An AI system that is augmented by the retrieval of data from an external source. Retrieval-augmented generation (RAG) enhances a large language model (LLM) by accessing an external data set.
A RAG example could be a chatbot platform for customer service that uses the LLM for natural language interaction and the external database for product specifics. Although AI agents can use RAG to access external databases, agents access data for many purposes, and they are more comprehensive and have greater flexibility than a RAG model. See
AI agent,
chatbot and
large language model.