New Copilot Agent: Add SharePoint knowledge source error / not working

Question

My customer wants to use Copilot Agents with available SharePoint content. But the content is not evaluated by the Agent. The customer is missing content from e. g. PDF files. What we tried so far:

• Created an Agent from the SharePoint document library directly. In this scenario our test prompt works and the agent replied with content extracted from the PDF file.
• Created a new fresh Agent in Copilot and add the site to the knowledge source. Published the agent multiple times. All tests failed. The exact same question from above test failed - the content from the PDF is not considered
• Uploaded the PDF directly to the agent. This worked too.

Today the whole system stopped working. We tried to add the SharePoint site to a new Agent, but we receiving every time an error during add content from SharePoint:

This item could not be found. Please use the Copy Link button from the Share menu

We tried all possible URL combination. In our first agent we could add the SharePoint Site URL to the knowledge sources.

I did the same in my test tenant and everything is working as expected - the pdf content is considered and the reply from the agent is complete. Also the source is showing correct.

Answer 1

Hi Marc,
I tried to repro the issue but at my end It is able to share the data from knowledge source.

Common Mistakes When Adding SharePoint to Copilot Agent:

1. Wrong URL Format – Using browser URL instead of "Copy Link" from Share menu.
   ✅ Use the official Copy Link option.
2. Using Site URL Instead of Library/Folder
   – Adding the site homepage instead of a document library or folder.
   ✅ Point directly to the document library or folder.
3. Permission Issues
   – Copilot Agent lacks access to SharePoint content.
   ✅ Ensure proper user and service permissions are set.
4. PDFs Not Indexed or Too New
   – Content not immediately searchable after adding.
   ✅ Wait for indexing or test with direct uploads.
5. Non-readable PDFs
   – Scanned or encrypted PDFs can’t be parsed.
   ✅ Use text-based, machine-readable PDFs.
6. Sensitivity Labels or IRM
   – Protected files may be ignored.
   ✅ Remove labels or test with unprotected versions.

Thanks,

Keshav Keshari

---

If the response is helpful, please click on "upvote" button. You can share your feedback via Microsoft Copilot Developer Feedback link. Click here to escalate.