Glossary
Retrieval-augmented generation
One term from the glossary, defined on its own page: what it means to anyone who meets it, and how we use it.
In short
Retrieval-augmented generation, or RAG, is a technique for grounding a language model's answer in a source corpus: passages relevant to the question are found at the moment it is asked and placed in the model's context, so the reply is built from retrieved text as well as from what the model already holds.
In practice
What it means for a business.
In our builds the corpus is your own material: your services, your prices, and the way you already answer clients are the usual starting point. It is retrieved at answer time and no model is retrained on it, and the access is agreed in writing before anything is connected.
See what each build works from, or see every term defined once.
Book a strategy call
Thirty minutes. A clear plan.
Thirty minutes on the question that brought you here, answered against your own figures rather than ours.
Book a thirty-minute call
Pick a time straight from the diary. If you would rather write first, email us and a person replies, usually the same working day.