Audit evidence review that keeps documents connected to the sample

Match documents before drawing conclusions

Document matching is a distinct audit step. A page can contain a relevant amount or account number without supporting the exact transaction, period, taxpayer, approval, or criterion under review. GovAuditAI lets the auditor review and adjust page-to-folio assignments before analysis.

The AI receives extracted text together with assigned page images. Text extraction remains important for amounts and identifiers, while page images provide layout and visual context. The workflow does not rely on vision alone for numeric OCR.

• Import spreadsheet sample identifiers
• Review extracted PDF pages
• Assign or remove pages for each sampled item
• Preserve evidence references with the resulting observation

Treat missing documentation as a review condition

Missing documentation is not automatically proof that a transaction is invalid. It means the available package does not contain the support needed for the stated test. That distinction matters when drafting requests and communicating with the auditee.

GovAuditAI separates missing evidence from verified facts and potential exceptions. Auditors can classify the item as Missing Documentation, request additional support, or override the AI suggestion with a documented reason.

Build a reviewable trail

Each AI result and auditor action can be recorded with the classification, action, override reason, and timestamp. Exports carry the final auditor decision rather than silently substituting the model’s initial suggestion.

That structure helps a reviewer retrace the path from sample to source page, observation, follow-up request, and final disposition.