Definition
Retrieval-augmented generation is an AI setup where the model fetches relevant documents from an external source, like the live web, before it writes an answer. It improves accuracy, recency, and how citable the answer is. Most modern AI search runs on it.
On this page
At a glance
- Categories Technical
- Abbreviation RAG
- Related fields AI Search, Technical
- Difficulty Advanced
How Retrieval-Augmented Generation Works
A plain model answers only from what it learned in training, which has a cutoff and no sources. RAG adds a retrieval step: the system searches for relevant material, feeds it to the model, and the model writes its answer grounded in those documents. That is why answer engines can cite pages and handle recent events at all.
Why Retrieval-Augmented Generation Matters
RAG is the reason your website can show up in an AI answer in the first place. Because these systems pull live content before answering, being retrievable and trustworthy drops your pages right into the path of the answer.
Frequently Asked Questions
-
Why do engines use RAG?
It grounds answers in real, current sources, which lifts accuracy and lets the engine cite where the information came from.
-
Does it stop hallucinations?
It reduces them by anchoring answers to retrieved sources. It does not remove them entirely.