A method where an agent retrieves documents and uses them to ground its answer.
When retrieval fails, the agent can answer from nothing, a common cause of hallucination.
Related: AI Agent Hallucination, Large Language Model
See also: What Is AI Agent Hallucination