Copilot Agent having trouble accessing sharepoint files

Hi everyone,

I’ve built an agent using Copilot Agent Builder for my job (unfortunately I don’t have access to the full Copilot Studio). The agent is supposed to retrieve insights from market research studies stored on a SharePoint site.

The agent does have access to the SharePoint location, but every time I query it, I get the same first response before actually going forward with answering my questions:

Unfortunately, the contents of the files could not be searched directly, but the snippets already provide a very clear picture.

This isn’t helpful, because I need the agent to actually extract and reference specific content from the documents. Ideally, it should mention the exact slide when referring to information, but it can’t do that.

Has anyone else run into this issue?

• Is there a way to improve how SharePoint files are indexed or made searchable?
• Are there specific file formats, metadata, or permissions that need to be configured?
• Any tips on how to get the agent to go beyond just snippets?

So far, I’ve tried uploading the SharePoint files directly to the knowledge base (which partially solved the issue, but then it started citing the wrong reports). Also, I’d prefer not to manually upload new reports every time one is added to SharePoint, since my whole team uses the agent.

tldr: Agent built with Copilot Agent Builder can access SharePoint, but only returns “snippets” instead of extracting real content (PDF, PPT, Word, some Excel). OCR doesn’t work, files are set to Internal/Public. Looking for tips on indexing, formats, or permissions to make documents fully searchable.