RAG

Retrieval-Augmented Generation (RAG) is an AI architecture that combines the generative capabilities of large language models with information retrieval systems. Instead of relying solely on the model's training data, RAG retrieves relevant documents, data, or context from external sources and includes them in the prompt, enabling the model to produce more accurate, up-to-date, and grounded responses. For database applications, RAG can retrieve relevant schema documentation, previous query patterns, business glossaries, and data dictionaries to improve text-to-SQL accuracy. RAG reduces hallucinations and keeps responses grounded in actual data.