# RAG

> Retrieval-augmented generation (RAG) combines information retrieved from an external collection with a model’s generation of an answer.

*Source: [https://contentcucumber.com/aeo-glossary/rag/](https://contentcucumber.com/aeo-glossary/rag/)*

*Topic: AI search basics. Also called: Retrieval-augmented generation, Retrieval augmented generation*

---

## Why it matters

The answer depends on both the retrieved material and the generated explanation. Check each one.

## Example (illustrative, not a client result)

An assistant retrieves a delivery-policy document, then uses that document to answer a customer’s question.

## What to check

When evaluating a RAG answer, inspect which material was retrieved as well as what the system wrote.

## What this does not tell you

A RAG answer is only as accurate as the collection behind it, which may be outdated, incomplete, or poorly matched to the question.

## Sources

- [Lewis et al.: Retrieval-Augmented Generation](https://arxiv.org/abs/2005.11401) (Research paper)

*Evidence type: Documented concept. Sources checked on 2026-11-19*

## Related terms

- [Grounding](https://llms.contentcucumber.com/aeo-glossary/grounding.md)
- [Hallucination](https://llms.contentcucumber.com/aeo-glossary/hallucination.md)
