Your customer increasingly sees an AI answer first. You do not know which questions they ask there, whether your name is in the answer, or what to do to get in it.
Your customer does not even have to open ChatGPT. For more and more searches, Google puts an AI answer above the ordinary results. Where that answer appears, your customer reads the advice before the list.
That advice has room for a few names. If yours is among them, your customer sees you. If it is not, that answer does not bring you to their attention, not even when you rank on page one of the ordinary results.
Search engines are not going away. The point is what your customer sees first, and that is increasingly a composed piece of advice rather than a list.
You can of course ask it yourself. But does Gemini name you too, on a concrete decision question, in your service area, and again next week? One answer is an observation, not a pattern.
And without a pattern you do not know what to tackle first. Write another blog post, sharpen your services page, make your service area and pricing clearer? You already put time and money into being found. Without knowing where you drop out of view, you do that on gut feeling.
This can be measured. Not with one question, but with the same questions, repeated, to several assistants.
Which installer in Utrecht can fit a heat pump in a 1930s house?
The answer names:
Van der Meer Installatietechniek, Warmtewerk, and a national comparison platform.
Not mentionedYourcompany.com does not appear in the answer.
ActionFirst move: add your service area, the property types you handle, your certification and an indication of pricing to your existing heat pump page. This is a question close to the decision and that information is not there yet, so a new page is not needed. At the next measurement you see whether the answer has changed.
Illustrative example. Your measurement runs on your own market, customers and competitors.
Want to see this for your own business?
Your first measurement consists of six buying questions on Gemini and is free of charge.
We put together a question set that runs from broad orientation to the concrete choice, supplemented with what people in your market ask Google. You add your own questions and freeze the set, so every following measurement is comparable.
No estimated visibility, but the answer exactly as the assistant gave it. Per question you see whether you are named, in which position, how you are described, which competitors do appear and which sources are shown. The score summarises that; the answer underneath stays visible.
Each measurement produces a prioritised list of what to tackle first, tied to the questions on which you were not named. You tick off what you pick up, and at the next measurement you see whether the answers have changed.
We separate three things that often get lumped together: being named, being described positively, and being explicitly recommended. Per question you see which of them occur in the answer.
AI answers change. A measurement is therefore a snapshot taken by a fixed method, not a ranking and not a guarantee. So we measure repeatedly: the change says more than any single result.
Behind every result you see the question asked, the full answer, the assistant, the measurement date and the sources shown. So you do not have to take a conclusion on our word; you can check it.
Where you start: your first measurement runs on Gemini with six buying questions. On a paid plan the same questions also go to ChatGPT, Perplexity and Google AI Overviews, and you measure continuously instead of once.
SEO and your business profile help you be findable and understood, and that work counts here too. But they do not show what an assistant does with it when it composes its own advice out of all those sources: are you named, how are you described, and who appears beside or instead of you? Asking ChatGPT one question yourself does not show that either: that is one answer to one phrasing, today.
We combine what people ask Google (People Also Ask) with questions from your buyer's perspective and AI question discovery. Every question gets a stage in the customer journey, from orientation to choice, and that stage weighs into your score. Your own questions simply go in alongside.
Before we advise anything, we check your site. If you already have a page on the topic, the advice is to strengthen it rather than create something new.
Selling in the Netherlands? Then the questions go in Dutch to assistants answering for the Dutch market. Selling across the border? Then we measure in that market. The same question yields different names and sources per language and country.
You work independently in the tool: you compose your question set, start your measurements and follow your action list. Whether you sell to consumers or to other businesses, locally or nationally: the measurement is relevant as soon as your customers compare providers online.
You sign up with an email address on the domain you want to measure, so we know you belong to it. And we check whether there is enough publicly available information about your business: if an assistant has nothing to cite, a measurement says little, and we would rather say so right away.
GEO-Crafter gets better through the people who use it. What you measure and report back sharpens the question sets and the advice for real markets. If something looks off in your measurement, we would like to hear it.
Enter your domain and an email address on that domain. If they match, you receive an invitation by email. If not, you hear from us within one working day.
Your first measurement consists of six buying questions on Gemini and is free of charge.
Measuring for a client, for instance as an agency? Enter your client's domain and the business email address of your own agency. You hear from us within one working day. See our privacy statement.