Source and evidence review
Evidence and Source Verification for AI-assisted documents.
A review layer for proposals, donor reports, tenders, evaluation reports and investment memos where every factual claim needs a source trail.
Claim registerEvery factual claim extracted
Source trailEvidence linked and flagged
Audit packReview notes and AI-use disclosure
Audit pack
AI can make writing faster. It does not make claims true.
Claim extraction
Factual claims are pulled out of the document and listed for review.
Source register
Each claim is linked to a source, with flags for weak, outdated or missing evidence.
Review trail
AI-use notes, reviewer decisions and unresolved risks are documented for the final file.
How it works
A source-controlled layer between AI drafting and final submission.
1
ExtractWe identify factual claims, figures, external references and unsupported statements.
2
VerifyClaims are matched to sources and flagged when evidence is weak or missing.
3
PackageThe team receives a source register, risk notes and a review-ready evidence pack.
Workflow scenarios
Where evidence verification changes the quality of AI-assisted work.
Grant proposal
Strong claims need a source trail before submission.
Issue
The proposal contains statistics, context claims and needs statements created from mixed research inputs.
Review
Extract claims, match each to a source and flag weak, stale or missing evidence.
Output
Claim register, citation trail and unresolved-risk list for the proposal lead.
Donor report
Narrative reporting must stay tied to project evidence.
Issue
Drafted narrative sounds convincing but does not clearly connect to activities, indicators and source documents.
Review
Map each material claim to evidence and separate verified points from interpretation.
Output
Evidence-backed report pack with reviewer notes and AI-use disclosure support.
Memo or report
AI-assisted analysis needs reviewer-visible assumptions.
Issue
A memo combines factual claims, interpretation and recommendations without a clear workpaper layer.
Review
Separate facts, sources, assumptions, limitations and human judgement points.
Output
Audit pack for client, evaluator, investor or internal leadership review.
Best fit
Use it when the document will be read by funders, evaluators, clients or investors.
- Grant proposals and donor reports with many factual claims.
- Evaluation or research reports that cite external evidence.
- ESG, due diligence or investment materials prepared with AI support.
- Teams that need AI-use disclosure support and source discipline.
Turn AI-assisted documents into reviewable workpapers.
We help your team keep claims, citations and AI use traceable.