Module 2 · Retrieval that reasons · Lesson 6
Agentic retrieval
How to let an agent run its own searches with search, grep and read tools, follow cross-references across a case's documents, and stay within a token and latency budget.
By the end of this lesson you can
- Explain why single-shot retrieval fails on multi-hop questions, and how interleaving retrieval with reasoning fixes it
- Design a small set of retrieval tools (list, search, grep, read) with scoped access, compact results and stable citations
- Decide when grep and file tools beat an index, using corpus size, vocabulary and the evidence on agentic search
- Evaluate an agentic retriever on evidence recall, answer quality, tool calls, tokens and latency