PageFinderv4.2.11.22Sign in

Blog

Copilot can’t see my old scans — what to do

Why an assistant that works on your Microsoft 365 files draws a blank on image-only PDFs, what fixes the next scan, and what reads the archive you already have.

November 24, 2026 · 5 min read · PageFinder team

PAIDSaved from Wordtext layer · search reads every wordOut of the scannera picture · search sees the file name

A common surprise: you ask the assistant in your Microsoft 365 apps to summarise a contract in OneDrive, and it answers that it cannot find the information, or gives you a summary of the file name. The file opens fine. The difference is that it is a scan.

What the assistant works on

As of this writing, Copilot and the assistants built into Microsoft 365 work on the text of the files you point them at or have open: Word documents, emails, PDFs that carry a text layer. That covers most of what was created in the last few years. It does not cover a PDF that is only a picture of a page — a scanned agreement, a faxed order, a photographed receipt — because there is no text in the file to work on. Microsoft's features change often, so check what your tenant has enabled; the principle does not change: no text layer, nothing to read.

Check which files are affected

Open a suspect PDF and try to select a line with the mouse. Words that highlight mean a text layer; a box drawn around the page means a scan. In most archives the scans are the signed, stamped and older documents, which are the ones a lawyer or an accountant asks about.

Fixing the next scan

  • Microsoft 365 offers an OCR service for OneDrive and SharePoint that an administrator can turn on, billed per page, which adds a text layer to new uploads. Check your plan and admin center; it is not available on personal OneDrive.
  • Scanner software and Acrobat can "recognise text" before a file is filed.
  • Either way, the result is only for files processed from now on.

Reading the archive you already have

The archive is the problem: thousands of scans nobody will open and re-save, and re-saving changes files that other people's links and records point at. The way through is a tool that reads the page image where the file is, without changing it.

That is what PageFinder does with OneDrive. It connects read-only, reads every file in the folders you choose — scans from the page image, the rest from their text — and catalogues each one: kind, the date it is about, the companies and people named, a summary. Then a question like "what does the 2009 warehouse lease say about renewal?" is answered from the scanned page, with the page cited so you can open the original. The files are not copied or moved, and nothing in OneDrive is re-saved. Use the built-in assistant for the document in front of you, and this for the collection, scans included. How it works with OneDrive.