Glimpse: advanced RAG patterns

Know the names and ideas you will hear: Contextual Retrieval (add a one-line summary of the document to each chunk before embedding — often a big gain for little work), Corrective RAG (check if retrieved chunks are relevant; if not, rewrite the query or fall back), Self-RAG (the model decides when to retrieve and checks its own answer), Adaptive RAG (choose no retrieval / simple / multi-step by question difficulty). Try Contextual Retrieval on your documents. The agent-driven versions (Agentic RAG) come in the Agentic AI module.