs run, the access model your governance requires. A project kicked off in Alloovium starts from the company playbook, not a blank folder — so the tenth project runs like the best one, not like whoever set it up. Video · \~2.5 min A new project on company rails Standing up a new project from company templates, watching the standard routines and user groups apply, and generating a first document that lands in the company format without anyone configuring it. ## Show the client the receipts Client trust is built or burnt in the gap between what you assert and what you can show. When the client questions progress, a cost, or a delay, you can share the cited answer itself — every claim in it linked to the document it came from — instead of a carefully worded reassurance. And every document exchanged with them carries its Exchange audit trail: sent, received, read, signed, dated. Disputes get shorter when both sides are looking at the same record. Relationships get longer when the client learns your answers always come with receipts. Video · \~90 sec Trust built on the record Sharing a cited answer to a client's cost question, then opening the Exchange trail on a contested transmittal — sent, read, and signed events with dates — and sending the client the link. ## A new team productive in a week An acquisition, a new division, a new region — a whole team landing at once is where onboarding usually collapses into months of “ask someone who knows.” In Alloovium they land in the workspace with user groups granting exactly the right access on day one, and the company knowledge base already answering the questions they'd otherwise queue outside someone's office to ask. _“How do we run variations here? What's the delegation for approving a subcontract?”_ — cited answers from your own standards, so the new team learns the company from the record while the old team keeps running their jobs. Video · \~2 min Landing a whole team at once Inviting a new division through groups, a new manager accepting and landing with the right project access, and asking their first questions of the company knowledge base — each answer cited to the standard it came from. Looking for a different seat? See the use cases for [project managers](/en/docs/use-cases/pm), [site managers](/en/docs/use-cases/site-manager), [safety & HSEQ](/en/docs/use-cases/safety-manager), [commercial managers](/en/docs/use-cases/commercial), and [subcontractors](/en/docs/use-cases/subcontractor). --- Canonical page: https://www.alloovium.com/en/docs/use-cases/vp-cxo