Retrieval-augmented generation (RAG)
RAG is a technique where a model retrieves relevant documents from a knowledge base and uses them to ground its answer, rather than relying only on what it was trained on.
In detail
RAG answers the question what does the knowledge base say about X. A memory layer answers a related but different question: what has this team or user decided and learned over time. They share machinery (embeddings, retrieval) but solve different problems; a team memory layer is closer to durable per-team state than to document question-answering.