Docs / Reference

Troubleshooting uploads & extraction

A document failed to extract

Retry it from its row on the Documents tab. Persistent failures are usually the file itself: password-protected PDFs, pure-image scans at very low quality, or files that aren’t actually PDFs. Re-scan at a readable quality and upload again.

The document classified as the wrong type

Classification follows the document’s content. A “chart note” that reads like an encounter visit note will classify as F2F notes — and the checklist may then say a chart-notes requirement is unmet. Check the Extraction tab to see what the AI read; the fix is usually uploading the actual clinical report (lab report, sleep study, office note) rather than a cover page or summary.

The checklist says something’s missing that I uploaded

Three usual causes, in order:

1. Still extracting — the checklist re-evaluates as documents complete; give it a minute
2. Classified differently than you expected — see above; the requirement names which document type satisfies it
3. The content really is absent — e.g., chart notes that mention a sleep study without the study’s values. The requirement’s citation says exactly what reviewers look for

A “Needs review” item won’t go green

That’s by design — some judgments (borderline clinical values, ambiguous dates) are flagged for a human rather than auto-decided. Review it against the cited source; it doesn’t block submission.

Case/document limits

Cases accept up to 20 documents and a bounded number of extraction runs — real cases never hit these; if you somehow do, contact us.

Still stuck?

Email hello@affirmdme.com or use the contact form — include the case ID (from the case URL) and we can see the same activity trail you can.