RAG

An AI architecture pattern that pairs a large language model with an external knowledge store (typically a vector index of text chunks) so that, for each user query, relevant documents are retrieved first and injected into the prompt before the model generates a response. RAG reduces hallucination, keeps answers grounded in up-to-date or domain-specific information (e.g. software documentation, medical guidelines, organisational policy), and allows the knowledge base to be updated without retraining the model. In accessibility tools, RAG is used to ground AI-generated assistance in authoritative, accessible documentation rather than the model's possibly stale internal knowledge.