Case management for documents
A case is a unit of work in Document Blueprint — one invoice, one permit, one contract. Cases hold files, extracted data, notes, and generated outputs, all grouped under a single identifier.
Last updated
A case is a unit of work in Document Blueprint — one invoice, one permit, one contract. Cases hold files, extracted data, notes, and generated outputs, all grouped under a single identifier.
If a workspace is the container and a template is the blueprint, a case is the instance. Files arrive, data gets extracted, workers add notes, outputs get generated — the case is the anchor that holds all of it together.
Core concepts
The case identifier
Every case has one field that names it uniquely within its workspace: the case identifier. For an invoice workspace, that's typically Invoice Number. For permits, Permit Number. The identifier is defined in workspace config and drives merge — when a new file arrives carrying the same value, it attaches to the existing case instead of creating a new one.
What a case holds
- Files — every document ingested or uploaded for this work item
- Case data — fields grouped by where each value came from (User Input, Extracted, Computed)
- Activity — notes, form submissions, generations, system events
- Outputs — PDFs produced by autofill against templates
- Team and Clients — workspace members and external collaborators, scoped by file category
Cases live inside workspaces
A case is always scoped to one workspace. Two workspaces can hold cases with the same identifier value without conflict. Switching workspaces in the dashboard sidebar changes which cases you see.
Walkthrough: a case's lifecycle
Picture an accounts-payable workspace. A new invoice from Acme Supply Co. arrives via Gmail.
- A Gmail automation watches a label and pulls in new attachments.
- The automation categorizes the PDF as
invoicebased on its filters. - AI extraction reads the PDF, finds
Invoice Number = INV-2026-0412andVendor Name = Acme Supply Co., and creates a new case with that identifier. - The case appears in the inbox with all extracted fields populated.
- A worker opens the case, double-checks the amount, and adds a note.
- The worker opens the autofill review modal, picks the AP cover sheet template, and a generated PDF lands in the case's outputs.
Common patterns
Identifier merging
When a second invoice with the same number arrives (a re-send, a corrected version), it merges into the existing case automatically. The new file is attached, extraction runs again, and the timeline records the merge.
Editing case data inline
Open any case and you can edit field values directly in the case-edit grid — no template trip required. At the end of the grid sits a dashed + Add field tile: click it, type a name like Tax ID and a value, and the field is added to this case. If a field with the same name already exists in the workspace, the value writes to that field instead of creating a duplicate. New fields auto-surface on this case and become available for matching on others.
Case-scoped collaborators
Add specific people to a case via the Team panel (workspace members) or Clients panel (external). Both panels scope by file category — pick which categories the collaborator can see; that drives files, activity, and inbox chips. Clients can also be given specific form templates to fill from the portal.
Moving cases between mailboxes
Cases live in mailboxes — Inbox, Archive, Trash, or any custom mailbox the workspace has defined. There's no automatic "stage" engine that moves cases when fields change; mailbox moves are user-driven.
- From the case detail panel — open the
⋯overflow menu in the case header and pick a destination from Move to mailbox. - From the case table — bulk-select cases and use the table's mailbox actions to move several at once.
- At creation — new cases land in their workspace's Inbox mailbox.
Troubleshooting
Two files about the same thing created two separate cases. The identifier value didn't match. Open either case, edit the Identifier field on one to match the other — the cases reconcile atomically.
A file ingested but no case appeared. Check the file's category. If it isn't categorized into something the workspace knows about, case creation doesn't trigger.
A case is missing fields I expected. Check the template that ran extraction. Fields have to be declared on a template — the AI doesn't extract anything that wasn't asked for. You can always add a one-off via + Add field on the case-edit grid.
An old case got an unexpected new file. Identifier merging is automatic. The new file's identifier matched an existing case's identifier — open the file's preview to see the extracted value.