Copilot Agent cannot read PDFs in a grouped by Sections view

(0) Share

Report

Posted on by CS-19030835-0

I am having trouble with my agent not being able to read the PDFs that are in my SharePoint library because there is a default view (needed for the company users) grouping the PDFs by Sections. If i ungroup them, the agent works flawlessly and can pull the PDFs.

How do I make it so the Agent can "see" the PDFs even if they are grouped by Sections?

Categories:

Bot administration

I have the same question (0)

All responses (5)

Answers (0)

Sort by

Suggested answer

Nivedipa-MSFT Microsoft Employee on at

Like
a
(0)

Report
Copy link

Link copied!
Hello CS-19030835-0,
SharePoint's grouped default view can affect how Copilot Studio's knowledge source indexes documents.

Recommended solution: Add a second flat (ungrouped) view to the library, while keeping the grouped view as the default for users. In Copilot Studio, set the knowledge source to either:

The flat view URL: https://tenant.sharepoint.com/sites/Site/Library/Forms/FlatView.aspx

Or the library root URL (not the view URL): https://tenant.sharepoint.com/sites/Site/Library

This approach lets users keep their grouped view, while the agent can access all PDFs.

Make sure PDFs are checked in, not in draft, and re-sync the knowledge source after any updates.

Was this reply helpful? Yes No
CS-19030835-0 8 on at

Like
a
(0)

Report
Copy link

Link copied!

I have now linked to the library root url with the flat view, and it works sometimes but not then it stops working again. For it to work again I need to remove the URL from knowledges and then add it again :/.

Was this reply helpful? Yes No
Nivedipa-MSFT Microsoft Employee on at

Like
a
(0)

Report
Copy link

Link copied!
Hello CS-19030835-0,
This is a known indexing and caching issue with Copilot Studio's SharePoint knowledge connector. Sometimes, the index becomes outdated or loses synchronization, and re-adding the URL prompts a fresh crawl.
Workarounds:
1. Use the Dataverse Indexing Method Instead of the URL
Rather than linking the SharePoint URL directly:

In Copilot Studio, go to Knowledge → Add knowledge → select SharePoint using the File upload method (Dataverse indexing)

Pick the desired SharePoint site or library

This approach uses Dataverse's search index, which tends to be more stable and updates automatically

This method is generally more reliable than using a library URL.
2. Add Individual File URLs
If your library is small:

Add each PDF's direct URL as a knowledge source instead of the entire library URL

This takes more effort but helps avoid inconsistencies at the library level

3. Use the SharePoint Site URL Instead of the Library URL
Point to the site URL:

https://yourcompany.sharepoint.com/sites/YourSite

Instead of the library URL:

https://yourcompany.sharepoint.com/sites/YourSite/YourLibrary

Using the site-level URL can help trigger a broader and more reliable crawl.
4. Power Automate Scheduled Re-sync
If the above options don't work, you can automate the process of removing and re-adding the knowledge source:

Leverage the Copilot Studio management API or Dataverse to refresh the knowledge source connection on a set schedule

Was this reply helpful? Yes No
CS-19030835-0 8 on at

Like
a
(0)

Report
Copy link

Link copied!

I tried to add the SharePoint library through the File Upload, but it does not change status from 'In Progress' to 'Ready'. I have waited 30 minutes while it was loading it. Is that normal? There are 99 pdf files in the library.

Was this reply helpful? Yes No
Nivedipa-MSFT Microsoft Employee on at

Like
a
(0)

Report
Copy link

Link copied!

Hello CS-19030835-0,
Yes, 30 minutes is normal for 99 PDFs — it's not stuck.

Was this reply helpful? Yes No