AI systems · Technique
Retrieval-augmented generation
Also known as: RAG
What is RAG?
Retrieval-augmented generation (RAG) is a technique in which a language model retrieves relevant documents first and then writes its answer grounded in them.
Most answer engines are RAG systems at heart. That is why chunk-level clarity matters: retrieval works on passages, so a passage that names its entity and states its fact without needing the rest of the page is easier to retrieve and safer to quote.
Related entities
- Answer-first writingAnswer-first writing opens every section with one sentence that fully answers the section's question and names its subject, before any context or caveat.
- AI crawlersAI crawlers are the user agents AI companies use to fetch web pages — some for model training, some for live search and answers — each controllable separately in robots.txt.
- Generative Engine OptimizationGenerative Engine Optimization (GEO) is the work of making a brand one of the sources a generative AI assistant names and cites when it composes an answer.
- Knowledge GraphA knowledge graph stores facts as entities connected by typed relationships, so a machine can answer questions by traversing connections instead of matching text.
- Google AI OverviewsAI Overviews are Google's generated summaries at the top of search results, composed by a Gemini model from retrieved web pages and shown with source links.
Field notes covering RAG
- AI Pages vs. AI Architecture: Why Volume LostCreating pages with AI multiplies documents. Having AI build entity hubs, internal links and wiki structure multiplies relationships — and that is what answer engines quote.
- How to Build a Site-Wide Entity GraphStable @ids, hub pages, about and mentions, sameAs, and one registry that keeps markup, links and llms.txt from drifting — a seven-step build guide.
- Your Website Is Not a Stack of Pages. It Is a Knowledge Base.Entities, relationships, schema, internal links, and answer-ready evidence: the architecture machines need when pages alone are no longer enough.
- Answer-First Writing: The Inverted Pyramid for MachinesPut the answer in sentence one under a question-shaped heading. The four rules that make a paragraph quotable by snippets, voice and AI assistants — with an interactive before/after.
- How AI Assistants Read a Website: Crawl to CitationThe five machines between your page and a recommendation — crawl, render, chunk, retrieve, compose — and an interactive robots.txt matrix for AI crawlers.
Sources & identifiers
Search disciplinesSite structureAI systemsLocal searchDistribution & growthAgentic commerceHover to trace relationships · click to open