Glossary: the language of AI visibility.

The practice of improving whether, and how accurately, a brand appears in the answers AI assistants and AI search features give. Where SEO aims for a place in a list of links, AEO aims for a place in the answer itself.

See also: GEO, AI visibility, AEO vs GEO vs SEO

A conversational tool, such as ChatGPT, Gemini, Perplexity or Copilot, that answers questions in natural language rather than returning a list of links.

A conversational search experience from Google that answers complex questions with an AI-generated response and links, and lets people ask follow-up questions.

An AI-generated summary Google shows at the top of some search results, with links to the pages it draws on.

How often, and how well, a brand appears in AI-generated answers. It covers whether the brand is named, how it is described and which sources support it.

See also: Answer Share

Be the Answer termRead the method →

The path from a buyer's need to a decision when AI is involved: buyer, trigger, job to be done, question, AI answer, sources, shortlist, decision.

Any system that responds to a question with a synthesised answer rather than a list of results. AI assistants and AI search features are answer engines.

Be the Answer termRead the method →

The difference between the questions where a brand should reasonably be considered and the questions where AI actually names it.

Be the Answer termRead the method →

The percentage of relevant buyer questions in which an AI assistant meaningfully names a brand. If AI names you in 31 of 100 eligible questions, your Answer Share is 31%.

A question that names a specific brand, such as asking what a named company offers. Brands usually appear in answers to branded questions; the harder test is unbranded ones.

See also: Unbranded question

A question a real buyer would ask an AI assistant while making a decision, written in their words and with their context.

See also: Question set

A link or reference in an AI answer that points to a source it drew on. Not every platform cites sources, and not every source used is cited.

Be the Answer termRead the method →

The percentage of answers that draw on a brand's own content or on third-party sources about it. We report owned and earned citations separately.

Software that visits web pages to collect their content. Search engines and AI companies use crawlers to index the web and, in some cases, to gather content for AI answers or model training.

See also: robots.txt

Experience, expertise, authoritativeness and trustworthiness: the qualities Google's search quality guidelines use to describe reliable content. It is a framework for assessing quality rather than a single measurable score.

Content about a brand published by someone else, such as reviews, news coverage, analyst notes or directory listings.

See also: Owned source

A distinct, identifiable thing, such as a company, product, person or place, that search engines and AI systems recognise and can connect to facts about it. Clear, consistent information helps systems identify a brand as the right entity.

Be the Answer termRead the method →

How often a brand is named first when an AI answer lists several options. It is not a ranking: AI answers rarely claim to rank the options they list.

The practice of improving how content and brands are cited and recommended by generative AI. The term was introduced in 2023 academic research and is often used interchangeably with AEO.

See also: AEO, AEO vs GEO vs SEO

AI that produces new content, such as text or images, in response to a prompt, rather than only retrieving existing content.

Connecting an AI answer to external, retrievable sources, such as web pages, so the response draws on current information rather than only on what the model learned in training.

See also: Retrieval-augmented generation

An AI response that states something false or unsupported as if it were fact, such as an outdated price or a service a company does not offer.

What a buyer is actually trying to achieve, as opposed to the product they end up buying. Starting from the job produces more realistic buyer questions than starting from keywords.

The point after which a model has no information from its training data. Answers about later events depend on the system retrieving current sources.

A structured network of entities and the relationships between them, which search engines use to understand and connect facts.

The type of AI model behind most AI assistants. It is trained on large amounts of text to predict and generate language.

A proposed file that summarises a website for AI systems, similar in spirit to a sitemap. It is not an official standard, and Google has said it does not affect search visibility.

The recurring descriptions an AI uses about a brand, such as premium, affordable or enterprise-focused. We report them with how often each appears.

See also: How we measure

Content a brand publishes and controls, such as its website, product pages and documentation.

See also: Earned source

The text a person gives an AI assistant: a question, an instruction or both.

A technique in which an AI search system breaks one question into several related searches and combines the results into a single answer. Google has described its AI Mode as working this way.

The fixed set of buyer questions used to measure AI visibility, tagged by buyer, buying stage and question type, and kept the same over time so results can be compared.

See also: How we measure

Be the Answer termRead the method →

How often a brand is presented as an option worth considering, not just mentioned.

A method in which an AI system retrieves relevant documents first and then generates its answer using them. It is one of the main ways AI answers are grounded in current sources.

A file on a website that tells crawlers which parts of the site they may visit. Many AI companies publish the names of their crawlers so site owners can allow or block them.

A brand's share of the total mentions, impressions or advertising in a market. Unlike Answer Share, it is not measured question by question in AI answers.

Be the Answer termRead the method →

How often a brand makes the finite list of options in answers to recommendation questions.

Code added to a web page, often using schema.org vocabulary, that labels information such as an organisation's name, products or FAQs so machines can read it reliably.

A question that doesn't name any brand, such as asking which providers suit a particular need. This is where shortlists are formed, before buyers know which names to ask about.

See also: Branded question

A search that ends without the person clicking through to a website, because the answer was shown directly on the results page or in an AI answer.