What’s your AI Visibility Score?
One number, 0 to 100, for how often ChatGPT, Gemini and Perplexity recommend your business when customers ask who to hire. We ask them five questions your customers ask and send you the score, every answer, who they named instead, and why.
Your score, and the evidence behind it
- Your AI Visibility Score. A single number from 0 to 100, worked out the same way every time so you can compare it month to month.
- The questions we tested. Five real questions built for your business — the way your customers actually ask them.
- What each engine said. ChatGPT, Perplexity and Google Gemini, with live web search enabled, and the answers reproduced verbatim.
- Whether you were named. And if so, in what position and with which page cited.
- Who was named instead. Usually the most uncomfortable and most useful part of the report.
- How your website and Google listing look. The obvious problems: pages Google cannot see, a thin Google profile, a slow site.
- Why. The main reasons behind the result, in plain language.
- What we would do first. Ranked, and honest about which items you could handle yourself.
Because describing this work does not land and demonstrating it does. Most owners do not know until they see it. We would rather show you than talk at you — and if you take the report and fix things without us, we are still glad you know.
Request your free score
Six fields. Your score in about two business days.
Fifteen answers, one number
We write five questions the way your customers ask them, never using your business name, and ask each one of ChatGPT, Google Gemini and Perplexity with live web search switched on. That gives fifteen answers. Each answer earns points for where you show up in it, and your score is the average.
| Where you show up in the answer | Points |
|---|---|
| Named first | 100 |
| Named in the top three | 70 |
| Named, but lower down | 40 |
| The assistant looked you up, then left you out | 10 |
| Not mentioned at all | 0 |
The “looked you up, then left you out” row matters more than its ten points suggest. Before answering, ChatGPT and Gemini run their own searches, and we can see them. When one searched for your name and still picked someone else, it already knows you exist — your website just did not give it a clear answer to repeat. That is usually the quickest thing to fix.
Once your five questions are set, they do not change. Re-tests use the same questions, the same assistants and the same scoring, so when the number moves, it is because your visibility moved — not because we asked something easier. AI answers do vary a little from day to day, so a few points either way is normal. We tell you when a change is real.
What the full report looks like
We publish a complete example so you can judge the format before you send anything. It uses a fictional industrial air filtration company and illustrative data — no real client is represented — but the structure is exactly what you receive.
| Section | What it contains |
|---|---|
| What we asked | The buyer questions tested, written for your category. None of them mention your company, because the point is to see whether you surface unprompted. |
| What the engines said | A question-by-question grid for ChatGPT, Perplexity and Gemini: named or not, and which companies were named instead. Model versions and the test date are recorded at the top. |
| What the engines searched for | The searches ChatGPT and Gemini ran before answering (Perplexity does not show its searches), which sorts every run into one of three states: named, searched for by name and then dropped, or never considered. |
| Why | The specific causes behind the result, each tied to something an engine did or cited — not a generic checklist. |
| What we would do first | A short ranked list of fixes, with the effort each one takes. |
| Full transcripts | Every response, unedited, so you can check our reading against what the engines actually returned. |
See a sample score See a full example report ›
Opens the example on our reports site. Your own report is private to you and is never published.
Score Watch
Your score is a snapshot. Score Watch re-runs it every month with the same five questions and emails you the number, who was named, and what changed.
- Your AI Visibility Score every month, scored exactly the same way
- Which competitors are gaining on you, and which sites the assistants are quoting
- A short plain-English email, not a dashboard you have to log into
- Month to month. No site work included — when you want the problems fixed, that is the SEO + AI Fix
Every care plan already includes it: Local Care re-scores quarterly, Website Care and SEO Growth monthly. All prices ›
Before you ask
Is the score really free?
Yes. There is no fee, no card, and no obligation. We run your score, send you the report, and that is the end of it unless you want to talk. It costs us a couple of hours and it is the most honest way we know to show what we do.
What exactly do I get?
A short written report showing the buyer questions we tested, what ChatGPT, Perplexity and Google Gemini each answered, whether your company was named, which companies were named instead, and the main reasons behind the result. You get the actual transcripts, not a summary.
How long does it take?
Usually two business days. If we are backed up we will tell you when you submit rather than leave you waiting.
Do I have to get on a sales call to receive it?
No. We email you the report. If you want to discuss it we are glad to, and if you would rather take the findings and act on them yourself, that is a perfectly good outcome.
What is a good AI Visibility Score?
Most businesses score low, because the assistants name only a handful of businesses per answer and most firms are not among them. A score of 30 to 50 means you are being named regularly. Above 60 means you are usually one of the first names given. The number that matters most is your own score next month compared with this one.
Can you guarantee a higher score?
No, and anyone who does is guessing. Nobody controls what ChatGPT or Google says. What we can do is remove the reasons you are being left out and measure the result honestly, including when it does not move.
What if the results are fine?
Then we will tell you that, and you will have a documented baseline you can re-test later. We would rather give you an honest all-clear than manufacture a problem.
Will you need access to my website?
No — everything we test is publicly visible. Access is only needed if you later engage us to make changes.