Multi-agent RAG and knowledge integration
AI-500 · Lesson 7 of 16 · 54:17
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the AI-500 exam (Microsoft Certified: Multi-Agent AI Solutions Expert). In this lesson: chunking (a 512-token start with overlap, structural, semantic, and parent-child chunking), embedding quality and re-embedding into a new index, retrieval precision with hybrid search, RRF, the semantic ranker, rerankers, MMR, and offline metrics (precision at 5, MRR, nDCG), multi-agent RAG with routing, fallback, merged results, and security trimming, and knowledge integration through the Azure AI Search tool, file search, MCP sources, and Foundry IQ knowledge bases. Includes three original practice questions.
Preview the topics
- Travelers ask Contoso’s visa agent about rule V-2041, and vector search keeps returning similar but wrong regulations. What should the team add?
- Contoso’s router sends each question to one of three indexes, but for vague questions it often picks the wrong one. What should the design add?
- A traveler uploads a 40-page corporate travel policy during a chat and asks what it allows. Which knowledge option fits?