Document Overview

Documents are at the heart of Campbooks. Every email attachment becomes a document that's classified, analyzed, and tracked through a review workflow — so the AI does the filing and you stay the approver.

Two status axes

A document has two independent statuses, which is what keeps the AI and the human cleanly separated:

• AI status — how far automated processing has got: pending → processing → completed (or failed).
• Review status — the human decision: pending → approved or rejected.

The AI never approves anything. It classifies and extracts; a person reviews and signs off every document.

Axis	Values	Set by
AI status	pending, processing, completed, failed	The AI pipeline
Review status	pending, approved, rejected	You

Document lifecycle

1. Ingested — an email attachment arrives and a document is created (ai_status: pending).
2. Processing — AI classifies the document and extracts its fields (ai_status: processing → completed).
3. Ready for review — it appears in your dashboard awaiting a decision (review_status: pending).
4. Approved / rejected — you sign it off; approval records you as the reviewer.
5. Exported — optionally push it to Google Drive, Zoho WorkDrive, or Notion. See Integrations.

Document types

Campbooks recognizes common types out of the box — and you can add your own:

• Invoices
• Receipts
• Contracts
• Statements
• Tax documents
• Reports
• Forms
• Correspondence

During onboarding, Campbooks can even suggest document types based on a short description of your organization. See Document Classification for how typing and extraction work, and how the system learns from your corrections.

Extracted fields

For recognized types, the AI pulls out structured data — vendor name, amount and currency, document date, invoice number, and any custom fields defined on the document type. Extracted values are editable: correcting a field never changes the review decision.

Search and filter

Documents are workspace-wide (every member sees them) and searchable by:

• Full text (Postgres + pgvector)
• Document type
• AI status and review status
• Date range
• Source email account
• Tags