Retrieval augmented generation

Retrieval augmented generation, or RAG, is a method where an AI model looks up relevant information at the moment it answers, rather than relying only on what it learned during training. The model retrieves the right facts and uses them to write the answer.

In agentic analytics, RAG is one way an agent finds the right context for a question. Faced with a question about churn, it can pull in the churn definition, the relevant tables, and the rules that apply, instead of guessing from a general sense of the schema.

How it relates to context engineering

RAG is a technique. Context engineering is the larger discipline of building and governing the knowledge the agent retrieves. Good retrieval does not help if the underlying definitions are missing or wrong, which is why the context itself has to be built with care.