Either the upload itself is refused, or the file uploads fine but the AI check reports it couldn't read the document.
The upload is refused
- The browser could not send the file. You see a message saying nothing was uploaded. The request never left your machine, so nothing was stored. This usually means the file sits in a cloud-synced folder and has not been downloaded to your device. Make sure the file is fully downloaded locally, or copy it to your desktop, then try again.
- File too large. Files are capped at 25 MB in the KYB Manager and the Document Library. Compress the PDF or split it.
- Unsupported type. The KYB Manager accepts PDF and image files (the library also accepts Word). Convert anything else to PDF first.
AI verification says it couldn't read the document
The document uploaded, but the automated check couldn't extract usable text from it. Likely causes, in order:
- Low-quality scan. Blurry, skewed, or low-resolution scans defeat reading. Re-scan at a higher quality, or photograph the original in good light.
- Image-only scan. A scanned PDF with no text layer is read with text recognition first (verification, AI review, review tables, DDQ import and Copilot indexing all do this, up to the first 12 pages). A scan that still cannot be read is usually too faint or too low-resolution for recognition: upload a clearer original, or a digitally issued copy if one exists.
If a better copy doesn't exist
A human review can always pass a document the AI couldn't read. Leave the file in place and let your reviewer (or the platform review step, where your workspace uses one) look at it directly; the AI verdict is a check, not a gate.