Every term in the AI citation and optimization stack — defined precisely. No jargon without explanation. Updated as the experiment produces new findings.
// maintained by GeoExperiment Team & CC
// updated 2026-06-07
// 24 terms defined
// direct answer
// what does this glossary cover?
This glossary defines every key term used in GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), LLM citation, schema markup, and technical SEO. Each definition is written to be directly extractable by AI systems — precise, complete in one paragraph, with no ambiguity.
// A
AEO — Answer Engine Optimization
// AEO
AEO is the practice of structuring content to appear as the direct answer in AI-powered search surfaces — specifically Google AI Overviews, featured snippets, and Bing AI answers. AEO requires placing the direct answer at the top of every page, using question-based headings, writing concise 40–60 word definitions, and implementing FAQ schema markup. Unlike GEO, AEO still depends on traditional Google ranking — you must rank to be featured.
// example: A page that answers "What is AEO?" with a one-paragraph definition at the top, before any background context, is structured for AEO.
content strategygoogle ai overviewsfeatured snippets
AI Overview is a Google Search feature that generates an AI-written summary at the top of search results, pulling from indexed web pages to synthesize an answer. Formerly called Search Generative Experience (SGE). Pages cited in AI Overviews gain visibility without requiring a click — but only pages Google has already indexed and ranked are eligible. Optimizing for AI Overviews is a core AEO objective.
// example: Searching "what is schema markup" on Google may trigger an AI Overview that synthesizes answers from 3–5 indexed sources, citing each with a link.
Answer-first structure is a content writing pattern that places the direct answer to a question at the very top of a page — before any context, history, or background. It is the single highest-impact signal for both AEO and GEO. AI systems scan for the first clear, complete statement on a page. If the answer is buried in paragraph three, the page is skipped in favor of one that leads with the answer. Every page on GeoExperiment applies this pattern.
// wrong: "Before we explore what GEO means, let's look at the history of search..." // correct: "GEO (Generative Engine Optimization) is the practice of structuring content so AI systems cite it in their responses."
Citation rate is the percentage of AI system queries on a given topic that result in a specific page or domain being cited as a source in the AI's response. GeoExperiment tracks citation rate as its primary experiment metric — querying six LLMs weekly with identical prompts and logging which pages are cited, how often, and in response to which queries. A citation rate of 0% means the AI never referenced the page; 100% means every query resulted in a citation.
// example: If 4 out of 6 LLMs cite geoexperiment.com/learn/ when asked "what is GEO," the citation rate for that page on that query is 67%.
Core Web Vitals are Google's standardized performance metrics used as a page experience ranking signal. The three metrics are: LCP (Largest Contentful Paint, target <2.5s), INP (Interaction to Next Paint, target <200ms), and CLS (Cumulative Layout Shift, target <0.1). Pages that fail CWV thresholds carry a ranking disadvantage. GeoExperiment tracks whether CWV scores affect AI citation rates as part of the technical experiment layer.
// example: A page with an LCP of 4.2 seconds fails the CWV threshold and may rank lower in Google — reducing its chance of being indexed and subsequently cited by AI systems.
E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness — Google's quality framework for evaluating content and its sources. Experience refers to first-hand engagement with the topic. Expertise means demonstrated knowledge. Authoritativeness means recognition by others in the field. Trustworthiness means accurate, transparent, reliable content with clear authorship. E-E-A-T is not a direct ranking score but informs Google's quality rater guidelines and algorithm design. It is also a strong LLM citation signal — AI systems favor sources with clear authorship and verifiable credentials.
// example: An About page listing the author's name, credentials, and linked professional profiles strengthens E-E-A-T signals for the entire domain.
An entity is a named, uniquely identifiable concept — a person, organization, place, product, or idea — that search engines and AI systems recognize and track across the web. Google's Knowledge Graph is built on entities. For GEO, entity consistency means your brand name, author name, and key facts must appear identically across your website, schema markup, social profiles, and third-party mentions. Inconsistent entity signals cause AI hallucinations about your brand.
// example: If your schema says "GeoExperiment" but your LinkedIn says "Geo Experiment" (with a space), these are treated as separate entities — fragmenting authority.
A featured snippet is a Google SERP feature that displays a direct answer — extracted from a web page — above the standard organic results. Featured snippets are the traditional AEO target. Pages that win featured snippets typically have a clear, concise answer near the top of the page, question-based headings, and structured formatting. Winning a featured snippet increases the probability of being included in a Google AI Overview for the same query.
// example: A search for "what is topical authority" may show a featured snippet quoting the first paragraph of a well-structured definition page.
GEO is the practice of structuring and optimizing content so that AI systems — including ChatGPT, Perplexity, Claude, and Gemini — cite it as a trusted source in their generated responses. GEO differs from SEO in that it targets AI citation rather than search engine rankings, and from AEO in that it targets standalone AI systems rather than AI features inside search engines. The five core GEO signals are: answer-first structure, definitive language, topical authority, original data, and structured JSON-LD markup.
// example: A page that answers "what is GEO?" with a direct definition above the fold, structured FAQ schema, and original experiment data is optimized for GEO.
JSON-LD (JavaScript Object Notation for Linked Data) is the recommended format for adding schema markup to web pages. It is placed in a <script type="application/ld+json"> tag in the page head or body, and is read by Google, Bing, and AI crawlers to understand the content type, authorship, and structure of a page. JSON-LD is preferred over Microdata because it does not require modifying HTML structure — it sits separately and is easier to validate and maintain.
// example: Adding a FAQPage JSON-LD block to a page tells Google and AI systems exactly which questions the page answers and what the precise answers are.
An LLM (Large Language Model) is an AI system trained on large volumes of text data to generate, summarize, translate, and answer questions in natural language. Major LLMs include GPT-4o (OpenAI / ChatGPT), Claude (Anthropic), Gemini (Google), Llama (Meta), and Mistral. LLMs answer questions either from their training data (static knowledge with a cutoff date) or by performing live web searches (real-time retrieval). GeoExperiment actively queries six LLMs weekly to track citation behavior.
// example: When a user asks Perplexity "what is generative engine optimization," the LLM performs a live web search, retrieves relevant pages, and cites the most authoritative source in its response.
LCP measures how long it takes for the largest visible content element — typically a hero image or headline — to render in the viewport. Google's threshold for a "good" LCP score is under 2.5 seconds. LCP is one of three Core Web Vitals and directly affects Google search ranking. Slow LCP also impacts AI citation indirectly — pages that are slow to load may be de-prioritized by AI crawlers like PerplexityBot.
// example: A hero image that is not preloaded and takes 4 seconds to appear gives a poor LCP score, signaling a slow page experience to Google.
Schema markup is structured data added to a web page to help search engines and AI systems understand its content, type, and relationships. It is defined by Schema.org and implemented using JSON-LD (recommended), Microdata, or RDFa. Schema markup is a required signal for both AEO and GEO — without it, AI systems must infer content meaning from raw HTML, which is less reliable. As of 2026, Google AI Mode uses schema as a trust signal, not just a display trigger for rich results.
// example: Adding Article, FAQPage, and BreadcrumbList schema to a learn page tells AI systems exactly what type of content it is, who wrote it, and where it sits in the site structure.
SEO is the practice of optimizing web content to rank higher in traditional search engine results pages (SERPs). Core SEO signals include backlinks, on-page keywords, technical site health, Core Web Vitals, and E-E-A-T. SEO is the foundation that makes AEO and GEO possible — a page that Google cannot crawl and index cannot be cited by AI systems that depend on Google's index. As of 2026, SEO alone is insufficient for AI visibility: fewer than 20% of top Google results are also cited by AI systems.
// example: A page with strong backlinks, fast load time, and clear heading structure ranks well on Google — but may still never be cited by ChatGPT if it lacks answer-first structure and schema markup.
SGE (Search Generative Experience) was Google's original name for its AI-powered search feature, now officially renamed Google AI Overview. The feature generates an AI-written summary at the top of search results for qualifying queries. The term SGE is deprecated as of 2024 — use "AI Overview" in current documentation and content. References to SGE in older guides still refer to the same feature.
// note: Any guide referencing "optimizing for SGE" is using outdated terminology. The feature and optimization principles are the same; only the name changed.
Topical authority is the degree to which a website is recognized as a trusted, comprehensive source on a specific subject by both search engines and AI systems. It is built through consistent, deep coverage of one subject area — interconnected pages covering every angle of a topic, not a mix of unrelated articles. Topical authority is one of the five core GEO signals and a primary factor in whether AI systems treat a source as citable. GeoExperiment builds topical authority by publishing exclusively about GEO, AEO, and LLM citation.
// example: A site publishing 20 interconnected pages about GEO — definitions, experiments, case studies, glossary, guides — builds stronger topical authority than a site with one GEO article among 200 unrelated posts.
RAG (Retrieval-Augmented Generation) is the technique AI systems use to fetch external information at query time before generating a response. When an LLM like Perplexity or ChatGPT (with search enabled) answers a question, it retrieves relevant web pages, extracts key information, and incorporates it into its generated response — citing the sources it used. RAG-based systems can cite your content within days of indexing, unlike training-data-only systems which require a full model retraining cycle.
// example: Perplexity uses RAG for every query — it searches the live web, retrieves the top results, and cites them directly. This is why Perplexity citation is the fastest proof point for GEO.