RAG (Retrieval-Augmented Generation)
Give the AI documents to look at before it answers. Fixes the "it doesn't know my stuff" problem.
An LLM only knows what was in its training data. If you want it to answer questions about your company's internal docs, your codebase, or information newer than its cutoff — it can't on its own.
RAG — Retrieval-Augmented Generation — is the pattern that fixes this. Before the model answers, a separate system searches a document store for relevant snippets and stuffs them into the prompt as context. The model answers based on what was retrieved.
Most "chat with your docs" products are RAG. So are many customer support bots and internal search tools. The retrieval piece usually uses a vector database. Retrieval quality often makes or breaks the whole system.