For two weeks we ask ChatGPT, Perplexity and Google AI the questions your customers ask. Then you know where you get named, who gets cited instead of you and what is worth doing.
A real chat with the agent. Brands and places replaced.
Aron
I watch from above who shows up in the answers.
What is an AI visibility audit?
In plain wordsWe repeatedly ask ChatGPT, Perplexity and Google the questions your customers ask before they decide, and show when the AI names you, who it names instead and which changes follow from that. Specialists also call this a GEO audit, after generative engine optimization.
Like mystery shopping, with questions
A mystery shopper asks for advice in a store. We ask the AI what your customers would ask, for example “Which tax advisor in Vienna is good?”, and write down who gets recommended.
Not once, but many times
The AI answers a little differently every day. So we ask on several days and in every system. Only then can you tell whether you get named reliably or just by chance.
A clear list at the end
You get no data dump, but one recommendation per question: leave it, improve the content or add ads. Plus a three month plan.
Why a single test is not enough
Ask ChatGPT the same thing today and tomorrow and you often get different sources. Deciding after one try means deciding by chance.
Your brand14% (6 to 28)
Strongest competitor38% (24 to 54)
Illustrative example. The line is the measured share, the shaded area the likely range. If the areas do not overlap, the gap is real.
40%of sources are new the next day, for the same question
4 in 10questions have a different top source the next day
95%of questions: no website is cited by all three AI systems
3 vs 9sources per answer in ChatGPT and Google AI, compared with Perplexity
Doing well in Perplexity does not mean you show up in ChatGPT. And because ChatGPT names only a few sources, space there is tight. That is why we report every system separately.
Sources: more than 2,300 AI answers with over 17,000 cited pages in the Atziry platform, December 2025 to October 2026. Day to day comparison from over 100 repeatedly measured questions, system comparison from around 40 questions asked on the same day.
How the audit works
Four steps, from the first measurement to a three month plan. Each member of our crew stands for one part of the work.
01Murmel
Measure
Together we define the questions your customers ask before buying. For two weeks we ask them on several days in ChatGPT, Perplexity and Google AI.
Result: every answer and every source recorded
02Fini
Compare
We see which websites get cited instead of yours. And we check whether AI crawlers like OAI-SearchBot and PerplexityBot are allowed to read your pages at all.
Result: competitors and technical setup at a glance
03Ulli
Decide
For each question you get one clear recommendation instead of a long list. So you see where effort pays off and where it does not.
Result: one recommendation per question
04Stoani
Act
You get a report and a three month plan. Implement it yourself or with us. If you like, the measurement keeps running.
Result: a plan with a metric for each step
One decision per question
The most important part of the audit. Every question lands in exactly one of three groups, depending on how reliably the AI names you today.
Keep
Reliably cited
The AI names you here regularly. Maintain the page, do not rebuild it.
Content
Sometimes yes, sometimes no
A clearer, more direct answer on your page turns chance into a reliable mention.
Content plus ads
Reliably others
Here the AI reliably recommends competitors. We improve your content, and until that works, ads in ChatGPT make sure you still show up.
Atziry platform
Measured with our own technology
The measurement runs in our platform, not in spreadsheets. It shows per subtopic where a little effort achieves the most.
The real “Three quick wins” view from the platform. Example pages replaced.
Demand and winnabilitySorted by the number of AI citations in the topic and by how easy the spot is to win.
Concrete stepsWhich page is missing, which terms it should cover and which page gets cited as a role model today.
Alerts on changesIf something changes for a question, you get notified.
Keep measuring after the audit
3×times a week
5×times a week
7×times a week
What you have in hand at the end
Ulli
Everything in one report, with reasons.
Citation rate with rangeoverall, per AI system and per question
Competitor comparisonwho gets cited instead of you and how reliably
Technical checkaccess for AI crawlers, structured data based on Schema.org, quotable text blocks
Recommendation per questionleave it, improve the content or add ads
Three month planwith a metric for every step
Results meetingwith marketing and management, 60 to 90 minutes
Pricing
Fixed price before we start. It depends on the number of questions, markets and competitors, with no extra charges later.
AI visibility audit
€2,900from, net
Two weeks of repeated measurement
ChatGPT, Perplexity, Google AI
Competitors and technical setup
Decision per question, three month plan
Fixed price after the first call
SEO and AI visibility audit
€3,900from, net
Both audits in one report
Google and AI search prioritized together
One results meeting
€900 less than booked separately
SEO audit
€1,900from, net
Technical, content, Search Console
Prioritized actions
Fixed price based on the size of your websiteSEO audit
In the first call we define questions, markets and competitors and give you the fixed price.Book a free call
Want to check for yourself first? The free AI quick check shows in 10 seconds whether AI systems may read your website.
FAQ about the AI visibility auditIncluding the questions our own AI measurement showed as open.
What is an AI visibility audit?
A two week measurement of whether ChatGPT, Perplexity and Google AI name your brand or cite it as a source for your customers’ questions. It includes a competitor comparison, a technical check and one recommendation per question.
Is it the same as a GEO audit?
Yes. GEO stands for generative engine optimization, meaning optimization for AI answers. A GEO audit checks where you stand in those answers today. We call it an AI visibility audit because that is clear without jargon. What is GEO?
How do I measure AI visibility per system without relying on one overall score?
By asking every question on several days in each system separately. The next day around 40% of sources are new, and for 95% of questions the three systems do not all cite the same website. One overall score would hide these differences. So every value comes with a range and per system.
Why is my website invisible in AI search when my SEO works?
Because each AI uses different sources and often names only a few per answer. ChatGPT and Google AI name around 3 sources per answer, Perplexity around 9. A good ranking helps but is not always enough. The audit shows exactly for which questions you are missing.
Why does AI recommend my competitors instead of me?
Usually because their pages answer the question more directly or show up as a source more often. The audit shows per question which page gets cited instead of yours, so you can see what that page does differently.
What makes a good audit different from a weak one?
A weak audit asks once and shows a screenshot. A good one measures repeatedly, reports each system separately, gives ranges instead of random values and ends with a decision per question, not a long list.
We added FAQs and structured data. Why is our AI visibility flat?
Because both alone rarely suffice. In our data test, FAQ schema was linked to citations only in Perplexity, and a before and after measurement of 1,885 pages found no increase after adding it. More often what is missing is a good Google ranking, the answer in the first sentence and mentions in the sources the AI cites. The audit shows per service which pages are named instead of yours. Data test
What should we track for AI visibility: prompts, brands or keywords?
Real customer questions without your brand name, asked several times per service. That gives three figures: how often your brand is named as a provider, how often your website is linked as a source, and your share against competitors. Keywords stay important for Google, but AI answers have no fixed positions like 1 to 10.
Is AI visibility already worth it for our business?
A measurement before any budget shows this: for how many of your customer questions does the AI name providers at all, and who is named? Where the AI names no providers, there is little to gain. Where competitors are recommended regularly, the work usually pays off. Provider analysis
Can I check this myself?
For a first impression, yes: ask ChatGPT the same five questions on three days and note the sources. For reliable statements across many questions, several systems and competitors you need a measurement like the audit. The free AI quick check shows whether AI crawlers may read your site. AI quick check
What does the AI visibility audit cost?
From €2,900 net. The fixed price depends on the number of questions, markets and competitors and is set before we start. Together with the SEO audit from €3,900 net.
What is the difference between “mentioned” and “cited”?
Cited means your website appears as a source in the answer. Mentioned means your brand name appears in the text without a link. We report both separately, because only a citation brings visitors to your site.
Are ChatGPT ads or GEO the better choice?
It depends on the question. Content takes weeks to months before AI systems pick it up. Where others are reliably recommended today, ChatGPT ads can close the gap right away. The audit shows for which questions that pays off and for which it does not. ChatGPT ads
What happens after the audit?
You get a report, a results meeting and a three month plan. Implement it yourself or with us. If you like, the measurement keeps running in the platform three, five or seven times a week, so you can see whether the actions work.
What do you need from us?
One contact person, your key markets and competitors and, if available, read-only access to Google Search Console. We work out your customers’ questions together.
