Glossary

The AI visibility glossary

The terms as this site uses them, one paragraph each, in the words the report uses.

Updated September 2026 · by Monish

AI visibility

How often AI answer engines (ChatGPT, Perplexity, Gemini, Google AI Overviews) name a company when a buyer asks who to consider in its category, and whether they cite its own pages when they do. Measured as a share of recorded answers, weighted by position in the list and by citation. How it is measured and moved.

Answer engine

A system that answers a question with a written answer rather than a list of links: ChatGPT with web search, Perplexity, Google's AI Overviews and AI Mode, Gemini with search grounding. How it is measured and moved.

AEO (answer engine optimisation)

The work of being named and cited in answer engines' answers. Interchangeable in practice with GEO. How it is measured and moved.

GEO (generative engine optimisation)

The same work under the name that emphasises the generated answer rather than the engine. See AEO. How it is measured and moved.

Buyer question

A question a person asks an engine while choosing a supplier: 'best X for Y', 'A vs B', 'A alternatives', 'X for a 40-person team'. The report derives about fifteen from a company's own website. How it is measured and moved.

Share of voice

The percentage of recorded answers that name a given company. One answer can name several, so the shares in a category do not add to 100. How it is measured and moved.

Citation

A page an engine shows or reports as the source of what it said. Perplexity and AI Overviews show them; ChatGPT with web search reports them in its response. Being cited is separate from being named. How it is measured and moved.

Named instead

In a buyer question where the company is absent, the company the engine named in its place. The report's 'where competitors win' table is a list of these. How it is measured and moved.

Named ahead

In an answer that names both, the rival placed above you. A loss the report counts alongside absences. How it is measured and moved.

Site readiness

The report's second score: thirteen technical checks on whether engines can read and quote a site (crawler access, sitemap, llms.txt, structured data and others), weighted. How it is measured and moved.

llms.txt

A Markdown file at a site's root that tells language models what the site is and which pages are worth reading. An index; llms-full.txt is the optional single file with the pages' text. How it is measured and moved.

AI crawler

The fetcher an AI company uses to read the web: GPTBot and OAI-SearchBot (OpenAI), PerplexityBot, ClaudeBot (Anthropic), Google-Extended (Google's AI training opt-out control), CCBot (Common Crawl). Named in robots.txt to allow or block. How it is measured and moved.

JSON-LD

The structured-data format engines read from a page: a script block declaring the Organization, its Products or Services, an Article's author and date. Where an engine learns what a company is without reading the prose. How it is measured and moved.

Grounded answer

An answer produced after a live web search, as opposed to one written from the model's training data. The report scores both but says which is which. How it is measured and moved.

Sharpest threat

In the report, the rival named ahead of you most often when you both appear: the company buyers are actually weighing you against, as opposed to the one with the largest share. How it is measured and moved.

Posture

Whether a company leads its category's answers by a clear margin (leader) or does not (challenger). The report's opening reads differently for each. How it is measured and moved.

See the terms measured on your own site

Your website, region and work email. The score appears in about three minutes, and the report is a link you keep.

Check my AI visibility