RAG (Retrieval-Augmented Generation)
Technique that connects an LLM to your own documents so it answers with your information, not its training.
RAG (Retrieval-Augmented Generation) combines an LLM with a database of proprietary documents to answer questions with up-to-date, client-specific information. Instead of asking the model to recall a manual, you first retrieve relevant passages and pass them as context to the LLM before generating the answer. It's the standard architecture for enterprise chatbots over internal docs, support assistants, and intelligent search tools. RAG is preferable to training your own model when content changes or is confidential.