Microsoft has expanded Foundry IQ’s capabilities by integrating Azure SQL Database as a knowledge source, now in public preview. Enterprise developers building Copilot, RAG, and agentic experiences can expose authoritative SQL tables and views directly as grounding data for AI models.
This solves a real pain point. Before this, getting structured relational data into an AI model’s context window meant building custom ETL pipelines: export to blob storage, index in Azure AI Search, wire up access control separately. Now Azure SQL tables become a first-class knowledge source in Azure AI Search, meaning the indexer reads directly from your database, respects column-level permissions, and surfaces results through Foundry IQ’s query interface. No intermediate files, no sync scripts.
The practical impact is biggest for teams running operational databases behind line-of-business applications. A customer support agent can look up a user’s order history from the live SQL database instead of a stale export. A financial analytics agent can query the general ledger table with the same freshness and access controls the application itself uses. Foundry IQ handles the chunking, embedding, and retrieval pipeline automatically once you configure the data source.
To try it: navigate to Microsoft Foundry, create a new knowledge source, select Azure SQL Database, and provide a connection string with read-only access to the tables you want to expose. The system indexes the data and makes it available to any agent or copilot configured to use that knowledge base. Start with a single authoritative table (product catalog, customer list, reference data) and let the agent answer questions from it before expanding to more sources.
What this means: relational databases have been the missing piece in the enterprise RAG stack. Vector stores handle unstructured data well, but most business logic lives in normalized SQL tables with foreign keys, views, and stored procedures. By making Azure SQL a native knowledge source, Foundry IQ closes the gap between operational databases and AI agents. Expect more Azure data services to follow the same pattern.
Sources: Azure update announcement, Azure AI Search documentation, Microsoft Foundry documentation.