Skip to content
anythingai.space
Retrieval & RAGbeginner12 min · 9 steps

RAG Fundamentals

Give a model the right facts at question time instead of hoping it memorised them — and know when not to.

grounded0%
context40tok
p95900ms
question“Refunds are available within 60 days.”(actual policy: 14 days)UserLLMguessing
requesterrordataresponsetoken
metrics are illustrative
1 / 9
ProblemStep 1 of 9

The model has never seen your data

A user asks: “What's our refund window for annual plans?” The LLM was trained on the public internet months ago. It has never read your policy docs.

So it does what language models do — produces a fluent, confident, plausible answer. It's wrong, and nothing in the reply tells the user that.

Watch outA wrong answer that sounds right is worse than “I don't know”.

Lesson 1 of 37 in Retrieval & RAG