Stop rebuilding RAG pipelines for every agent.

Foundry IQ introduces a new abstraction for retrieval: reusable Knowledge Bases that handle query planning, hybrid search, reranking, and grounded responses out of the box. Instead of stitching together vector stores, prompts, and orchestration, developers can focus on how agents use knowledge, not how it’s retrieved.

In this session, we’ll walk through how Foundry IQ integrates with the Microsoft Agent Framework to enable predictable, multi-source retrieval in real-world scenarios. You’ll see how to design retrieval that supports agent reasoning, structured outputs, and enterprise requirements like grounding and traceability.

We’ll demo a production-style agent and highlight practical patterns for steering retrieval behavior, combining sources, and avoiding common pitfalls.