AI search is changing how decisions get made. Users are no longer scanning results. They're accepting answers. Which means: If you're not cited, you're not considered.
This glossary breaks down the terms that actually matter. Not just definitions, but what they mean in practice.
When an AI system selects your content as a source in its generated answer. This is the closest equivalent to "ranking" in AI search, but it is more selective. AI does not show 10 options like Google. It shows a handful. If you are not one of them, you do not exist in that moment.
Citations can take several forms: a direct brand mention by name, a link to your URL as a source, a paraphrased reference to your content without naming you, or inclusion in a comparison or recommendation list. Each carries different weight and visibility.
Search experiences powered by large language models that return direct answers instead of a list of links. Examples: ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews.
The key shift: users are skipping the results page entirely. Instead of scanning 10 blue links and clicking one, they read a generated answer that synthesises information from multiple sources. If your content was used to construct that answer, you were cited. If not, you were invisible.
The next evolution beyond AI search. Instead of just generating an answer, an AI agent takes multiple autonomous steps: searching across many sources, cross-referencing data, evaluating options, eliminating weak candidates and presenting a curated shortlist or taking direct action (booking, purchasing, subscribing).
The critical difference from AI search: the user never sees the options that were filtered out. Learn more about agentic search.
The practice of making your brand more likely to be selected and cited by AI systems. Unlike SEO, which focuses on ranking on a results page, GEO focuses on being chosen, trusted and used in AI-generated answers.
GEO covers: crawler access (can AI reach you?), content structure (can AI extract clean answers?), entity clarity (does AI understand what you do?), authority signals (does AI trust you?), and ongoing monitoring (is your visibility improving?).
A measurement of how likely your brand is to be cited by AI systems, typically scored 0-100. The score is based on accessibility, content structure, schema markup, entity clarity and authority signals.
Most sites score below 40, which means they are rarely or never cited. A score above 70 indicates strong AI visibility. The trend over time matters more than any single score. This is what tools like SearchScore measure.
Your presence within AI-generated answers across platforms like ChatGPT, Perplexity, Gemini and Google AI Overviews. Not rankings, impressions or traffic. Whether you are actually mentioned when AI systems answer questions relevant to your business.
This is the metric most teams are currently blind to. Traditional analytics dashboards do not track it. You can have strong Google traffic and zero AI visibility.
The phenomenon where your AI citation rate changes over time without you changing anything on your site. Caused by model retraining, citation criteria updates, competitor optimisation and shifts in query patterns.
Drift is why a single audit is not enough. You need ongoing monitoring to catch downward trends early. Learn more about AI visibility drift.
Any factor that influences whether AI selects your content for citation. Includes: crawlability, structured data, clarity of answers, external authority, consistency across sources, content depth and freshness. Most teams optimise these in isolation. AI evaluates them together as a composite.
Google's quality framework that also influences AI citation selection. AI engines assess whether content comes from a credible, experienced source. Signals include: named authors with real credentials, cited data sources, professional presentation and consistent expertise over time.
E-E-A-T matters more for Google Gemini and Google AI Overviews than for ChatGPT, because Gemini is built on Google's quality infrastructure.
How clearly and consistently your organisation is defined across the web. AI engines need to understand what your company does, what category it belongs to and what makes it distinct. This comes from: your homepage copy, Organisation schema, consistent branding across directories and clear positioning on your about page.
If your homepage opens with vague branding language like "empowering tomorrow's solutions," AI has poor entity clarity on you. If it opens with "SearchScore is an AI visibility audit tool," entity clarity is strong.
A plain-text file placed at your domain root (yourdomain.com/llms.txt) that tells AI systems which pages matter most and summarises what your business does. Think of it as robots.txt for AI understanding rather than AI crawling.
Adoption is still below 1% of websites, which means creating one gives you an immediate advantage over nearly every competitor. It takes under 5 minutes to create. Learn more in the ChatGPT SEO guide.
Machine-readable code (JSON-LD format) embedded in your pages that helps AI engines understand what your content is and how to interpret it. Common types: Organisation (defines your company), Article (marks up published content), FAQ (wraps Q&A sections), Product (marks up products with pricing and features), HowTo (marks up step-by-step guides).
Schema does not "rank" your site. It reduces ambiguity. Which increases the chance of being selected and cited accurately. Gemini uses schema more heavily than other AI platforms.
The perceived credibility of the person behind the content. AI systems increasingly look for: named authors with real credentials, consistent expertise across published work, verifiable professional backgrounds and cited qualifications.
Anonymous or generic content is less likely to be trusted by AI engines. Adding author bios with specific expertise markers to every published page is a quick win.
How widely your brand is recognised and referenced across the web. Includes: mentions on other sites, backlinks from authoritative sources, press coverage, consistent positioning and reviews on third-party platforms. AI uses this to answer: "Is this a credible source?"
How useful and specific your content actually is. AI prefers content with specific answers, real data points, original insights, cited sources and actionable detail. Generic summaries and surface-level content get skipped in favour of pages with genuine expertise and specificity.
The process where AI retrieves information from external sources (the web, databases, documents) before generating a response. This is how citations happen. AI does not rely purely on training data. It retrieves relevant content, evaluates it and uses it to construct an answer.
If your content is not retrieved during the RAG process, it cannot be cited. This is why crawl access and index presence are the most basic requirements.
The AI systems powering AI search. Examples: GPT-4, Claude, Gemini, Llama. They are trained on vast datasets and refined through retrieval and feedback. These are the platforms where citations matter most. Each LLM has different citation behaviour and preferences.
The practice of surrounding your content with signals that reinforce credibility. Includes: inline citations to data sources, author credentials near the content, supporting evidence and references to peer-reviewed or authoritative sources. It helps AI understand not just what you are saying, but why it should trust it.
Content structured to lead with a direct answer to a specific question, followed by supporting detail. AI engines extract from the top of pages first. If your key information is buried 800 words in, it may never be reached. Answer-first structure dramatically increases citation likelihood.
This means: your opening paragraph should answer the page's primary question. Supporting context comes after. The opposite of traditional copywriting that builds to the reveal.
The percentage of AI citations in your topic area that belong to your brand, compared to competitors. If you are cited in 3 out of 10 relevant queries and your main competitor is cited in 7, your share of AI voice is low even though your absolute presence looks decent.
This metric is most useful for competitive analysis. It tells you whether you are winning or losing the AI search battle relative to your actual market competitors.
Whether AI crawlers (GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot) can reach and read your content. Controlled primarily by your robots.txt file, but also affected by server response codes, security plugins and firewall rules.
This is the most common audit failure. Over 40% of websites accidentally block AI crawlers through broad wildcard rules. If GPTBot is disallowed in your robots.txt, ChatGPT cannot cite you. The fix takes 30 seconds. Check yourdomain.com/robots.txt for any lines blocking AI-specific user agents.
A file at yourdomain.com/sitemap.xml that lists all the pages on your site you want search engines and AI crawlers to discover. It helps AI engines find your content efficiently, especially pages that might not be well-linked internally.
Most CMS platforms generate sitemaps automatically. If you do not have one, create it. It is a basic hygiene factor: without a sitemap, AI crawlers rely entirely on following links, which means they may miss isolated or deeply nested pages.
A protocol that lets you instantly notify search engines and AI platforms when your content changes, rather than waiting for them to recrawl on their own schedule. Supported by Bing, Yandex and other indexes that AI engines pull from.
Without IndexNow, content changes can take days or weeks to surface in AI answers. With it, updates propagate near-instantly to participating indexes. Implemented by adding a simple API key file and sending POST requests when pages change.
The depth and breadth of your coverage on a specific subject area. Distinct from brand authority (which measures recognition). Topical authority measures whether you have comprehensive, interconnected content that covers a topic thoroughly.
AI engines reward topical authority because it signals genuine expertise. A site with 20 well-structured pages covering every angle of a topic is more likely to be cited than a site with 2 shallow pages on the same topic, even if the smaller site has stronger backlinks.
A cluster of interconnected pages organised around a central topic. The hub is a comprehensive pillar page that overviews the topic, with sub-articles that dive deep into specific subtopics. Each sub-article links back to the hub and to related sub-articles.
Content hubs signal topical authority to AI engines. They show you cover a subject completely rather than having isolated, disconnected pages. SearchScore checks for content hub structure as part of its topical authority audit.
A section of your content that is structured specifically to be extractable by AI engines. A citability passage is a concise, factual statement that directly answers a question or defines a concept, written clearly enough that an AI engine can extract and paraphrase it accurately.
Not all content is equally citable. Long, meandering paragraphs without clear structure get skipped. Short, direct statements with specific data or clear definitions get cited. Deliberately crafting citability passages into your key pages increases citation likelihood.
A schema.org type that tells AI assistants and voice platforms which sections of your page are suitable for being read aloud. It uses CSS selectors to identify the most relevant content sections, typically headings and key paragraphs.
While originally designed for voice assistants, SpeakableSpecification is increasingly used by AI search engines to identify the most important content on a page. Adding it signals to AI platforms which parts of your content are the definitive answers.
Meta tags (og:title, og:description, og:image, og:url) that define how your page appears when shared on social platforms and when parsed by AI systems. Originally created by Facebook, now widely used as a content identification standard.
AI engines read Open Graph tags as part of understanding what a page is about. Missing or generic OG tags mean AI has less structured information to work with. Every key page should have complete, accurate Open Graph metadata.
Having your business name, address and phone number identical across every platform where you appear: your website, Google Business Profile, directories, social profiles and review sites.
Inconsistent NAP data undermines entity clarity. If your website says "Acme Corp" but directories list "Acme Corporation" or "Acme Corp Ltd," AI engines cannot confidently associate all references with the same entity. This reduces citation likelihood and accuracy.
SEO answers: "Can you rank?"
GEO answers: "Will AI choose you?"
Those are not the same question. A site can rank position 1 on Google and be invisible to ChatGPT. A small niche site with clean structure can be ChatGPT's go-to source while ranking nowhere on Google. The disciplines overlap but are distinct.
The brands that understand this distinction early are the ones getting cited now, while their competitors are still optimising for a world where AI search does not exist.
Most teams do not know if they are visible in AI because there is no standard analytics dashboard for it. Platforms like SearchScore measure whether you are mentioned, which queries trigger citations, which competitors appear instead of you, and what is specifically holding you back.
Run a free audit to get your AI visibility score and see where you stand across ChatGPT, Perplexity, Gemini and Google AI Overviews.
Related: What is GEO? | AI Visibility Audit | ChatGPT SEO Guide