—Glossary
Retrieval-augmented generation (RAG)
Retrieval-augmented generation (RAG) is a method in which an AI system first finds relevant passages in a set of documents, then uses them as context to write its answer.
Retrieval-augmented generation, usually shortened to RAG, gives a large language model the right reference material at the moment it answers. Instead of relying on what the model absorbed in training, the system looks up your own content and asks the model to answer from it.
How it works
- Prepare the content. Your documents, such as service pages, policies, price lists and FAQs, are split into short passages and indexed so they can be searched by meaning, not just keywords.
- Retrieve. When a question comes in, the system finds the passages most relevant to it.
- Generate. The model receives the question plus those passages, with instructions to answer only from them.
Why businesses use it
RAG keeps answers specific to your business and current. Update a policy page and the next answer reflects it, with no retraining. It also makes answers traceable, because you can see which passage an answer came from.
RAG reduces AI hallucination but does not remove it. Answers are only as good as the source content, and the model can still misread a passage. Good setups tell the assistant to say it does not know, and hand off, when no passage covers the question. Live data such as calendar availability or order status is better fetched from the system of record through an API than stored as documents.
How it relates to Worcia
RAG is how assistants in Concierge, Worcia's AI receptionist, answer from your approved business information rather than from general knowledge.