AI Discovery/ Discovery Library/ Glossary/ RAG (Retrieval Augmented Generation)

RAG (Retrieval Augmented Generation)

The technique of grounding LLM responses in retrieved content from a live index. RAG is the mechanism by which web-enabled AI systems return up-to-date answers.

Definition

The technique of grounding LLM responses in retrieved content from a live index. RAG is the mechanism by which web-enabled AI systems return up-to-date answers.

Why this matters

Understanding RAG is essential to understanding why content recency and answer shape matter so much for AI Discovery.

Where it sits in The Standard™

This term is part of the canonical AI Discovery vocabulary, published in the Body of Knowledge alongside the Eight Discovery Signals and The Methodology Handbook. The full glossary contains 56 terms — the operative vocabulary for AI Discovery practitioners.

Practitioner note

Practitioners certified through Certified™ are tested on the AI Discovery vocabulary. This glossary is the institutional reference.

Find out where you stand in 7 days.

The AI Discovery Audit™ delivers a 0–1000 score, an eight-signal breakdown, and a 90-day roadmap. From $800.

Get Your AI Visibility Audit