In the year of the Agents, the industry is shifting its attention to autonomous workflows and orchestration—but Retrieval-Augmented Generation (RAG) remains just as essential as ever. What’s changing is how RAG is built and how it integrates with new technologies and APIs, especially Microsoft Foundry.

This session explores how to design modern RAG architectures using Foundry, clarifying the roles of both AI Search indexes and the new Knowledge Base concept introduced together with AI Search. You’ll learn when a simple vector-based index is enough, when a Knowledge Base becomes necessary, and how Knowledge Bases work internally to enhance grounding and retrieval quality.

We’ll also demonstrate how agents can use either approach—whether linking directly to an AI Search index or leveraging a Knowledge Base as a richer grounding source. By the end, you’ll understand how to choose the right RAG pattern for your scenario and how Foundry elevates retrieval in agent-driven systems.