What is RAG?
RAG (retrieval-augmented generation) is a technique in which an AI system first searches your own documents for relevant passages and then has a language model write the answer from them. The term comes from a 2020 research paper by Lewis and colleagues at Facebook AI Research.
Also called: retrieval-augmented generation
Why it matters
If you need an AI assistant that knows your products, policies and contracts, RAG is usually a quicker route to it than retraining a model. Answers can cite the source document, so staff can check them. When a price list or policy changes, you update the index, not the model.
1How it worksDocuments are split into chunks, stored with embeddings, and the closest chunks are added to the prompt for each question.
2RAG vs fine-tuningRAG gives the model facts at question time; fine-tuning changes how the model writes, not what it knows today.
3Quality starts upstreamOutdated, duplicate or badly scanned documents produce wrong answers.
ExampleAn equipment distributor in Karaganda has years of manuals, price lists and support emails in Russian and Kazakh. With RAG, its support assistant finds the matching manual section and past fixes for each new request and drafts a reply with links to them. Support staff review the draft before it goes out.
On this topic
What we do
Related terms
LLMA large language model: an AI model trained on huge amounts of text to read and write language.Vector databaseA database that stores text or images as vectors so it can find items by meaning, not exact words.AI agentSoftware that uses a language model to plan the steps of a task and carry them out in your tools.AI hallucinationWhen an AI model states something false or invented as confidently as if it were a fact.Fine-tuningFurther training of an existing AI model on your own examples so it handles one task in a set way.
Need to put this in place?
Describe the task, we will name the scope, price and timeline. The price is fixed in the contract.
Your selectionAI agent developmentfrom 480 000 ₸
The team gets your request; we reply during business hours.Price fixed in the contract before you pay.No calls unless you ask.