Statuses & what they mean

Every document carries two status values, each tracking a different question. Knowing which one answers which question saves a lot of confusion later.

The Documents view shows both as separate columns. The Approval Status filter strip filters by the human-review status; the Document Status column shows the technical-pipeline status.

The two tracks

Track	Question it answers	Values
Document Status	Has the system finished its work on this?	Needs action, Processing, Done, Exported
Approval Status	Has a human signed off on this?	Pending, Validated, Approved, Rejected, Duplicate

They're independent. A document can be Done (extraction finished) but still Pending (no human has touched it). It can be Exported (sent to the ERP) and Approved (signed off). It can be Done and Rejected (extracted fine, but rejected on business grounds).

Document Status — the technical track

Set automatically by the system. You don't change it manually — Recognito moves the document through these states as the work proceeds.

Needs action

The document exceeded the project's auto-scan page threshold. Recognito holds off on extraction until you open the document, click Process, and pick which pages to actually scan.

This guards against accidentally running a 200-page PDF through OCR. The default threshold is 5 pages; admins can raise or lower it.

Processing

Extraction is running. OCR plus machine learning is reading the document and populating field values. This usually takes seconds to a couple of minutes depending on document complexity.

You can open a document in Processing but the fields haven't populated yet — wait for Done.

Done

Extraction finished and the project's Initial_Workflow has run. The document is ready for human review.

This is the longest-lived Document Status — documents stay Done for as long as the human-review track keeps them there. A document can be Done and Pending, Done and Validated, or Done and Approved, depending on where it is in the approval track.

Exported

The document has left Recognito — either an outbound integration delivered it, or someone manually marked it as exported.

Exported is not the same as Delivered

A document can be marked Exported manually via the bulk action Mark as Exported, even if nothing was actually pushed downstream. This exists for orgs that export out-of-band (printing for a paper archive, pasting into accounting software) and want to track which documents they've already handled.

Approval Status — the human-review track

This one moves through human action. The states track where a document sits in the review and approval pipeline.

Pending

The default Approval Status for a freshly extracted document. No human has signed off yet. The assignee opens the document, verifies the data, and either validates, rejects, or — in small orgs without separate approver groups — approves directly.

Validated

The assignee has signed off on the data. The document is ready for an approver to review.

If the project has Approver Groups configured, the document waits in Validated until the configured approvers act. If no groups are configured, the assignee can move straight to Approved.

Approved

The final sign-off has happened. The document is ready to leave Recognito — outbound integrations fire, the data flows to wherever you've configured it to go.

Approved documents can still have their Document Status change (to Exported) but their Approval Status stays at Approved.

Rejected

An assignee or approver decided the document shouldn't move forward — wrong data, wrong document type, business reasons. The document stays in the project for audit purposes but doesn't export.

Reversing a rejection requires the Override permission. Default users can reject but can't un-reject — they need an admin's help.

Duplicate

The system flagged this document as a duplicate of one already in the project, based on the field-value duplicate checker your admin configured (typically invoiceNumber plus vendorTaxId).

Two different duplicate paths

File-hash duplicates (byte-for-byte identical files) are rejected at upload and never appear in the Documents view at all. Field-value duplicates — what you see flagged as Duplicate here — get saved, extracted, and surfaced for inspection. They cost you a page quota; the file-hash path doesn't.

If a Duplicate is in fact a real document (false positive on the duplicate check), the assignee can return it to Pending and validate it normally. If it's a confirmed duplicate, reject it.

The state diagram

   UPLOAD
     │
     ├──► (file-hash duplicate) → rejected at upload, never appears
     │
     ▼
   Document Status: Needs action ──(user picks pages)──┐
                                                        ▼
                                                  Processing
                                                        │
                                                        ▼
                                                      Done
                                                        │
                                                        ▼
                                                  (assignee opens in Studio)
                                                        │
   Approval Status: Pending ──► Validated ──► Approved ──► Document Status: Exported
                       │             │
                       └─────────────┴──► Rejected
                       │
                       └──► Duplicate (field-value match detected)

The two tracks move in parallel but on different beats. The system moves the Document Status track; people move the Approval Status track.

What to do at each status

If a document is…	The next action is…
Needs action	Open it, click Process, pick the pages to scan
Processing	Wait — the engine is still running
Done + Pending (and you're the assignee)	Open it, verify the fields, click Validate
Done + Validated (and you're an approver)	Open it, sign off — click Approve or Reject
Done + Approved	Nothing — wait for export
Done + Rejected	Investigate. If it should move forward, ask an admin to override
Done + Duplicate	Check if it's actually a duplicate. If yes, reject. If no, return to Pending and validate normally
Exported + anything	The work is done; this is the archived state

What's next

• Finding & filtering documents — narrow the table to whichever statuses you care about.
• Bulk actions — what you can do to multiple documents at once.
• How Recognito works — the broader mental model that puts both status tracks in context.