liveBook · Manning

Agentic RAG

As explained in Designing AI Agents

Retrieval-augmented generation in which an agent can make decisions about whether and how to retrieve information.

Agentic retrieval-augmented generation lets a model decide whether to retrieve, revise a query, and route across sources. That flexibility adds model calls and expands the verification surface. section 4.3.1

A governed pipeline separates corpus preparation from evidence assembly. Preparation records provenance, permissions, and versions while building indexes or navigation; assembly selects retrieval methods, reranks results, follows required relationships, and packages cited evidence. A demonstrated implementation uses query rewriting, hybrid search, fusion, and context assembly. Prefer deterministic lookup when the task supplies an exact key. section 4.3.1

Figure 4.3 RAG Pipeline from basic retrieval to governed compile and assemble paths. Provenance, version, permission, and source identifiers survive into the evidence package.

Trace how provenance, versions, permissions, and source identifiers carry through the governed retrieval pipeline into its evidence package. section 4.3.1

How Designing AI Agents covers it

How deep it goes
Implements query rewriting, hybrid search, fusion, and context assembly within a broader governed pipeline.
Examples
Uses query rewriting, hybrid search, result fusion, and context assembly as an implementation example.
What you can do after
Apply it by letting a model choose whether to retrieve, revise queries, and route across sources.
How it connects
Connects retrieval-augmented generation with query rewriting, hybrid search, reranking, and context assembly.

Also explained in

Compare how 9 books explain agentic RAG →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Designing AI Agents. The section links point to the parts of the book it draws on.