# RAG (Retrieval-Augmented Generation)

> Retrieval-augmented generation (RAG) is a technique where an AI system first searches a set of trusted documents or records for passages relevant to a question, then passes those passages to a large language model to write the answer. Because the model answers from retrieved sources rather than memory alone, responses can reflect current, company-specific information and cite where each fact came from.

## Why it matters for your business

RAG is how most businesses put AI on top of their own knowledge, such as SOPs, contracts, product manuals, policies and past tickets, without training a model. Content is updated by editing the source documents, and access can follow existing permissions. Answer quality depends mostly on retrieval: how documents are split, indexed and filtered, and whether the right passage is found. When a RAG assistant gives a wrong answer, the cause is often a retrieval problem or a stale document rather than the model itself.

## Example

A property management company loads its lease templates, maintenance policies and vendor list into a RAG assistant. When a leasing agent asks whether a tenant can install a washer, the assistant retrieves the relevant lease clause and policy page, answers in two sentences and links to both documents so the agent can confirm before replying to the tenant.

Related terms: https://www.lumeor.studio/glossary/embeddings/, https://www.lumeor.studio/glossary/vector-database/, https://www.lumeor.studio/glossary/chunking/, https://www.lumeor.studio/glossary/grounding/, https://www.lumeor.studio/glossary/knowledge-base/

---

Source: https://www.lumeor.studio/glossary/rag-retrieval-augmented-generation/