# Cited Document Intelligence Stack

> A traceable document AI stack for parsing complex files, extracting decision fields, and producing cited briefs for human review.

- Canonical: https://gptnavi.com/stacks/cited-document-intelligence-stack
- Estimated monthly cost: $0-$500/month plus document volume
- Last materially updated: 2026-08-24
- Best for: Consultants, Operations teams, Procurement teams, Research teams

## Problems this stack solves

- Slow document review
- Broken table extraction
- Untraceable summaries
- Missing fields
- Conflicting source evidence

## Recommended tools

### Primary parsing

Recommended: Cohere Parse, LlamaParse

Why: Preserve useful layout, tables, forms, and page references before downstream AI work.

### Structured extraction

Recommended: Reducto, Unstructured

Why: Extract only the fields needed for a comparison and flag uncertainty instead of guessing.

### Cited answering

Recommended: PageIndex, Perplexity

Why: Keep important conclusions tied to the source set and visible context.

### Review workspace

Recommended: Notion, Airtable

Why: Track file versions, extracted fields, owners, questions, and final decisions.

## Beginner setup plan

1. Start with one low-risk document family.
2. Require page or section references for material claims.
3. Keep original files available to reviewers.
4. Escalate legal, financial, and compliance interpretation to qualified people.

## Included workflows

- https://gptnavi.com/workflows/complex-documents-to-cited-decision-brief
- https://gptnavi.com/workflows/multi-agent-research-to-decision-brief
- https://gptnavi.com/workflows/governed-agent-with-connected-tools