Foundry IQ SharePoint (indexed) knowledge source is now generally available
The SharePoint indexer and SharePoint (indexed) knowledge source in Foundry IQ are generally available. They bring libraries, lists and pages into a managed pipeline that carries SharePoint user, Entra group and site group permissions down to each chunk. Also GA: semantic chunking and image descriptions, document-level permissions for ADLS Gen2, Blob and the Push API, private ingestion, cross-tenant customer-managed keys, and Azure SQL and File knowledge sources. OneLake permission preservation and Microsoft Retriever 1 are in public preview.
- status
- Generally available
Permission-trimmed retrieval is what makes it safe to ground agents in SharePoint. The catch is that permissions are captured at indexing time, not at query time. Item-level changes appear after the next sync, and a change inherited from a parent scope can need an explicit resync, so plan for a window where someone who lost access can still get answers from that content. The index stores SharePoint access lists, not SharePoint's full policy model; the remote SharePoint source (preview) checks permissions and sensitivity labels live at retrieval.
Official sources
- techcommunity.microsoft.comFoundry IQ SharePoint (indexed) knowledge source is now generally available