How to get cited by ChatGPT search: a practical playbook
Eight-step playbook covering robots.txt, schema, content shape, third-party mentions and llms.txt for getting your domain into ChatGPT search citations.
The AEO + GEO knowledge base
Every term, metric, engine quirk and tactical playbook we've documented across 24 glossary entries and 1 pillar guide. Search the lot or browse by section.
24 glossary entries
1 guide
Updated regularly
No results. Try a different term, or browse the full glossary.
Guides
Pillar pages: practical playbooks longer than a glossary definition.
Concepts
3 terms
AI search rankings
AI search rankings track how often AI engines (ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews) name a brand in their answers and cite the brand's domain as a source, across a fixed set of bu
AI search vs SEO
Where SEO measures ranking in a list of blue links, AI search measures whether your brand is named and cited in the synthesised answer. Different user behaviour, different metrics.
GEO (full form)
GEO stands for Generative Engine Optimisation — the practice of optimising content so generative AI engines cite and recommend your brand in their synthesised answers.
Engines
6 terms
How ChatGPT chooses citations
ChatGPT chooses citations by running a web search via its tool-use system, then re-ranking results through its own relevance model before generating the answer.
How Claude Web Search works
Claude's web search is a tool-use capability where the model decides when to query the web, what to query for, and which results to use as evidence. Often runs multiple search hops per answer.
How Gemini grounding works
Gemini grounding lets the model anchor answers in real web sources via Google's groundingMetadata. When enabled, Gemini returns the answer alongside a structured list of source URLs.
How Google AI Overviews pick sources
Google AI Overview picks sources via a retrieval pipeline that overlaps with classical Google ranking but weights different signals — particularly structured data, freshness, and passage extractabilit
How Perplexity ranks sources
Perplexity is a search-native AI — every answer is grounded in live web sources with inline numbered citations. It weights freshness, factual density, and primary-source signals more heavily than othe
Most cited domains in ChatGPT
The domains ChatGPT cites most across categories — consistently Reddit, Wikipedia, G2, Capterra, TechCrunch, Forbes, and the brand's own site, plus category-specific trade publications.
Metrics
5 terms
AI Overview mention
An AI Overview mention is when your brand name appears in the synthesised answer prose, distinct from a citation where only your URL appears in the source list.
Brand hallucination prevention
When AI engines fabricate false claims about your brand — invented features, wrong pricing, made-up partnerships. Caused by information vacuums where authoritative content is missing.
Citation rate (AI search metric)
Percentage of AI answers that include your domain in the source citation list. Measures who the AI trusts as evidence.
Share of voice in AI answers
Your brand visibility relative to the other brands the AI mentions in the same answer set. The most defensible AEO metric — normalises for query selection bias.
Visibility (AI search metric)
Percentage of AI answers that name your brand in the response prose. Measures who the AI recommends to the buyer.
Tactics
7 terms
AI Overview tracking API
An AI Overview tracking API captures Google AI Overview content for a fixed query set on a schedule, then exposes the data (text, citations, brand mentions, score deltas) for ingestion into a dashboar
AI reputation management
AI reputation management is the practice of monitoring and shaping how AI answer engines describe a brand: what they say, what they get wrong, and which sources they cite. The AI-search analogue to cl
AI visibility tools (the AEO tooling landscape)
AI visibility tools are software platforms that track how AI answer engines (ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews) describe a brand, name it in answers, and cite its domain. The AE
Brand hallucination monitoring
The detection pass that catches fabricated AI claims about your brand after they appear — recurring factual queries across engines, compared against ground truth.
Google AI Overview API
There is no official public API for Google AI Overview. Programmatic access goes through third-party Google search APIs that scrape the rendered SERP — SearchApi.io, SerpApi, Apify, DataForSEO — at ro
llms.txt examples + templates
Real-world examples of llms.txt files from sites that publish them, plus a working template you can copy. Updated as the standard evolves.
Structured data for AI engines
Structured data (JSON-LD) tells AI engines exactly what your page is about in a format they can parse unambiguously. The schemas that move the needle for AI citations are different from the schemas th
Google AI surfaces
3 terms
AI Overview rollout status (where it is live, where it isn't)
Google AI Overview is live across roughly 100+ countries as of mid-2026, on roughly 25 to 48 percent of eligible queries, with rollout still expanding. Specific availability depends on your account, r
AI Overview vs Knowledge Panel
AI Overview is a generative, per-query answer synthesised from live web sources. Knowledge Panel is a structured, per-entity card pulled from Google's Knowledge Graph. Different pipelines, different s
How to turn on AI Overview
AI Overview cannot be manually turned on — Google decides per-query based on query type, region, account signals, and Search Labs status. There is no user toggle.