Your customers ask AI. We measure whether you appear in the answer, show what works, and prove whether it pays off.
ChatGPT, Gemini and Claude answer the questions people used to google you for. Geony measures whether your brand appears in those answers, which competitor does appear, and which sources decide that, so you know what to do to get in there too.
Free check across ChatGPT, Gemini and Claude · no account needed
Built by marketers and brand builders who do the work themselves. That is why the outcome is a worklist you pick up on Monday, not a data dump.
Measure, improve, prove
Not a loose dashboard but a loop: you see where you stand, do the work that pays off, and measure whether it returned.
Measure
Every week across ChatGPT, Gemini and Claude, every figure with a confidence interval. So you know whether a shift is real or noise.
Improve
A worklist by impact: which source pages you need to appear on and which questions still belong to no one. Focused work, no tricks.
Prove
After every action Geony keeps measuring, so you see in black and white whether it worked. If nothing moves, you stop doing work that does not pay off.
Invisible in AI is the new page two
Three things that have changed, and why your current SEO report does not show them.
One answer, not ten links
An AI assistant gives no list of results but a single recommendation. If you are not in it, you do not exist for that question. There is no position eleven to fall back on.
Every assistant its own truth
From our measurements: 316 of the 398 cited sources appear with exactly one assistant. What ChatGPT reads, Gemini does not. Measuring one assistant means missing three quarters of the playing field.
Your Analytics does not see it
Whoever gets an AI answer often clicks nowhere. The question was asked, the advice given, and your statistics stayed empty. So measuring has to happen at the source: in the answers themselves.
This is what we found when we really started measuring
One Dutch market, twenty questions, three assistants, every question asked three times. These are our own figures, not a brochure.
The full methodology, including why every figure carries a confidence interval, is laid out in the open on the methodology page.
From question to worklist in three steps
Your questions, asked every week
You choose the questions your customers ask ('which agency should I pick for…'). Geony puts them to ChatGPT, Gemini and Claude every week, each multiple times, because AI answers differ every time.
Every answer dissected
Which brands are mentioned, in which spot, and which sources are cited. Who is not there is recorded too, because that absence is usually the finding.
A worklist, not a data dump
The sources that decide the answers and where you are missing, sorted by impact, with a status per action. On Monday you know what to do and why.
For seven industries the starter set of customer questions is ready and waiting, from marketing agency to healthcare practice. See your industry
The measurement plays
From customer question to proof: the open dot is the brand that is not mentioned, and it only fills once the measurement proves it. Played with fictional brands; the real screens are below.
This is how it really looks
This screen comes straight from the app, with an actual measurement. What you see here is what you get.
We also report what we are not sure of
Tools that measure once report chance as if it were a trend. Geony asks every question multiple times and computes with the same statistics clinical studies use. Drag the number of measurements up beside this and watch the confidence interval narrow: this is not an animation but the real formula.
Read the full methodAt 3 measurements the true visibility could lie almost anywhere. Anyone who sees a trend here is reporting chance.
The measurement points, you improve
The measurement tells you exactly where effort pays off. Then you do the work, and the measurement proves whether it returns.
Do it yourself
The worklist says which source pages you need to appear on and which questions still belong to no one. Every action gets a status, so you always know what is still open.
How improving works →See whether it works
After every action Geony keeps measuring weekly, with a confidence interval on every figure. If your visibility moves demonstrably, you see it. If nothing moves, you stop doing work that does not pay off.
How Geony keeps measuring →Figures you dare to show
Results, not applause: quotes are about what the measurement changed. (Examples; we'll fill them with real customer quotes once we have them.)
Example“Voor het eerst kan ik een klant laten zien wat AI over zijn merk zegt, mét een interval eronder. Het rapport is niet langer een discussie, het is het begin van de opdracht.”
Example“We stopten met gokken. De werklijst zei precies op welke bronpagina's we moesten staan, en de wekelijkse meting bewees week na week dat het bewoog.”
Example“Dat Geony ook zegt wat het níét zeker weet, is precies waarom een kritische klant het gelooft. Cijfers die je durft te laten zien, dat verkoopt zichzelf.”
First just this
What exactly does Geony measure?+
Whether your brand (and your competitors) are mentioned in the answers ChatGPT, Gemini and Claude give to the questions your customers ask. Plus which sources feed those answers, because that is what you steer on.
Why multiple measurements per question?+
The same AI gives different answers to the same question. Measure once and you report chance. Geony asks every question multiple times and shows a 95% confidence interval on each percentage.
Is measuring one AI assistant not enough?+
No, and that is measurable. In our own data the same brand appeared in 36% of Claude answers and 1% of Gemini answers. Steer on one assistant and you know nothing about the others.
What is GEO (generative engine optimization)?+
The field that optimizes visibility in AI answers, the way SEO does for search engines. It starts with measuring: you cannot optimize what you cannot see.
Does Geony also help improve visibility?+
Yes, along four levers: appear on the sources that feed the answers (Geony tells you exactly which), claim the questions where no one is mentioned yet, remove technical blockers, and prove with the weekly measurement whether it worked. No time to carry it out? The team behind Geony can take that work over.
Who is Geony for?+
For brands that want to know where they stand, and for agencies that offer AI visibility as a service: white-label with their own logo, shareable reports under their own name, and if wanted on their own domain such as report.youragency.com.
Within a day you'll know whether AI mentions or ignores your brand
The first measurement runs on your own prompts, across three assistants, with real figures. No demo data, no sales call first.
Prefer to talk it over first? Book a call