m365_retrieve
Retrieve raw text chunks from Microsoft 365 for AI analysis and cross-document reasoning. Get relevance-scored excerpts from SharePoint and OneDrive to maintain control over content synthesis.
Instructions
Retrieve raw text chunks from M365 for YOUR AI to reason over.
Returns relevance-scored excerpts from SharePoint/OneDrive—you control synthesis.
Use for:
- Custom analysis and cross-document reasoning
- When you need source text, not just answers
- Deep research where you want control over synthesis
Use m365_chat instead for:
- Quick Q&A where M365's answer is sufficient
- Calendar/email questions
- People lookupInput Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | Natural language query for enterprise content. Be specific: include document types, projects, or topics. E.g., 'Q4 revenue projections for ACME deal' not just 'revenue'. | |
| data_source | No | Where to search: 'sharepoint' (team sites, wikis), 'onedrive' (personal files), 'connectors' (external systems via Copilot connectors) | sharepoint |
| filter_expression | No | Optional KQL filter to narrow scope. Examples: 'path:https://contoso.sharepoint.com/sites/HR', 'FileType:pdf', 'LastModifiedTime>2024-01-01' | |
| max_results | No | Number of text chunks to return (1-25). More chunks = more context but longer processing. |