RAG lets an AI answer from your own data. Before the AI answers, the system finds the relevant documents and sends them along. Domino IQ can use RAG from Notes 2026.
RAG stands for Retrieval Augmented Generation. In plain words: find first, answer after. RAG lets an AI answer from your own data and not only from what it has learned on the internet.
Why do we need RAG?
A language model knows a lot about the world. But it does not know your customers, your contracts or your own rules. If you ask about them, it guesses. This is called hallucinating. RAG solves it by giving the model the right facts just before it answers.
How does RAG work?
- Your documents are turned into numbers that describe their content. These are called vectors.
- When a user asks a question, the system finds the documents that are most similar to the question.
- Those documents are sent to the language model together with the question.
- The model answers from them and can show where the answer comes from.
RAG in Domino IQ
Domino IQ arrived in Domino 14.5. From Notes and Domino 2026, Domino IQ can use RAG. You point to a Notes database, and Domino creates vectors from the documents. Then a Notes application can ask questions and get answers from the content of the database.
The smart part is that RAG in Domino follows the database ACL. A user only gets answers from the documents he is allowed to read. Data stays in house, because the language model can run on your own Domino server.
What can you use it for?
- A search in manuals and procedures where you ask in everyday words.
- A quick summary of a customer's history before a meeting.
- Help to write replies to requests based on earlier cases.
We are happy to help you use Domino IQ and RAG in your own Notes applications. Read more about Notes development or in the article on Notes 2026 and Domino 2027.