AI visibility is whether AI assistants — ChatGPT, Google Gemini, Perplexity, and the rest — find your business, describe it accurately, and cite sources you control when customers ask the questions you could answer. It’s the AI-era counterpart to search rankings, with one structural difference: a search engine gives your customer a list to choose from, and an assistant gives them an answer. You’re no longer competing for a position on page one. You’re competing to be in the answer at all.
This page is our working definition of the term — what AI visibility is, how it differs from rankings, and how it’s honestly measured.
An answer is not a list
For twenty years, being findable meant ranking: your business appeared somewhere in a list of results, and a customer scrolling that list might pick you. Position mattered, but the list was long, and lots of businesses got some share of the attention.
An assistant collapses the list. Asked “who does ceramic tint near me,” it names a business or two, says something about each, and stops. There is no page two. The businesses it names get the call; everyone else was never seen. That concentration is what makes this layer worth measuring on its own — a business can rank respectably in classic search and still never appear in assistant answers, because the assistant is not reading the ranked list the way a human does. It’s reading sources, weighing them, and composing a short answer.
And when the answer is wrong about you — old hours, a dead phone number, a service you don’t offer — nobody sends you a notice. The customer just acts on it.
Where the answers come from
AI has no memory of your business. An assistant answers from what it retrieves in the moment: your Google Business Profile, review platforms, directories, local press, your website — fetched and read in the seconds before it responds. (The mechanics are worth understanding, and we walk through them in Why doesn’t ChatGPT know my business?)
That’s why AI visibility is a property of your sources, not a property of the AI. You can’t edit an assistant. You can edit what it reads.
How AI visibility is measured
A definition you can’t measure is a slogan. We measure AI visibility as three separate numbers — deliberately separate, because they move at different speeds and mean different things:
Discovery Rate — when a customer asks without naming you (“who does this near me”), how often do you appear in the answer? This is the number that feels like visibility, and it’s the one everything else ultimately serves. It’s also the slowest to move.
Answer Accuracy — when the assistant does talk about you, is each fact right? Hours, phone, services, prices — every claim in the answer lands in one of three buckets: correct, wrong, or missing. A business can be perfectly visible and still bleed customers to a wrong phone number.
Citation Share — of the sources assistants cite when answering your kind of question, how many do you control, how many belong to competitors, and how many are neutral directories? This is the early signal. Sources change before answers do, so Citation Share usually moves before Discovery Rate — it tells you whether the ground is shifting under the answers.
One honest complication: assistants vary their answers. Ask the same question twice and you may get different responses. So a single test proves nothing — real measurement means repeating each question and reporting rates with confidence intervals, so a change you report is a real change and not noise that landed in your favor. It also means nobody can guarantee you a mention, and any honest definition of AI visibility has to say so.
Why this layer is the one to watch
Customers who used to scroll search results are starting to ask assistants instead — and take the recommendation. The habit doesn’t need to be universal to matter to a local business: every customer who asks an assistant and hears a competitor’s name is a call that never came, and there’s no analytics dashboard where you’d see it happen. Search visibility failed loudly — you could watch your traffic drop. AI visibility fails silently.
The encouraging part is that the inputs are ordinary. The sources assistants read are the same profiles, listings, reviews, and pages that local businesses have always maintained — now with a reader that is faster, stricter about consistency, and unable to shrug off contradictions the way a human can.
AI visibility, SEO, AEO, GEO
You’ll see three acronyms in this space: SEO (ranking in search results), AEO (structuring content so answer engines can quote it), and GEO (improving how generative models retrieve and recommend you). They’re three layers of the same job. We use AI visibility as the plain-English umbrella for the AI layers, so you don’t need to keep the acronyms straight to know what’s being done for you.
Where to start
Start by finding out where you stand — you can check what AI says about your business yourself, free, in about twenty minutes. If you’d rather have it measured properly — repeated runs, accuracy checked claim by claim, every finding cited to a dated transcript — that’s what we do, and the first look is free.