RAG (Retrieval-Augmented Generation)
A technique that lets an AI look things up before it answers. It retrieves the relevant documents first, then generates from them, instead of relying on memory.
What it is
RAG, retrieval-augmented generation, is a way of giving an AI the right information at answer time. Instead of relying only on what the model learned in training, the system first retrieves the most relevant documents, your notes, your docs, your design system, and feeds them into the context. Then the model generates an answer grounded in those documents.
Simple analogy
It is open-book instead of closed-book. Rather than answering from memory, the model is handed the exact pages it needs, then writes the answer from them.
Why this matters for designers
RAG is how you make an AI actually know your system instead of guessing about it. Point a retrieval setup at your component docs, token files, and decision logs, and the model answers from your real material. This is also the strongest defense against hallucination: grounded answers are far harder to invent.
How it works in practice
- Your documents are indexed so they can be searched by meaning.
- When you ask something, the system retrieves the most relevant pieces.
- Those pieces go into the context, and the model answers from them.