Hi Stephan Bren,
Thank you for raising your issue.
I tried to reproduce the issue you described using a SharePoint-based Copilot Agent connected to a document library containing a similar company benefits PDF (with info like holidays and vesting schedule). In my testing, the Copilot Agent was able to correctly extract and respond to simple questions, and it accurately cited the document contents. I'll attach a screenshot showing a working example from our environment for reference.
here are a few things worth checking-
- PDF Text Accessibility
- Agent Description
- Permissions
- Crawling and Indexing Delay
Thanks,
Keshav Keshari
*************************************************************************
If the response is helpful, please click "Accept Answer" and upvote it. You can share your feedback via Microsoft Teams Developer Feedback link. Click here to escalate.