AI glossary
Retrieval-augmented generation (RAG)
What is a Retrieval-augmented generation?
Retrieval-augmented generation (RAG) is a technique where an AI system first searches a trusted knowledge source, such as company documents or a database, and then gives the relevant passages to a language model so its answer is grounded in that information and can cite sources.
Last reviewed · Dexorn Technologies
RAG reduces made-up answers and keeps responses current without retraining a model. The knowledge base is split into chunks, indexed with embeddings and searched for each question.
It is the standard approach behind internal knowledge copilots, customer-support assistants and document Q&A tools.
Examples
- Copilot that answers HR policy questions
- Support assistant grounded in product manuals
- Search across contracts with cited clauses
Put it to work
Custom AI Software & Products
Copilots, RAG search, AI-powered web and mobile apps, built end to end.