Unable to Access PDF Files from SharePoint Folder via Copilot Studio

(3) Share

Report

Posted on by TS-15051642-0

Hello everyone,

I'm working on building a Copilot agent in Microsoft Copilot Studio that should access and retrieve information from PDF files stored in a SharePoint folder. The goal is for the agent to search and extract data from these PDFs based on the date present in the filename (e.g., "13.05.2025.pdf") and respond to user questions in Microsoft Teams accordingly. The PDFs are stored in a SharePoint Online folder in the cloud and each PDF file is relatively small, not exceeding 800 KB, so file size is not the issue, but i am using Copilot Studio in the Portuguese language.

Despite everything being set up, the agent is not accessing the PDF content correctly. It either doesn't return any result or fails to match the file name correctly using the user’s input date.

Is there a limitation or specific configuration needed to allow Copilot Studio to access PDF files in SharePoint Online. What is the recommended method to parse or match file names by date (formatted like DD.MM.YYYY) and use that for retrieving the correct file? Are there best practices for setting up plugin actions to search and return PDF content dynamically from SharePoint?

Any help or documentation would be greatly appreciated!

Thank you!

Categories:

General topics

I have the same question (0)

All responses (2)

Answers (0)

Suggested answer

Veddd007 41 on at

Like (0)

Report
Copy link

Link copied!

hi, this seems to be a product issue that my team also faced recently. Please raise a ticket to microsoft regarding this. It maybe due to a newer version of PDF wrapper that copilot studio is using. However in our environment it has started working well recently which wasnt the case a few days back. If needed to be fixed immediately, raise a ticket else it should get fixed in time

Was this reply helpful? Yes No
Suggested answer

Romain The Low-Code... 2,562 Super User 2026 Season 1 on at

Like (0)

Report
Copy link

Link copied!

By default no agent know the date or time. Public agent like GPT have special exception and service for it. (but private GPT out of the box, don't know too, since copilot studio is a 4o ATM ;) )
So first if you wana work with date you have to add it to the contexte like give the date of day.

Then your pdf need be real PDF, often in company they are fake one ; either scanned image or bad one generated by low cost pdf generator and cannot be read : easy try : open the pdf, try to outline the text and copy past elsewhere : often it work only for few line or 0

Then and least : copilot studio "at the moment" know nothing about metadatas ! know only file content : if you wana make request based on file name you have to try other feature like the add document libary to dataverse (it s recently added on the knowledge adding home page, little blue link in the middle of the page ^^)

Was this reply helpful? Yes No