The essential glossary for understanding how brands get discovered, understood and recommended by AI. Search is changing: people no longer just type keywords and pick from ten blue links. They ask AI systems questions and expect a synthesised answer. That shift has introduced a new vocabulary around Answer Engine Optimization (AEO), Generative Engine Optimization (GEO) and AI visibility.
This glossary defines the terms shaping the next era of digital discovery: from query fan-out and entity authority to AI citations, share of answer, and recommendation rate. Every entry is written the way the AEO/GEO playbook prescribes: answer-first and self-contained, so the glossary itself is built to be cited by the engines it describes.
Definitions are presented as self-contained, answer-first blocks so each concept can be understood independently.
An AI citation is the reference an answer engine gives to a source it draws on when composing a response. In AI-first discovery it is often the only impression a brand earns, since users rarely click through. Winning citations, not rankings, is the core objective of AEO and GEO.
AI crawlers are the bots AI systems use to read web content for answers. The key ones are GPTBot and OAI-SearchBot (OpenAI/ChatGPT), ClaudeBot (Anthropic) and PerplexityBot (Perplexity). Allowing them in robots.txt is a prerequisite for being eligible to appear in any AI-generated answer.
AI discovery is the process by which an AI system encounters and identifies a brand, organisation, person, product or source as relevant to a user's query. It is the first step in AI visibility — before a brand can be cited or recommended, it must be discoverable.
AI-first content is written and structured for both human readers and machine interpretation at once — clear, contextual, evidence-based and easy to retrieve. It is designed so AI systems can understand, extract and reference it, not only so people can read it.
An AI-first content strategy is planned from the outset for both human audiences and AI-mediated discovery. It considers how every piece can be understood, retrieved, synthesised and referenced by AI systems — treating machine legibility as a first-order goal, not an afterthought.
AI-first discovery is the model in which users get answers directly from AI systems rather than clicking through ranked links. The synthesised answer arrives before the sources, and the citation inside it is often the only impression a brand receives. It reframes visibility from ranking on a page to being named in the answer.
An AI mention is the appearance of a brand, organisation, person, product or concept in an AI-generated response, whether or not it carries a direct citation. Mentions without attribution still shape perception and count toward AI share of voice.
AI Overviews are Google's AI-generated summaries that appear above traditional results, resolving a query on the results page itself. They reach over two billion people monthly and now appear on roughly half of all searches. Users rarely click inside them, which makes the citation itself the primary brand impression.
AI reputation is the way an AI system consistently represents, describes and contextualises a brand or entity across relevant queries. Unlike a review score, it is inferred from the whole web of sources a model has learned to trust — making corroboration and consistent messaging central to managing it.
AI search is any search experience in which artificial intelligence interprets a query and generates, summarises or synthesises an answer, rather than simply returning traditional links. It spans Google AI Overviews, ChatGPT, Perplexity and similar tools.
AI share of voice is the proportion of relevant AI-generated responses in which a brand appears, compared with competitors in the same category. It is the industry-generic term for what Gutenberg tracks as Share of Model.
AI visibility is the extent to which a brand, organisation, product, expert or point of view is discovered, understood, mentioned, cited or recommended by AI systems. It is the overarching outcome that AEO and GEO are built to improve.
An AI visibility score is a measurement designed to assess a brand's visibility across a defined set of AI queries. Depending on methodology, it may combine mentions, citations, prominence, recommendations and competitive visibility into a single trackable figure.
An answer engine is any system that responds to a query with a direct, synthesised answer rather than a list of links — Google AI Overviews, ChatGPT, Perplexity and similar tools. Optimising for answer engines means structuring content so it can be extracted and cited within that answer.
Answer Engine Optimization is the practice of structuring content so AI answer engines can extract, trust and cite it. It relies on answer-first passages, clean heading hierarchy, tables, FAQs, schema markup and text that renders without JavaScript. AEO extends SEO; success shifts from where a page ranks to whether it is cited.
An answer-first block is a short, self-contained passage — typically 40 to 60 words — that answers a question directly at the top of a section, before any elaboration. It is the single most extractable content unit for AI engines, written to be lifted whole into a generated answer.
An authoritative source is one AI engines treat as reliable enough to cite, based on corroboration, credentials and consistent presence. Engines weigh authority differently — ChatGPT leans encyclopaedic, Perplexity favours community and recency — but across all of them, becoming an authoritative source is the end goal of AEO and GEO.
A brand entity is the uniquely identifiable representation of a company, organisation, product or brand within digital information systems. Establishing a clear, consistent brand entity lets AI systems recognise and correctly attribute references to it.
A brand footprint is the overall digital presence of a brand across websites, publications, platforms, communities, structured databases and AI-generated experiences. The wider and more consistent the footprint, the more corroboration AI systems have to draw on.
A brand mention is any reference to a brand across the web, linked or not. In AI visibility, mentions matter more than backlinks: analysis of tens of thousands of brands found they correlate roughly three times more strongly with AI visibility. Where a link once passed authority, a mention now passes trust.
Citability is how easily an AI engine can lift and attribute a passage as a source. A citable passage is self-contained, factually clear, attributed to credible data and structured to stand alone when extracted. High citability is the on-page goal of AEO — making content the machine wants to quote.
Citation rate is the percentage of relevant AI responses in which a particular source or domain is cited. It measures how often a brand's own content is doing the work of earning citations, as distinct from being mentioned without attribution.
Citation-worthy content contains enough originality, authority, specificity, evidence or expertise to be valuable as a source for AI-generated answers. Original research, proprietary data and a clear expert point of view are what most reliably make content worth citing.
Content chunking is structuring content into concise, self-contained sections so individual passages can be retrieved and understood independently by search and AI systems. It is the on-page technique that makes passage-level retrieval possible.
A conversational query is a question expressed in natural language, the way someone would speak it, rather than in clipped keyword syntax. AI search is optimised for these, which is why question-phrased headings and direct answers now matter more than keyword strings.
Corroboration is the repeated, consistent mention of a brand across many independent sources, which signals to AI engines that it is trustworthy. Generative models reward the most corroborated brand, not the loudest. Building it means earning presence across the specific platforms models learn to trust — press, community and reference sites.
Crawlability is the ability of automated systems to access and discover content on a website. If a page cannot be crawled, it cannot be retrieved, cited or recommended — making crawlability the technical precondition for all AI visibility.
Digital authority is the credibility an entity builds through authoritative content, independent references, citations, links, research and third-party validation across the web. It is what moves a brand from merely present to trusted enough to cite.
Digital PR is the practice of earning brand mentions and coverage across third-party sites, communities and publications. In AI visibility it functions as discovery infrastructure: the web-wide mentions that generative models learn to trust. Earned media has moved from an awareness tactic to a technical requirement for citation.
E-E-A-T stands for Experience, Expertise, Authoritativeness and Trust — Google's framework for assessing content credibility. It is not itself an AI ranking formula, but its principles — signalled through author credentials, original research and consistent reputation — shape whether AI systems treat a source as reliable enough to cite.
An entity is a uniquely identifiable person, organisation, product, place, concept or other distinct subject that AI and search systems can track. Entities, not keywords, are how modern systems organise and connect information.
Entity association is the set of relationships an AI or search system establishes between an entity and related people, topics, products, attributes or concepts. Strong, accurate associations are how a brand becomes linked to the topics it wants to own.
Entity authority is the strength and credibility of an entity's digital presence across trusted sources and knowledge systems. Built through consistent representation and corroboration, it is among the strongest predictors of whether AI systems will cite a brand.
Entity clarity means describing an entity consistently everywhere it appears, so AI systems can confidently connect the references to one thing. Clear, consistent entities are among the strongest citation predictors on engines like ChatGPT.
Entity disambiguation is the process of distinguishing between different entities that share the same or similar names. Consistent identifiers, structured data and knowledge-graph presence help AI systems attribute information to the right entity.
Entity recognition is an AI or search system's ability to correctly identify a person, company, product, place or concept within content. It is the first step that lets a system connect content to an entity and, in turn, to a brand.
FAQ schema is a structured-data type that marks up question-and-answer pairs so AI systems recognise them as direct answers. Applied to question-phrased content, it makes passages explicitly machine-readable as responses to real queries — a high-value, low-effort AEO signal.
Generative brand presence is the visibility and contextual representation of a brand within AI-generated experiences. It captures not just whether a brand appears, but how accurately and favourably generative systems describe it.
A generative engine is an AI system that composes original responses by synthesising information across many sources, rather than retrieving a single page. Because it assembles rather than ranks, brands earn visibility by being widely mentioned and corroborated across the web, not by holding a single top position.
Generative Engine Optimization is the practice of earning presence and mentions across the open web so generative models surface a brand when they compose an answer. It is the reputation layer beneath every citation — built through earned media, community presence and consistent entity signals, not on-page structure alone.
Generative search uses generative AI to construct answers by synthesising information from multiple sources, rather than pointing to a single ranked page. It is the mechanism behind AI Overviews and answer engines.
Grounded generation is producing AI responses using retrieved or verified information rather than a model's internal knowledge alone. It is the practice that turns grounding into output — and it is why well-structured, citable content gets quoted.
Grounding is the practice of anchoring an AI's answer to verifiable external sources rather than its internal parameters alone. Grounded answers carry citations, which is where brand visibility is won. Content that is clear, structured and attributed is easier for a model to ground its answer in — and therefore to cite.
A hallucination is a confident but false or unsupported statement produced by an AI system, generated without sufficient grounding in reliable sources. For brands it is a reputation risk: a model may misstate facts it cannot corroborate. Clear, consistent, well-attributed content reduces the chance of being described incorrectly.
Information gain is the degree to which a piece of content adds genuinely new, useful or distinctive information beyond what already exists. High information gain — original data, fresh analysis, expert insight — is what makes content citation-worthy.
Information retrieval is the process by which a system identifies and returns relevant information from a collection of sources. It is the retrieval half of retrieval-augmented generation, and the reason retrievability matters.
Interpretability is how easily an AI system can understand the meaning, context, entities and relationships within content. The more interpretable a page, the more reliably it can be extracted and cited — clear structure and language raise it.
A knowledge graph is a structured map of entities and the relationships between them that AI systems use to understand and verify facts. Presence in graphs such as Google's Knowledge Graph and Wikidata helps models recognise a brand as a legitimate entity, strengthening its eligibility to be cited.
A knowledge panel is a search-interface element that displays structured information about an identified entity. It is a visible sign that a brand is recognised as an entity in a provider's knowledge graph.
llms.txt is a proposed plain-text file that maps a site's key pages for AI systems, similar in spirit to a sitemap. It is a useful, low-cost hygiene signal, but as of 2026 no major AI provider has confirmed reading it in production. Treat it as good practice, not a guaranteed channel.
Machine readability is how easily automated systems can parse, interpret and process digital content. Structured data, clean HTML and server-side rendering all raise it — and content a machine cannot read is invisible to AI.
Mention rate is the percentage of tracked AI responses in which a particular brand or entity appears. It is a core AI-visibility metric, capturing presence whether or not a citation is attached.
Passage-level optimization is the practice of making each passage independently extractable, rather than optimising a page as a whole. Because AI engines lift individual passages to compose answers, every section leads with a direct response, so any part of the page can stand alone as a citable unit.
Passage retrieval is the retrieval of a specific section of a document rather than a whole webpage. Because AI systems increasingly pull passages, not pages, structuring content into self-contained chunks directly improves what gets surfaced.
Prompt coverage is the percentage of strategically relevant AI prompts for which a brand achieves measurable visibility. It answers a blunt question: across the prompts that matter, how often do you show up at all?
A prompt set is the curated list of real questions a brand's buyers put to AI systems — the prompts that decide whether it is mentioned. Mapping the prompt set, rather than keywords alone, is the starting point of AEO: it defines the answers a brand needs to own.
The prompt-to-brand journey is the path from a user's initial AI query through discovery, consideration and potential recommendation of a brand. Mapping it shows where in the AI conversation a brand needs to be present to influence the outcome.
Query expansion is the process of adding related concepts, synonyms, entities or contextual terms to a search to improve retrieval. It is one reason breadth of relevant, well-linked content helps: it gives systems more ways to match you.
Query fan-out is the process by which an AI search system breaks a complex query into multiple related sub-queries, gathering information from different sources before composing an answer. It means a single user question can be won or lost across many underlying searches.
Query intent is the underlying objective or information need behind a user's search or AI prompt. Matching content to intent — not just to words — is what makes it semantically relevant and eligible to be surfaced.
A recency signal is any visible marker that content is current — a publish date, a last-updated date or a referenced year. AI engines, and Perplexity especially, favour fresh content: a page revised this quarter beats an identical page last touched two years ago. Recency is itself a citation factor.
Recommendation likelihood is the probability that an AI system will recommend a particular entity for a specific user intent. It shifts the goal beyond being mentioned to being actively put forward as the answer.
Recommendation rate is the percentage of relevant AI prompts in which a system recommends a particular brand, product, service or source. It is the sharpest end of AI visibility — the point where discovery converts to endorsement.
Retrievability is how easily an AI or search system can discover and retrieve a piece of content. It depends on crawlability, structure and machine readability — and without it, quality content still goes unseen.
Retrieval-Augmented Generation is the technique by which an AI system fetches relevant documents in real time and uses them to compose its answer, rather than relying on training data alone. It is why fresh, well-structured, crawlable pages can be cited — the model retrieves and quotes them at query time.
robots.txt is the file that tells crawlers which parts of a site they may access. For AI visibility it must explicitly allow the AI crawlers — GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot — otherwise content is excluded from the systems where citations are won.
Schema markup is code, usually written in JSON-LD, that labels content so machines understand what each element means — a definition, an FAQ, an author, a date. It makes a page legible to AI systems, improving the odds that its content is correctly extracted, attributed and cited.
Schema.org is the collaborative vocabulary standard used to structure information about entities and content on the web. It is the shared dictionary that schema markup draws its types and properties from.
Search Engine Optimization is the practice of improving a page's ranking in traditional search results. It remains the technical foundation that feeds AEO and GEO — crawlability, content and authority still matter — but its success metric, position on the page, no longer captures visibility inside AI-generated answers.
Semantic relevance is the degree to which content meaningfully corresponds to the concepts and intent behind a query, beyond matching exact words. It is what semantic search rewards.
Semantic search evaluates the meaning, relationships and context behind a query rather than relying on exact keyword matches. It is why entity clarity and topical depth now outperform keyword density.
Server-side rendering delivers a page's full content in the initial HTML, before any JavaScript runs. It matters for AI visibility because many crawlers do not execute JavaScript: content that appears only after scripts load can be invisible to them. SSR ensures the machine can read what users see.
Share of answer is the proportion of the information, recommendations or references within a single AI-generated answer that can be attributed to a particular brand or source. Where Share of Model counts appearances across answers, share of answer measures dominance within one.
Share of Model is the measure of how often a brand is cited or mentioned across AI engines for its priority prompts, relative to competitors. It is the AI-era equivalent of share of voice — tracking presence in generated answers rather than rankings or clicks, engine by engine.
Source authority is the perceived reliability, expertise and credibility of a source an AI system retrieves from. Higher source authority raises the chance that a page is chosen to ground an answer.
Source diversity is the breadth of independent, credible sources that support or reference an entity, claim or topic. Diverse corroboration signals trustworthiness more strongly than repetition from a single source.
Structured data is machine-readable information that helps search and AI systems understand the meaning and attributes of a page's content. Schema markup is the most common way to add it.
Third-party validation is independent recognition of a brand, claim or expertise through credible publications, institutions, analysts, reviews or research. It is the earned, external proof that AI systems weigh more heavily than self-description.
Topic authority is the depth, breadth and credibility of a brand's content and expertise around a subject area. Comprehensive, interlinked, expert coverage is how a brand becomes the source AI systems associate with a topic.
Visibility is the extent to which a brand or entity can be discovered across a given search, digital or AI ecosystem. AI visibility is the subset of it that plays out inside AI-generated answers.
A visibility gap is the difference between a brand's desired presence and its actual visibility across relevant queries, AI responses or competitive categories. Measuring the gap is the starting point of any AI visibility programme.
Zero-click search is a query that ends without the user clicking any result, because the answer is delivered on the surface itself. Well over half of searches now end this way. It shifts the value of content from driving traffic to earning the citation that shapes the answer.
Mmeasure where you stand, build your presence, then own the category. Definitions below are drafted from the tier names and playbook framework — confirm against final scoping before publication.
Visibility Pulse is Gutenberg's entry AEO/GEO tier: a diagnostic that measures how visible a brand currently is across AI engines. It maps the priority prompt set, benchmarks Share of Model against competitors and identifies the citation gaps to close — establishing the baseline every AI visibility programme is built on.
Citation Partner is Gutenberg's managed AEO/GEO tier: an ongoing programme that actively builds a brand's presence in AI answers. It combines answer-first content, structured-data implementation and digital-PR signal-building, tracked monthly against Share of Model — moving a brand from measured to actively cited.
Citation Command is Gutenberg's enterprise AEO/GEO tier: a full-scale programme to make a brand the default answer in its category. It runs content, PR and digital as one integrated system — original research, entity and knowledge-graph management, per-engine optimisation and continuous Share of Model tracking — to own the AI conversation.
Use this glossary as the foundation for understanding AI discovery, citation, recommendation and the visibility systems built around them.
Explore AI Visibility