Knowledge Connectors

Google Drive and Microsoft SharePoint work differently from other Connectors. Instead of fetching data in real time, they index your files (process and catalog them for fast search) and store them on our platform so Work can search them instantly, similar to how Libraries work.

This means your team can query documents across your entire drive using natural language, with results based on the actual files and respecting your existing access permissions.

1. An administrator configures the Connector and initiates indexing of your organization's files.
2. Files are indexed and stored in our European data centers.
3. Each user connects with their own account and starts querying indexed content from any task.
4. The index syncs regularly with the source. Edits, new files, and deletions at the source are reflected during each scheduled sync.

This is a one-time process. Once indexing is complete, your team can connect and start using the Connector.

Selective sync lets you choose exactly which folders, sites, or libraries to index. To configure it:

1. Open the Admin Controls tab on the Connectors page.
2. Click on the Connector card.
3. Browse and check/uncheck individual folders or sites to include in indexing.

To add new folders to an existing index, re-run the indexing process. New content is added to the existing index.

The platform monitors changes at the source (creation, modification, deletion) and updates the index during each scheduled sync. Deleted files are removed from the index and can no longer be queried through Work.

Once your administrator has completed indexing, you can connect with your own account:

1. Open the Connectors page from the sidebar.
2. Find the Google Drive or SharePoint card and click Connect.
3. Sign in with your own Google or Microsoft account.
4. Confirm the connection.

A green Connected indicator confirms you're set. You can now query indexed content from any task by enabling the Connector as a tool.

We replicate user, group, and permission structures from the source system. If you don't have access to a file in Google Drive or SharePoint, you can't access it through Work either.

• Permissions are based on explicit sharing (user, group, or domain level).
• Files shared as "anyone with the link" are not automatically visible to your entire organization through Work. They must be explicitly shared with specific users or groups.

• Permissions rely on Microsoft Entra ID groups. Older SharePoint-only groups may not be recognized.
• Files shared exclusively with external guests are generally not indexed.

What happens when you disconnect depends on who does it:

• A user disconnects (from My Connectors): your ability to query the data is removed, but the index stays intact for other users in your organization.
• An admin disconnects (from Admin Controls): the Connector is disabled organization-wide and the indexed data is scheduled for permanent deletion.

Re-enabling after an admin disconnection requires a full re-setup and re-indexing.

Both give Work access to your documents, but they work differently. Pick the one that fits your situation:

• Files are indexed and stored in our European data centers.
• The index is retained only while the knowledge Connector is active. If an admin disables the Connector, data is immediately scheduled for deletion.
• Data accessed through Connectors is never used to train or improve our models, regardless of your plan.
• Conversations that reference Connector data follow your plan's data policies.

For more details, see our privacy policy and trust center.