SharePoint Connector

Governed, real-time access to enterprise knowledge stored in SharePoint, protected by Content Intelligence security models, is now available to your AI Agent teams.  This connector offers:

  • Natural-language search and semantic retrieval
  • Reads and syncs content to hybrid search index
  • Support for multi-site architectures

Business Benefit

  • Faster knowledge discovery and decision-making
  • Reduced manual search and content triage
  • Improved content reuse and consistency

Steps to enable and configure

Register a new application

You will need to register Fusion as a new application in your Microsoft Entra ID platform.  This is because the connector is going to connect to SharePoints APIs via oAuth 2.0 and needs sufficient access to read the files you are hoping to connect to.  When you register the application you will need to capture these pieces of information:

  • clientKey
  • clientCert
  • SharePoint Application Client ID - this is a GUID

Additional information to collect from your SharePoint site(s)

  • SharePoint Tenant ID - this is a GUID that can be found on your Entra overview page
  • Site name - should be visible in your SharePoint url immediately after /sites/
  • Folder paths - your SharePoint site may contain A LOT of files, make sure that the ones you are syncing are the files you really want by selecting the specific folder(s) that contain your files.  Start small, you can always add more folders later.

Create one or more SharePoint connectors

  • Access AI Agent Studio
  • Navigate to Credentials > Connectors
  • Add a new connector
  • Select "SharePointO365"
  • Fill in all of the required fields with the information collected above

Schedule process to sync files

The "Content Intelligence SharePoint Connector Job" scheduled process should be scheduled to run regularly in order to maintain the synchronization of files.  If you are synchronizing a large quantity of files, note that the first execution may take quite a bit of time.  Subsequent runs will be incremental and therefore should complete more quickly.  It is recommended, as you are first setting up the connector, that you run this job ad hoc until you have everything configured correctly and your first sync completed.  Afterwards, run it no more frequently than every 30 minutes so it has time to complete between executions.  Note that this same scheduled process is required for the web connector's crawling, you only need to schedule it once to cover both.

(Optional) Update user groups of sync'ed files

If you want the sync'ed files to have different user groups from the one selected in the connector setup you will need to access authoring to review and update the user groups.  You can also mass update user groups.  User group updates will be retained even if the file is updated on the SharePoint side.  You can access authoring by going to https://<your domain>/fscmUI/redwood/knowledgeauthoring/main?contentType=<connector code>.  Make sure to replace the bolded parts of the example url with your actual data.

Use connector in one or more agent teams

  • Create a tool for every SharePoint connector that you have configured. 
  • Add your connector tool(s) to whatever agent(s) they will be useful in
  • Via your agent's prompt, tell it to use the tool in the circumstances where you want that data consulted

Tips and considerations

Ingesting and indexing large amounts of content will contribute to the total size of Content Intelligence content, which has a 4TB limit.

Uploading content to SharePoint will count against the SharePoint storage limit (usually 1TB per user).

SharePoint sync activity can be reviewed from the logs available in Scheduled Processes.  If articles are not being created, this is the first place to look. 

If articles are created, but they do not have attached files, review the articles as they may contain an explanation as to why the file is missing.  The most likely causes are the file being greater than 100mb OR the file having an extension that is not marked as allowed in the "Manage Knowledge Files and Images" admin.