Building Agents
Retrieval-Augmented Agents
How agents search private knowledge, decide what to retrieve, and ground answers in evidence.
Retrieval supplies context
A retrieval system finds relevant passages or records at decision time instead of placing an entire knowledge base in the prompt. The agent can search iteratively as its understanding of the task improves.
Search is a tool
Treat retrieval like any other tool with query, filter, and result contracts. Metadata filters, access controls, and source dates are as important as semantic similarity.
Require evidence
Ask the agent to connect important claims to retrieved sources and to say when evidence is insufficient. Citation correctness needs its own evaluation rather than being assumed from fluent output.