Retrieval-Augmented Generation
A framework that retrieves relevant passages first, then feeds query plus evidence to the model so it answers from real text.
Retrieval-Augmented Generation, or RAG, is the backbone of how modern AI assistants stay accurate and up to date. First proposed in 2020, this framework works by fetching relevant real-world information before the model even starts writing.
When you ask a question, the system routes your query through a vector index to find the best matching source passages. It then feeds both your query and this live evidence directly into the model's context window. This retrieve-then-generate order builds factuality right into the process, allowing the model to reason over actual text instead of inventing facts.
This approach has major advantages over generate-then-ground methods, which write a draft first and try to bolt citations on afterward. With RAG, if the system can't find any relevant information, it can fail cleanly and let you know right away. It is also faster because it only needs a single pass to generate a complete, accurate response. Today, this technology powers major real-world applications, including Amazon's Rufus shopping assistant, proving that grounding AI in real-time search is the key to reliable answers.
Retrieval-Augmented Generation (RAG) is a framework that fetches relevant evidence before the model writes, then feeds the query plus that evidence into the model in a single context window. First proposed by Lewis et al. in 2020, it routes a query through a vector index, pulls the top passages, and lets the model reason over real text instead of inventing facts.
The retrieve-then-generate order is what makes it robust: factuality is built in because the model works from supplied text; failure is cleaner because if retrieval finds nothing the system can say so early; and it is faster, needing one pass rather than draft-then-fix. We have argued this beats Google's generate-then-ground approach, which bolts citations on after drafting.
RAG underpins how modern assistants stay current, and it leans on embeddings and vector search to supply the "retrieve" step. Amazon's Rufus shopping assistant is a production example built on RAG.