SharePoint · Connector
How do we connect SharePoint to AI agents on our own infrastructure?
Katonic connects to SharePoint through Microsoft Graph, reads the documents in the sites you name and indexes them in the Knowledge Engine on infrastructure you control. Agents then answer from those documents with a model you serve, so the documents and the questions are not sent to a public AI service.
What it reads.
From SharePoint.
- 01Sites and their document libraries
- 02OneDrive, personal and business
- 03Shared drives
Through: Microsoft Graph API v1.0
Where the content goes
Into a knowledge store you run.Content is pulled into the Knowledge Engine on infrastructure you own or control, and indexed there. Agents answer from it with a model you serve. The documents, the index and the questions are not sent to a public AI service.
How it signs in
OAuth 2.0 with an Azure AD application. You register the application in your Microsoft tenant and give Katonic its details.
The set-up dialog asks for
- SharePoint site URL or URLs
- Tenant ID
- Client ID
- Client secret
Permissions
Access is set on the source in Katonic. In addition, Katonic reads the permissions of each item from SharePoint and re-syncs them, every 24 hours by default. Whether an item's own permissions restrict who can retrieve it, or only add to the access set on the source, is a setting an administrator turns on for the cluster.
Katonic reads permissions through the Graph permissions API and resolves Azure AD groups to their members. A document shared outside the organisation is blocked and flagged for an administrator.
Sync
Incremental, using Microsoft Graph delta queries, and full.
You set how often the source is read again, in minutes, hours, days or weeks.
Limits
- A site scope is required: you name the site or sites to read. The connector does not read every site the account can reach.
Use case pages that name SharePoint. More are in the gallery.
Does Katonic read SharePoint permissions?
Yes. It reads the permissions of each document and re-syncs them, every 24 hours by default. Whether those permissions restrict retrieval, or only add to the access set on the source in Katonic, is a setting an administrator turns on for the cluster.
Does it read every site in our tenant?
No. You name the site or sites. A site scope is required.
Where are the documents indexed?
In the Knowledge Engine, on infrastructure you own or control. They are not sent to a public AI service.