NLP & Language · intermediate · concept 143 of 311 · Updated Nov 4, 2026

🗂️Agentic RAG

Agentic RAG puts a language model in charge of the retrieval loop: it decides whether to search, rewrites the query, routes it to one of several sources, grades the passages, and searches again when they fall short. It matters for multi-step questions and mixed sources, where single-shot RAG often answers from the wrong passage. Example: a question across three annual reports becomes one search per report, then one merged answer.