AI visibility for knowledge-intensive SMEs
Your next client asks ChatGPT.Three names come back.Not Google alone.
Is your name in the answer?
In ChatGPT, Gemini and Perplexity.
Try it now, free, in twenty seconds
Pim van den BoomenBram Kohlensystems architect with a finance backgroundconversion specialist from the coaching world
Budel, Noord-BrabantAbout us
Ik zoek een goede verkoopmakelaar in Eindhoven, wie raad je aan?
Als je je woning in Eindhoven wilt verkopen, zou ik niet blind voor de grootste makelaar kiezen. Op basis van actuele Funda-beoordelingen, aantallen verkooppreviews en lokale reputatie zou ik deze shortlist maken:
- LISTED. Makelaarskantoor Eindhoven — mijn eerste keuze om mee te praten. Sterke lokale reputatie en momenteel 4,9/5 uit 42 Google-reviews.Listed +1
- Debbie Mels Lommers Makelaars — interessant als je een makelaar zoekt met een zeer sterke reviewscore. Een actuele ranglijst vermeldt 9,7/10 over 99 reviews.Ik Zoek De Beste Makelaar
- Van Uffelen — meer gevestigde partij. Op Funda staat momenteel 9,2/10 op basis van 118 beoordelingen in de afgelopen 24 maanden.Funda
Real ChatGPT answer (in Dutch) · 8 Sept 2026 · Ask it yourself
Why answers differDoes your business appear in ChatGPT?
Three AI models are asked who the best in your trade in your city is, and this screen shows you whether your name comes up.
That was one question.
Your clients ask dozens of them, each worded slightly differently. The full measurement asks all of them across four AI platforms and in Google, and shows who does get named and why. Nobody calls you: you get the report.
You come back in the answers.
On this one question, today. What we have not measured is on how many other questions, in which position, and whether it still holds next month. That is worth half an hour, with this result on the table.
See the full measurementSchedule a conversationSchedule a conversation
Your trade
Where the week leaks away
Three trades, the same complaint three times over in different words. Pick yours and see whether it fits.
Pick your trade. Only this block changes.
Estate agent
You have Realworks. And then you type it in a second time.
Now
With DDD
The enquiry arrives by WhatsApp or voicemail. In the evening you retype it into Relations, and then chase the diary.
90 to 150 minutes a week
We read that inbox, put the contact into Realworks through the API and send the confirmation. Realworks stays in charge.
Following up after a viewing: a message when you remember, and otherwise nothing.
60 to 90 minutes a week
A fixed follow-up a day later, and the viewer who is house-hunting themselves moves on to buying.
The evening update to the seller, typed from memory in the car or on the sofa.
40 to 60 minutes a week
You dictate it. A message goes out in the same shape as last week.
The word you use yourselfretyping
Your office takes out the Realworks API itself. Without it we build only the reading half.
Coach
The gap in your diary is not free time.
Now
With DDD
Preparing the intake and chasing the form. If it is not filled in, you start the session from zero.
60 to 90 minutes a week
The form and the reminder go out by themselves, and a briefing card is waiting for you.
A cancellation at a quarter to four for the four o'clock hour. You send a message and leave the gap.
40 to 70 minutes a week, plus the empty hour
The cancellation reports itself to your waiting list, and a payment link is ready for whoever cancels late.
The invoice for the programme that started last week is still a draft. You want no fuss.
45 to 75 minutes a week
After the session the invoice is ready for you to approve, with the reminder already queued behind it.
The word you use yourselfno-show
Your diary and your bookkeeping stay yours. We tie them together, we do not replace them.
Accountant
You chase paperwork. Every deadline again.
Now
With DDD
Chasing paperwork with personal emails, lists and phone calls. The PDFs arrive by WhatsApp, not through the portal.
8 to 14 hours a week, spread across the office
A checklist per file, reminders on fixed days in your own tone, and escalation to the case handler when nothing comes in.
Every exception from scan-and-recognise goes back to the client as a separate email.
3 to 5 hours a week
One queue of exceptions, and the same standard question per receipt instead of an email you write.
Getting a new client live runs on a checklist someone works through by hand.
2 to 3 hours per client
The signed engagement queues the tasks and sends the portal invitation.
The word you use yourselfchasing
Exact stays Exact. What we build is a layer around your software, not a second set of books.
The time bands are indicative figures from trade research of September 2026, not from a measurement at a client of ours. They say what the work costs now, not what we promise to save.
The other side
And if they do find you, the enquiry shouldn't be left sitting in WhatsApp.
- On the Map: eight pages are live, and ChatGPT says your name.
- Autopilot: the follow-up happens by itself. Reviews, measurements and posts keep running.
- Headquarters: one screen, Monday morning. Not eleven tabs.
We put you On the Map. Once that stands, we switch on the Autopilot. And then you open Headquarters.
What we build for thatMeasure, assess, build
We start with a measurement, not a conversation. That's not out of modesty: without a figure, we'd both just be talking about a hunch.
Measure
Whether you're mentioned when someone asks about your trade in your city, and who comes back instead.
Assess
Whether what's on your site is accurate, and what needs to change before anything gets built.
Build
The pages that are the answer, on your own domain and in a repo under your name.
You do not have to start with the scan. If you already know your site is the problem, the Website Stress Test is the direct route.
How we measureOur own scoreThe full measurement across four AI platformsWhat it costs
The full measurement gives you your TAZM score: one figure saying how much of your own search market you hold. This is how we work it out.
Why us
What makes us different
The whole chain, not just the page
Anyone who rewrites your page delivers a page. We measure first, assess what is there, build what needs building and tie it to your way of working, from the answer in ChatGPT to the enquiry in your inbox.
What is yours stays yours
Your repository, your data, your domain: it is all in your name and not in ours. The subscription is cancellable monthly, there is no minimum term, and there is nothing you are locked into.
We say what is not included
No advertising, no CRM, no complete IT department. Every step on the pricing page says what does not belong to it, and it says so before you sign rather than after.
EU, and we name the names
Your data sits in Europe: our servers run in Germany and Finland. We name every supplier on our stack page, so you can check it instead of having to believe it.
Boundaries
What we do not do
- No advisory report that ends up in a drawer. We build it, or we do not do it.
- No eight-week discovery. Your system diagnosis is there in a minute and a half.
- No platform you cannot leave. We build on a dedicated server of ours in the EU, with open source, and the code and the data are yours.
- No exit fee. If you leave, you get the code, the data and an export of the server, within ten working days, at no cost.
Our own figure
We measure ourselves too
Our own score stands at 7.1 out of 2 of 4 components, measured on 28 August 2026, with the same method we use for you. We publish every measurement, including the ones that disappoint.
See our own measurementOur own bill
We run on the same stack we build
This is what our own fixed software costs look like, measured, and what goes when we leave GoHighLevel in October.
Today
Everything together, measured
- GoHighLevelthe CRM the client sees today€270
- n8n Cloudcarries all nine production flows today€20
- Other fixed costsunchanged by the exit€521.53
Total€812/mo
After the exit
Without GHL and n8n Cloud
- Twenty CRMthe CRM, self-hosted€0
- n8n self-hostedthe successor to n8n Cloud, flows still to migrate€0
- Other fixed costsunchanged by the exit€521.53
Total€522/mo
Frequently asked questions
- What does it cost?
- The path has three steps and you can stop after any of them: measure, assess, build. What you already paid is deducted from the next step. Alongside that runs a subscription with which we measure and adjust again each month, cancellable monthly. You don't have to build first: if your site is already sound and you just want to know if it's moving, start there. All amounts are listed by name with explanation on the pricing page, excluding VAT, and they're listed there and nowhere else, so there's no second place that falls behind.
- What if ChatGPT still doesn't mention me after three months?
- Then we show why, and adjust the plan without an extra invoice until the next measurement. We don't give a guarantee that an AI model will mention your name, because no one can deliver on that, not even the agencies that do promise it. What we do fix is the rhythm: the same questions every month, the same yardstick, and the difference between two measurements in your inbox. Expect three to four months before a model consistently returns your name; what you have straight away is the starting point, with the date attached.
- Can I get just automation without visibility?
- Yes. The systems side works independently: if your problem is that the enquiry sits unanswered, not that no one finds you, we build that piece and nothing else. We charge for that by complexity, with an amount that's fixed once the scope is fixed. Start on the systems page; there are three examples there from an estate agency, a coaching practice and an accounting office.
- What if my current site is the problem?
- Then you don't need to start with the scan: the Website Stress Test is the direct route, and it sits as step 2 on the pricing page. It shows whether your site is repairable. If it isn't, we build a new one, on the same technology as the citation layer: Astro on Cloudflare, in a repo under your name. We discuss that in the conversation after the stress test and not before, because without that assessment it's a gamble, for you and for us.
- Why not do it myself, with ChatGPT and Zapier?
- You can, and for a single action it's often the fastest route. Where it goes wrong is at the second one: a flow no one monitors goes quiet without notice, and then you notice it through the client who heard nothing. What we build runs on its own server with monitoring and error alerts in place, alongside documentation of what happens. Visibility is different: there the problem isn't the build but the measuring. Without a baseline measurement using the same questions every month, you don't know if it's moving, and then you're optimising on a feeling.
Recent on our blog
What we build, what it costs, and what we learn from it. Articles are in Dutch.
Hoe meet je AI-vindbaarheid?
De meetlat-methode: vaste zoekvragen, een nulmeting over meerdere platforms en een maandelijkse hermeting waarvan de…
SigiWat zijn Google AI Overviews?
Wat het AI-antwoordblok boven de zoekresultaten is, hoe Google zijn bronnen kiest en wat het met je doorkliks doet.
SigiWat is GEO en hoe verschilt het van SEO?
GEO wil geciteerd worden waar SEO wil ranken. Wat hetzelfde blijft, wat verschuift, en wat volgens onderzoek…
SigiStart with the measurement
One question to three AI models, a result on screen, within twenty seconds. After that you'll know whether there's anything to gain here, and only then is a conversation worth having.