AI answer engine
A system that responds to questions with a synthesized answer rather than a list of links — e.g. ChatGPT, Perplexity, Google AI Overviews, Gemini, Copilot, Claude, Grok.
AI Overviews (AIO)
Google's AI-generated answer layer in Search, powered by Gemini, grounded in the Google index via query fan-out; merged with AI Mode at I/O 2026. 2.5B+ monthly users as of Q2 2026.
AI visibility
The degree to which AI engines name, cite and correctly describe a brand in answers; measured by the four-dimension framework (Visibility, Citation Rate, Semantic Coverage, Brand Cognition).
Answer block
A 40–80 word passage that directly answers a specific question, placed under a question-shaped heading; the unit of content AI engines lift into answers.
Answer engine optimization (AEO)
A near-synonym of GEO emphasizing the answer surface; GEO additionally stresses the generative synthesis mechanics and citation economy.
Brand Cognition
ChinaGEO standard dimension 4: the factual accuracy, sentiment and cross-engine consistency of what AI answers say about a brand.
ChatGPT-User
OpenAI's on-demand fetch agent, retrieving a specific page when a user pastes or asks about it.
Citation Rate
ChinaGEO standard dimension 2: how often owned/earned sources are credited in AI answers, weighted by source authority tier.
ClaudeBot
Anthropic's training crawler; separate from user-triggered fetch agents used by Claude's web features.
Corroboration
The synthesis-stage filter by which claims repeated across independent sources are stated as fact while single-source claims are hedged or dropped.
Entity resolution
The process by which an engine determines which real-world thing a name refers to; clean resolution requires consistent facts and schema across sources.
Entity collision
The failure mode where an engine confuses your brand with a better-documented namesake, answering about them when asked about you.
E-E-A-T
Experience, Expertise, Authoritativeness, Trust — Google's quality framework, operationalized for AI as declared authorship, dates, corroboration and consistency.
GEO (Generative Engine Optimization)
The practice of making a brand findable, understandable and citable by AI answer engines so they name it in synthesized answers.
Google-Extended
Google's robots.txt token controlling use of content for Gemini/model training; does not affect Search indexing or AI Overviews grounding.
GPTBot
OpenAI's training-data crawler; independent of OAI-SearchBot, which serves ChatGPT search answers.
Grounding
Constraining a model's answer to retrieved source passages; the mechanism that makes AI search optimizable.
Hallucination (brand)
Confident false statements engines make about thinly-documented brands, generalizing from patterns instead of facts.
hreflang
HTML/link annotations declaring a page's language/region and its alternates; the technical layer of multilingual GEO.
IndexNow
A ping protocol that notifies Bing (and partners) of new/updated URLs immediately; shortens the path from publishing to AI-answer reflection.
Knowledge Graph
Google's structured entity database, feeding AI Overviews with entity facts; fed by schema, Wikidata and corroborated sources.
llms.txt
A Markdown convention at site root briefing AI systems on what the site is and where its citable content lives; complements robots.txt and sitemap.xml.
OAI-SearchBot
OpenAI's search-retrieval crawler for ChatGPT answers; the GEO-critical OpenAI agent.
PerplexityBot
Perplexity's live-retrieval crawler; Perplexity-User handles on-demand fetches.
Prompt set
A graded, versioned set of buyer questions used to measure AI visibility; the GEO replacement for the keyword rank list.
Query fan-out
Google's expansion of one question into hidden sub-queries, each retrieving its own passages for synthesis into an AI Overview.
RAG (retrieval-augmented generation)
The retrieve-then-generate architecture behind grounded AI answers; passages are the unit of competition.
Semantic Coverage
ChinaGEO standard dimension 3: the share of a category's buyer-question matrix in which engines can answer with the brand present.
Source graph
The web of third-party sources (directories, press, reviews, community) that corroborates a brand for AI synthesis.
Visibility
ChinaGEO standard dimension 1: the percentage of a graded prompt set in which an engine names the brand, scored 0–100 per engine.
Zero-click search
A search session resolved without any click to a website; roughly two-thirds of US Google searches in 2026 panels, higher where AI answers appear.