Microsoft Foundry IQ: Knowledge Grounding for AI Agents via MCP
Microsoft announced Foundry IQ in public preview on March 18, a new service within Azure AI Foundry that gives AI agents structured access to enterprise knowledge. Rather than relying solely on what's baked into the model, agents can now query centralized knowledge bases through a single grounding API.
Foundry IQ reimagines retrieval-augmented generation (RAG) as a dynamic reasoning process rather than a one-time lookup. Powered by Azure AI Search, it centralizes RAG workflows into a single API endpoint with auto-chunking of documents, vector embedding generation, permission-aware retrieval, semantic reranking, and citation-backed responses.
The key integration for the agentic ecosystem is MCP support: Foundry IQ now integrates with the Foundry Agent Service via Model Context Protocol, enabling developers to build AI agents that are grounded in enterprise knowledge while respecting user permissions and data classifications. This means agents can access real documents and data rather than operating on stale training data.
Foundry IQ is part of Microsoft's broader IQ series alongside Work IQ (for Microsoft 365) and Fabric IQ (for data analytics), creating a unified intelligence layer across the enterprise. For agent builders, this provides a standardized way to connect agents to organizational knowledge without building custom RAG pipelines.
Documentation: https://learn.microsoft.com/en-us/azure/foundry/agents/concepts/what-is-foundry-iq
← Back to all articles
Foundry IQ reimagines retrieval-augmented generation (RAG) as a dynamic reasoning process rather than a one-time lookup. Powered by Azure AI Search, it centralizes RAG workflows into a single API endpoint with auto-chunking of documents, vector embedding generation, permission-aware retrieval, semantic reranking, and citation-backed responses.
The key integration for the agentic ecosystem is MCP support: Foundry IQ now integrates with the Foundry Agent Service via Model Context Protocol, enabling developers to build AI agents that are grounded in enterprise knowledge while respecting user permissions and data classifications. This means agents can access real documents and data rather than operating on stale training data.
Foundry IQ is part of Microsoft's broader IQ series alongside Work IQ (for Microsoft 365) and Fabric IQ (for data analytics), creating a unified intelligence layer across the enterprise. For agent builders, this provides a standardized way to connect agents to organizational knowledge without building custom RAG pipelines.
Documentation: https://learn.microsoft.com/en-us/azure/foundry/agents/concepts/what-is-foundry-iq