Glossary

RAG

Retrieval-Augmented Generation — a pattern that grounds LLM responses in retrieved source documents to reduce hallucination.

Why it matters

Business and delivery impact.

RAG lets AI assistants answer from your actual docs instead of guessing, making design-system chatbots trustworthy.

How to measure

Proof you can track.

Retrieval precision, answer faithfulness and hallucination rate.

Example

What it looks like in practice.

A Slack bot that retrieves the latest token docs before answering 'which color token should I use for errors?'

Keep the context connected.