How we measure AI visibility

What the score means

Your AI Visibility Score is a single number from 0–100. It is how often AI assistants actually recognise and recommend your business when customers ask: sixteen questions, three engines, one pass, on the day we scan. It is not a vanity metric — a high score means AI reliably puts you in the answer; a low score means you're effectively invisible at the moment of choice.

Here is the arithmetic, because it is not complicated. Seventy of the hundred points come from the discovery and comparison questions — where a customer is still choosing, and we count whether the assistant puts you in the answer at all. The other thirty come from the direct and booking questions — where a customer already has your name, and we count whether the assistant actually knows who you are. We do that separately for each engine, and your score is the average of the three.

What we ask the AI

We ask the assistants the questions a real customer would ask — not keywords. Four kinds:

• Discovery — “best [your service] in [your city]”, “who's good for [need]”, “recommend somewhere near me”.
• Comparison — “compare the top [your trade] in [your city] for [need]”, “is [you] any good, or is there somewhere better?”.
• Direct — asking about your business by name: are you any good, what do you offer, are your details right.
• Booking — “can I book with [you]”, “are they taking new customers”.

We run these across the assistants people actually use — ChatGPT, Google Gemini and Perplexity — in one pass: each question asked once of each engine, on the day of the scan. We don't repeat the run and average it, so we won't pretend the number is steadier than it is — ask the same engine the same question tomorrow and it can answer differently. That is why a subscription re-measures every month. The movement between scans is the finding, not an error in it.

What the score measures

The answers the assistants gave us, and nothing else. Your Google profile, your website and your reviews are not added into the number — we read them, we show you what they say, and we tell you where they disagree with the assistants, but they are not terms in the sum. Neither is anything you buy from us. A subscription changes how often we measure you. It cannot change the measurement.

What we think moves it

Assistants build their answers out of what they can find and trust about you across the open web. So we think: change what they find, and you change what they say. That is our reading, not a measurement — the score moves only when their answers do, and an answer can also move on its own, as the paragraph above says.

• Your Google Business Profile and your own website — the two sources our scan sets beside what the assistants say. Other listings may count too; we do not check them yet.
• Reviews — Rating, volume and freshness. How much each one counts, we have not measured.
• A website AI crawlers can read — server-rendered, machine-readable content (many site builders hide everything behind JavaScript that AI never runs).
• Accurate, consistent facts — the hours, address, services and prices AI repeats about you. Your report shows each one as the assistants state it, engine by engine, so you can see which are wrong or missing.

What we publish

We turn these measurements into public, citeable rankings — the AI Visibility Index, city by city — and every row carries the date it was scanned. The data behind the rankings is open for anyone — including AI assistants — to read and cite.