Brand Mentions and AI Visibility Glossary

The terms

These are the terms brand owners meet when they work on how AI engines see their brand. They are listed alphabetically. Where a term needs more than a paragraph, the entry links to the guide that covers it fully. The entries on paid links, disclosure and Google’s spam policies are general information, not legal advice.

AI Mode

Google’s conversational search mode, where you ask a full question and get a written answer with supporting links instead of a list of results. Google says AI Mode passed 1 billion monthly active users by July 2026 and has merged it with AI Overviews into one experience. See how brands get named in AI Overviews and AI Mode.

AI Overviews

The AI-written summaries Google shows at the top of some search results, with links to the pages it drew on. Google says there are no special optimizations needed to appear in them beyond being indexed and eligible for a snippet. See how brands get named in AI Overviews and AI Mode.

AI referral traffic

Visits to your website that come from a link inside an AI answer. ChatGPT adds utm_source=chatgpt.com to its links, so those visits show up in analytics. Other assistants appear as referrers. It is still small next to Google traffic, but growing. See does AI visibility bring real traffic and sales.

AI search

Any search experience where an AI model writes the answer instead of only listing links: ChatGPT search, Perplexity, Gemini, Copilot, Claude with web search, and Google’s AI Overviews and AI Mode. Most combine what the model learned in training data with pages retrieved live. See how AI engines decide which brands to mention.

AI visibility

How often, and how well, AI engines name and describe your brand when people ask about your category. It is usually measured as a mention rate across a prompt set, plus position and citations. It is a measurement, not a fixed rank, because answers vary run to run. See how to measure your brand’s AI visibility.

Answer engine optimization

Often shortened to AEO: work aimed at getting a brand or page used in direct answers, from featured snippets and voice assistants to AI chat. It overlaps heavily with generative engine optimization and with good SEO: clear facts, indexable pages and a brand other sites talk about. See GEO, AEO and SEO compared.

Brand mention

Any appearance of your brand’s name on a page someone else publishes, or inside an AI answer. It can be a linked mention or an unlinked mention. The most useful mentions sit on real, indexed publications and name you next to your category and use case. See what brand mentions are and why AI runs on them.

Citation

A source an AI answer links to, usually shown as a numbered reference or a card under the text. Being cited means the engine used a page as evidence. BrightEdge found ChatGPT mentions brands about 3.2 times more often than it cites them, so the two are measured separately. See the sources AI cites most.

Co-occurrence

Your brand name appearing near the words that describe what you do, such as your category, use case or city. When many independent pages put the same brand and topic side by side, it becomes easier for an engine to connect them. It is reasoning from how language models learn, not a confirmed ranking rule. See context beats volume.

Crawler

A program that visits web pages automatically and copies their content, either to build a search index or to collect training data. Examples are Googlebot, GPTBot, OAI-SearchBot, PerplexityBot and ClaudeBot. Site owners control most crawlers through robots.txt. See where ChatGPT gets its information about brands.

Digital PR

Earning coverage for a brand in news sites, magazines and blogs, usually by pitching journalists with stories, data or expert comment. Good digital PR produces strong brand mentions, but it is slow and unpredictable, because each journalist decides what to cover. See brand mentions vs digital PR.

Entity

A distinct thing a search engine or AI can identify and keep apart from others: a company, person, product or place. Being a clear entity means engines know that your brand name refers to you, what you do and how you relate to other things. See make your brand a known entity.

Generative engine optimization

Often shortened to GEO: work aimed at getting a brand named or cited in answers from generative AI such as ChatGPT, Perplexity and Google’s AI features. Despite the new name, much of it is familiar: crawler access, clear facts, consistent entity data and many credible mentions. See GEO, AEO and SEO compared.

Google-Extended

A robots.txt control, not a separate crawler, that tells Google whether your content may be used to train Gemini models and ground Gemini app answers. Google says it does not affect inclusion in Google Search or ranking, so blocking it does not remove you from AI Overviews. See Gemini, Claude and Copilot.

GPTBot

OpenAI’s crawler for collecting training data. OpenAI says disallowing it means a site’s content should not be used to train its foundation models. Blocking GPTBot does not remove you from ChatGPT search; that is controlled by OAI-SearchBot. See where ChatGPT gets its information about brands.

Hallucination

An AI answer that states something false with confidence, such as a wrong founding date, a product you do not sell or a source that does not exist. Thin or conflicting information online leaves a model more gaps to fill with guesses, so consistent, accurate facts across many sources reduce the room for errors (reasoning; no study has yet measured brand errors directly). See when AI gets your brand wrong.

Indexing

A search engine storing a page in its index so it can be found and shown. A page that is not indexed is invisible to search, and to AI features that rely on search. Google requires a page to be indexed and snippet-eligible before it can be a supporting link in its AI features. See where brand mentions should appear.

Knowledge cutoff

The date after which a model’s built-in knowledge stops. Anything newer, such as a launch or rebrand, reaches the model only through live search. Anthropic’s newest Claude models list a training data cutoff of June 2026, months before October 2026. See how a new brand gets into AI answers.

Knowledge Graph

Google’s database of entities and the facts that connect them: who founded a company, where it is based, what it makes. It is built automatically from many sources across the web. Being in it helps Google show accurate facts about your brand, including in a knowledge panel. See make your brand a known entity.

Knowledge panel

The information box Google shows for some people, organizations and things, built automatically from its Knowledge Graph. Some panels can be claimed by proving you control an official profile Google already links to the entity. Local businesses use Google Business Profile instead. See make your brand a known entity.

Large language model

Often shortened to LLM: the kind of AI behind ChatGPT, Claude, Gemini and Copilot. It learns patterns of language from huge amounts of text, then writes answers word by word. It does not look facts up unless connected to search, which is why training data and retrieval both matter. See how AI engines decide which brands to mention.

Linked mention

A brand mention with a hyperlink to your website. Links help people click through and help search engines discover pages. AI answers mainly read the words around your name, so the context of a mention often matters as much as the link. Paid links need rel=”sponsored” or nofollow. See unlinked brand mentions.

Mention rate

The share of AI answers that name your brand, across a fixed prompt set run several times. If you appear in 12 of 40 answers, your mention rate is 30%. It is the steadiest measure of AI visibility, because single answers change from run to run. See how to measure your brand’s AI visibility.

Native advertising

Paid content made to match the look of the publication it appears in, such as a sponsored article. It is legal when clearly labeled. The FTC wants such content labeled plainly as an ad, and UK and EU rules require similar disclosure. This is general information, not legal advice. See paid brand mentions and disclosure rules.

Nofollow

A link attribute, rel=”nofollow”, that tells search engines not to treat a link as an endorsement. Google accepts it on paid links, though it prefers rel=”sponsored” for ads and paid placements. Nofollow is also the catch-all for links a site does not want to vouch for. See paid brand mentions and Google’s link policies.

OAI-SearchBot

OpenAI’s crawler for ChatGPT search. OpenAI says sites that block it will not be shown in ChatGPT search answers, apart from possible bare navigational links. If you want to appear in ChatGPT, check your robots.txt does not block it. See where ChatGPT gets its information about brands.

PerplexityBot

Perplexity’s crawler for its answer engine. Perplexity says it is used to surface and link websites in its answers and is not used to train AI models. Blocking it makes a site less likely to appear as a cited source. See how Perplexity picks its sources.

Prompt

The question or instruction someone types into an AI tool. For brand measurement, the useful prompts are the ones buyers really ask, such as “what is the best CRM for a five-person sales team?” Small changes in wording can change which brands appear. See how to measure your brand’s AI visibility.

Prompt set

A fixed list of prompts you reuse every time you measure, built from real buyer questions about your category, use cases and market. Keeping the set fixed lets you compare results over time; running each prompt more than once evens out the randomness of AI answers. See how to measure your brand’s AI visibility.

rel="sponsored"

A link attribute Google asks sites to use on advertisements and paid placements. Google’s spam policies treat paid links that pass ranking credit as link spam, but accept them when marked sponsored or nofollow. It is a search rule, separate from the disclosure law that applies to readers. See paid brand mentions and Google’s link policies.

Retrieval-augmented generation

Often shortened to RAG: an AI model searching for relevant pages first, then writing its answer from what it found. It is how ChatGPT search, Perplexity and Google’s AI features can mention brands newer than their knowledge cutoff. It is also why indexing matters for AI visibility. See how AI engines decide which brands to mention.

robots.txt

A plain text file at the root of a website that tells crawlers which parts they may visit. It is how you allow or block GPTBot, OAI-SearchBot, PerplexityBot or Google-Extended. Blocking the wrong one can quietly remove you from an AI search tool. See why AI doesn’t mention your brand.

sameAs

A property in schema markup that lists your official profiles elsewhere, such as LinkedIn, Wikidata or a social account. schema.org defines it as a link that unambiguously indicates the item’s identity. It helps engines confirm that those profiles and your website are the same entity. See make your brand a known entity.

Schema markup

Structured data added to a web page’s code that describes things in a standard format search engines read, such as Organization, Product or LocalBusiness. Google says Organization markup helps it disambiguate your organization in results. It labels facts clearly; it does not replace mentions elsewhere. See make your brand a known entity.

Share of voice

Your brand’s share of all brand mentions in AI answers for your category, compared with competitors. If AI names your brand 15 times and three rivals 45 times across the same prompt set, your share of voice is 25%. It shows who the engines treat as the default choice. See how to measure your brand’s AI visibility.

Site reputation abuse

Google’s term for third-party content published on a host site mainly to benefit from rankings the host earned with its own content. Google has enforced it since May 5, 2024. Google says editorial work, and advertorials whose purpose is to share content directly with a publication’s readers rather than to manipulate rankings, are not covered. This is general information, not legal advice. See paid brand mentions and Google’s link policies.

Topic context

The subject matter around a mention: the category, use case, audience and market a page connects your brand to. A mention in a guide to vegan meal kits says something different from one in a list of logistics firms. Engines can only link you to topics you are actually written about. See context beats volume.

Training data

The text a large language model learns from before it is released, mostly public web pages, books and licensed sources up to a knowledge cutoff. What the web said about your brand before that date shapes what the model “knows” without searching. Newer facts must come from retrieval. See how AI engines decide which brands to mention.

Unlinked mention

A brand mention with no hyperlink. Google’s John Mueller has said a mention without a link does not pass link signals the way a link does. AI models, though, read words, not links, so an unlinked name in the right context can still inform an answer. See unlinked brand mentions.

Wikidata

A free, public database of structured facts that Wikipedia and many other projects draw on. A company can qualify without a Wikipedia article if serious, public references describe it. Keep any entry factual and sourced, and disclose any connection to the company. See make your brand a known entity.

Zero-click search

A search that ends without the person clicking any result, because the answer was on the page. Even before AI Overviews, SparkToro found nearly 60% of Google searches ended without a click. It is why being named in the answer matters. See does AI visibility bring real traffic and sales.