From our work
Customers increasingly ask ChatGPT, Claude and Gemini instead of Google. AI visibility is the work of getting those tools to know you, describe you correctly and have a reason to recommend you.
249 € one-time + 119 €/mo price breakdown →
I want AI visibility →Where your customers ask today
Example scenario
Who do you recommend in Bratislava for [your service]?
This is how someone who has decided to buy asks. They do not work through ten links, they take the recommendation.
I would recommend three options: company A, company B and company C. All three clearly describe what they do on their websites, and I also found independent reviews of them.
Your business is not in the answer and the customer never hears about you. Not because you are worse, but because the model has nowhere to read about you.
I would recommend your company, company A and company B. The first does exactly what you are describing and spells it out on its website, with ballpark prices.
The same question, the same three places in the answer. All that changed is what the model builds it from.
A diagram, not a screenshot. It shows the mechanism we work on, meaning which sources a model builds its answer from. It is not a guarantee of ranking in any particular tool.
What's included
The technical names for this are GEO and AEO, optimization for generative search and for answers. In practice it is these eight things.
Access for AI models
We check whether AI tools can reach your website at all. Blocking at the hosting level, or a site that only renders in the browser, means you are invisible to them no matter how good your work is.
Facts in visible text
Prices, cities, service names and opening hours have to be in the text of the page. A model takes what it sees when the page loads, not what sits in an image.
Content written around customer questions
A customer does not ask for the name of your service, they ask about their problem. We rewrite headings and opening paragraphs so they answer those questions directly.
Definitions of terms in your field
Short definition pages and FAQ sections that models quote from, because the answer is right there in the first sentence.
Structured data and company identity
Schemas that tell a model who you are, what you sell and for how much, and that tie mentions of you from different places back to one company.
Fixing what models say about you incorrectly
An old price, a service you no longer offer, the wrong city, or a mix-up with a similarly named company. We find it and fix the sources the model draws it from.
Monthly measurement across nine tools
The same set of questions every month in ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude, Copilot, Meta AI, Grok and DeepSeek.
A report with four components
The exact wording of the answers, whether your business is in them, how the model described you, and a comparison against last month. You see movement, not hunches.
How we assess it
Most companies ask a single question: will AI mention me? That is only the first of four. A model can mention you and at the same time credit you with a service you do not offer, quote an old price or present you as the cheaper substitute for a competitor. So we track four things separately, and with each one you know exactly where you stand. We also unpack what the term means in the glossary.
Visibility
Does your name come up?
How often the model names you on the questions your customers actually type. Not just the ones where you are the only possible answer.
When it doesn't fit, the model says so
“I do not know that company, but I can recommend several others in [city].”
Accuracy
Is what it says about you right?
Whether the description matches what you really do and for whom. This is where the most expensive mistakes happen, because the model does list you, just with someone else's service or an out-of-date price.
When it doesn't fit, the model says so
“Yes, I know that company. They do [someone else's service] and cost roughly [old price].”
Structure
Can it read your website?
Schemas, headings, llms.txt and access for crawlers. The technical layer that decides whether a model has anything to build an answer from at all.
When it doesn't fit, the model says so
“I did not find enough information on their website to be able to recommend them.”
Authority
Does it have a reason to trust you?
Whether anyone other than you backs up what you say. A model favors a company whose claims it can verify from the outside.
When it doesn't fit, the model says so
“They say that about themselves on their own website, but I have no way to verify it from another source.”
Why now
Audit
We find out how AI tools see you right now and where there is room to improve.
Content
We create content specifically for AI to consume: facts, definitions, answers to real questions.
Technical layer
Implementing schema markup and structured data for maximum readability by AI models.
Monitoring
Regular visibility checks in ChatGPT, Perplexity, Google AI Overviews and others.
Methodology
Once a month we run the same set of questions through every tracked tool and record the answers verbatim. The set is built for your field and your city, and we do not change it from month to month.
It sounds like a detail, but it is the whole difference. Without the same input there is no way to tell whether your website improved or the model was simply in a different mood. That is why we always measure with the same thing.
We mix the questions across types so that one strong area does not cover up a weakness somewhere else.
What is in the monthly report
The exact wording of the answer to every question, whether your business is in it at all, how precisely the model described you, which sources it picked for that, and a comparison against the previous month. That this is not speculation is confirmed by Google in its own documentation on AI features in search.
Comparative
“Which company is better at this?”
Shows whether the model puts you next to the competition or below it.
Recommendation
“Who do you recommend in Bratislava for…?”
The closest thing to how someone with money in hand asks.
Problem-based
“I have this problem, how do I solve it?”
The customer describes a problem, not a service. Most companies are missing here, because their content is written around the names of services.
Price
“Roughly how much does it cost?”
Reveals whether the model works with current numbers or fills in the blanks.
Non-branded
“Who does this in Slovakia?”
A question without your company's name. This is where your real position shows, because the model has to pick you on its own.
Branded
“What do you know about the company…?”
Straight at you. Here it is accuracy that shows, meaning whether the model tells the truth about you.
Staying power
The most common objection, and a fair one. That is why we do not build on how one particular model behaves today.
What changes
The models and their versions
Today's ChatGPT is not the one from a year ago, and a year from now it will not be today's.
The format of the answers
Sometimes a list, sometimes a paragraph, sometimes an answer with links to sources.
The list of tools
Every few months another assistant turns up and claims a share of the questions.
What stays
The content and data on your website
We work with the content and structured data of your website, meaning the thing all the models draw on at once. That survives a change in any one of them.
The same set of questions
When the result moves, you can tell whether it is your website behind it or a change on the tool's side.
The way we measure
We add a new tool to the measurement. We do not rewrite the strategy because of it.
AI visibility from us means two things. Work on your website so that models can put an answer together from it: access for AI crawlers, facts in visible text, structured data, and fixing what they say about you incorrectly today. And monthly measurement of the same set of six question types across every tracked tool, the answers recorded verbatim, tracking of four components (visibility, accuracy, structure, authority) and a report comparing against the previous month. Initial setup from 249 € one-time, monthly measurement after that from 119 € and with no lock-in.
A model does not go through the web the way a person does. It builds its answer from what it can read and what it trusts. If a page has no structured data, does not directly answer the questions people type, and nobody links to it, the model has nothing to draw on. A website can look great and still be practically empty to a model.
We went through the most common causes in the article why ChatGPT does not recommend you.
From three places. From your own website, meaning the copy, the structured data and the llms.txt file. From pages that mention you, meaning directories, articles and reviews. And from what the model learned during training, which nobody can reach any more.
The first source is entirely in your hands, the second partly, the third not at all. That is why only the first two can be worked on, and why no specific placement in the answer can be promised.
Your website and its structure we influence completely. Mentions elsewhere partly, that is work on authority and it takes the longest. The behaviour of the model itself we do not influence at all, that is where the version changes and so does the way it puts answers together.
That is why we measure with the same set of questions every month. When the result moves, you can tell whether it is your website behind it, or only a change on the tool's side. And that is why we sell this as a head start, not as a guarantee of first place in the answer. Anyone who claims otherwise either is not measuring it or is lying.
What it costs
Setup249 €one-time.
Then measurement119 €a month, no lock-in.
The setup is one-time work on your website. The measurement is the part that repeats, which is exactly why it can be stopped at any time. The price is final, we are not VAT registered.
01 In the setup
02 In the monthly measurement
+ What pushes the price up
It is the same price as SEO, and that is no accident. It is the same kind of work, only for a different way of searching. You can order either on its own or both together. An overview of all services is in the pricing.
From our work
Related services
The future of search
