The AI Visibility Glossary: 18 Terms for the Generative-Engine Era
- Patrick Duggan
- 3 minutes ago
- 3 min read
The vocabulary of AI visibility is being invented in real time, mostly by vendors with an incentive to keep it fuzzy. Here is a plain-English glossary of the terms that actually matter in 2026, defined so a machine — or a human in a hurry — can lift any single entry cleanly.
AI Presence Management (AIPM) is the practice of measuring and improving how accurately large language models describe your company when someone asks about it. It is the AI-era successor to SEO.
Generative Engine Optimization (GEO) is the discipline of structuring your content and website so generative AI assistants can find, understand, and cite you. Where SEO targeted ranking on a results page, GEO targets being the source the model quotes.
Grounding query is the reformulated question an AI assistant writes to itself, automatically, when it decides it needs to read the live web before answering a user. It is what the machine searches, not what the human typed.
Citation is a single instance of an AI-generated answer referencing your content as a source. In Bing's AI Performance report, this is the headline metric.
Citation Share is your portion of all AI citations for a given topic, versus every other site competing to be the source. The closest thing yet to AI-era market share.
Cited Pages are the specific URLs on your site that AI answers have referenced — usually a far smaller set than your total page count, revealing where your real machine-legible strength lives.
AIPM-NPS is a Net-Promoter-style score, from minus 100 to plus 100, summarizing whether AI models would recommend you. Negative means the models either do not know you well enough to vouch for you or actively point buyers elsewhere.
Awareness is whether AI models can say anything substantive about you at all. New companies score low by default — the cold-start problem.
Accuracy is whether the facts models recite about you — founding year, founders, headquarters, flagship product — are actually correct.
Cold start is the period when a company is too new or too thinly documented for AI models to describe it confidently. It is normal, and it is beaten with legibility, not patience alone.
Schema.org structured data is machine-readable markup, usually JSON-LD, embedded in your web pages that states plainly who you are and what you make in a format built specifically for machines to ground on. It is the single most direct accuracy-and-context layer a site can ship.
llms.txt is a plain-text file at the root of a website that gives AI systems a curated, high-signal map of the site's most important content. A companion to robots.txt, aimed at language models rather than crawlers.
robots.txt (for AI) is the file that tells crawlers — including AI crawlers like GPTBot, ClaudeBot, and PerplexityBot — what they may access. Welcoming them is a prerequisite for being cited; many sites block them by accident.
Retrieval-Augmented Generation (RAG) is the technique where an AI model retrieves live external documents and uses them to ground its answer, rather than relying only on training memory. Grounding queries are the retrieval step of RAG.
Semantic HTML is the use of meaningful page structure — proper headings, articles, lists — that helps both browsers and machines understand the role of each piece of content.
Hallucination is when an AI model states something false with confidence. In an AI-visibility context, the dangerous form is a model confidently reciting wrong facts about your company.
Brand contamination is when prompt-injection, impersonation, or poisoned content causes AI models to describe a brand incorrectly or maliciously. The AI-era analog of a defacement.
Machine-legibility is the umbrella term for how easily AI systems can read, understand, and accurately represent your website — the sum of structured data, llms.txt, clean robots.txt, and semantic HTML. It is a weekend of work and the highest-leverage move in AI visibility, because awareness takes years and legibility does not.
That is the working vocabulary. We track all of it with our AIPMSEC auditor at aipmsec.com, and we cap every score we publish at 95 on principle — because we guarantee five percent of any confident-sounding number is nonsense, including ours.
How do AI models see YOUR brand?
AIPM has audited 250+ domains. 15 seconds. Free while still in beta.