What is AI saying about your brand?
More and more customers ask an AI before they decide. You can see your Google ranking — but nobody tells you where you stand inside AI answers: recommended, skipped, or misquoted. Luxcerta measures it.
This is no fringe behavior: a June 2026 Pew Research Center survey found about half of U.S. adults now use AI chatbots, with 24% using them daily — and ChatGPT usage has more than doubled since 2023. Back in 2024, Gartner predicted traditional search engine volume would drop 25% by 2026 as users shift to AI assistants — a forecast, not a measurement, but one pointing the same way as Pew's measured data.
Sources: Pew Research Center, survey report, June 17, 2026 · Gartner press release, Feb 19, 2024 (forecast)
What we do
We take questions real users actually ask — mined from public forum posts, not invented keywords — and put them to the AI platforms repeatedly, each time in a fresh conversation. Then we count: who gets named, in what order, which websites the AI cites, and whether it gets your facts wrong.
The counting involves no AI judgment at all — fixed name matching, a fixed method. Ask the same questions yourself and you will see the same thing.
What a real survey looks like
In August 2026 we ran our first category survey — dental implant clinics in greater Taipei: 15 real questions × 3 fresh conversations each, 45 AI answers in total. Two findings:
Visibility is extremely concentrated — the long tail can vanish overnight
The 45 answers named 55 clinics and hospitals, but the vast majority appeared once or never. Only 6 appeared consistently. In other words: if the AI mentioned your clinic today, it was probably luck — a different phrasing tomorrow and you're gone, and you'd never know.
The AI reads your website — then recommends someone else
| Clinic website (anonymised) | Answers citing it | Of those, naming the clinic |
|---|---|---|
| Clinic A's site | 17 | 1 |
| Clinic B's site | 15 | 0 |
| Clinic C's site | 15 | 0 |
| Clinic D's site | 12 | 0 |
| Clinic E's site | 9 | 0 |
In the same answer, the AI cited one clinic's website — and recommended another. The content gets consumed as category reference material, and the credit never returns to whoever wrote it.
The public version anonymises these five clinics — the point of the survey is to show the phenomenon, not to embarrass anyone. Wondering if you're one of them? Email us; we only tell the clinic itself.
What a diagnosis gives you
A named measurement of your own brand, delivered in the email body and on a web page — never as an attachment:
- Your appearance count and relative positionacross the same question set: how often you and your peers each get named
- What the AI cites when it mentions youyour own site, forum threads, or content someone else wrote about you
- Whether the AI gets your facts wrongprices, services, credentials — with the exact sentence and which answer it appeared in
- Full method and limitationswhich platforms, how many repetitions, what we could not measure — stated plainly
Our principle
What we didn't measure, we don't claim. Insufficient samples are labelled insufficient. Unmeasurable is written as unmeasurable — never dressed up as a precise-looking zero. Every number you get can be verified by you, in about two minutes.
Want to know what AI says about you?
Email us your brand name and website. We usually reply within one to two business days — every diagnosis is currently run and reviewed by hand.
Request a diagnosis