When your customer asks ChatGPT, Google’s AI or Perplexity who to hire, it names two or three businesses. We find out whether you’re one of them — and if you’re not, exactly why.
We put six to eight real buyer questions to the actual AI engines — phrased the way your customer would phrase them, not the way a marketer would. Then we record, word for word, which businesses get named.
Illustration of the format. Your report carries your real queries and the untruncated responses.
Every check runs against your live site, listings and reviews. Nothing is guessed, and anything we can’t verify is labelled as such.
If these fail, nothing else matters
Being readable isn’t the same as being recommended
The five checks that most decide whether AI names you count triple. High-impact areas count double. The rest count once. It’s a weighted score, not a flat average, because the things aren’t equally important.
Anything we cannot verify with a tool is marked Inferred and capped. Your score never looks better than the evidence behind it, and the report tells you how many checks fell into that bucket. We’d rather hand you an uncomfortable number you can trust than a flattering one you can’t.
Four fields are required. The optional ones make the audit noticeably sharper — especially the questions your buyers actually ask.
We’ll run the audit and come back within 24 to 48 hours. Keep a copy of what you sent, below.
No automated email. No generic PDF with your name pasted into it. A person reads your business.
The automated checks run against your live site, then we put your buyers’ questions to the real AI engines and record what comes back.
Scores get sanity-checked against what we know about your market. Anything a tool couldn’t verify is marked honestly rather than guessed at.
You get the score, the findings and the transcripts — yours to keep, and yours to hand to any agency you like. If we’re a fit we’ll say what we’d do. If we’re not, we’ll say that too.
Free, and no call required. You get the score, the findings and the live answer-test transcripts whether or not you hire us, and you’re welcome to hand the whole thing to another agency. The catch, if there is one, is that most people who read their own transcript want it fixed.
Automated checkers are useful and several are genuinely good. They tell you how often AI mentions your brand — the outcome. This diagnoses the cause: thirty weighted checks against your actual site and listings, a live answer test, a competitor share-of-answer tally, and a person who reads it before it reaches you.
We use it to run the audit and reply to you. We don’t sell it, we don’t add you to a drip sequence, and we don’t pass it to anyone else. Full detail is in our Privacy Policy.
No. And if the audit shows your foundations are sound, we’ll tell you that rather than invent work. Retainers are month to month if you do want help — no contract, no lock-in.
It means we judged that check from indirect signal rather than confirming it with a tool. Those get capped, and the summary tells you how many there were. It keeps the score honest — your number should never look better than the evidence behind it.
Twenty-four to forty-eight hours either way. Larger sites take longer to crawl but the turnaround is the same — we cap the crawl and tell you if we had to.