Retrieval-augmented generation (RAG) is a technique where an AI system searches an external source, such as the live web, retrieves relevant documents, and uses them to ground its generated answer. It lets a model draw on current information and cite sources beyond what it learned during training.
Most AI search experiences use RAG to fetch and cite live pages. For brands, RAG is the reason crawlable, well-structured content can be surfaced and cited even for topics a model did not know from its training data.