Agentic RAG
Agentic RAG: RAG in which an agent plans searches, tools and checks instead of following a single fixed pipeline. The system iteratively chooses what to retrieve and when to stop, which is useful for multi-hop questions. Budgets, permissions and traces are required to control cost and reliability.
In italiano: Agentic RAGLegge I · Embedding, feature e spazi vettorialiThe system iteratively chooses what to retrieve and when to stop, which is useful for multi-hop questions. Budgets, permissions and traces are required to control cost and reliability.
Examples
- Searching multiple sources
- Multi-hop queries
- Verification and renewed retrieval