Automation glossary

Retrieval-Augmented Generation (RAG)

Retrieval-augmented generation (RAG) is a technique where relevant documents are retrieved from a knowledge source and given to a language model so its answer is grounded in that information.

In plain language

Before answering, the AI looks up your own help articles or policies and answers from them instead of from memory alone.

Example

A support assistant retrieves the three most relevant help articles and answers using only those, with links.

Embedding · Large Language Model (LLM) · Context Window · Hallucination

Common misconception

RAG does not eliminate hallucinations; poor retrieval or outdated documents still produce wrong answers.

Implementation guidance

Keep source content current, split documents into meaningful chunks, cite sources in answers, and test retrieval quality on real questions.

Primary references

← Browse every glossary term