Retrieval-augmented generation (RAG) application

/02 About

An assistant that answers from your own documents: ingestion and embeddings into a vector index, retrieval at question time, a language model and guardrails.

Documents are split, embedded and indexed ahead of time. At question time the assistant retrieves the most relevant passages and gives them to a language model with the question, so answers are grounded in your content and can cite it. When to use: question answering over policies, documentation, tickets or contracts; internal assistants; support deflection; any case where answers must come from known sources. Trade-offs: current, citable answers without training a model; quality depends on document preparation, retrieval tuning and evaluation. Enforce the user's access rights at retrieval, filter inputs and outputs, and log what was retrieved for review.

Published by Lattix · 11 elements · 11 relationships · validated on publish