Before a team tests software, it often asks ChatGPT which tools are worth a look. We measure for which use cases and comparisons your product gets named and build the pages that answer those questions.
Which software helps a team of 20 sort support requests?
For a team of that size, these are often named:
Tool A, with automatic routing
Tool B, affordable for small teams
A comparison article on a software portal
Your product covers exactly that, but is not named.
Schematic illustration, not an original screenshot.
Aron
I see when your product is missing.
What does AI visibility mean for SaaS companies?
In plain wordsAI names software it can clearly assign to a use case: what the tool is for, which team size, what it costs and what it integrates with. A list of features alone rarely answers those questions.
AI cannot place the product
A team looks for a tool to sort support requests. AI names others, although your product covers exactly this workflow.
Features instead of guidance
A tech lead compares two developer tools. The vendor page lists features but does not say which team they fit.
Informed, but no click
A buyer learns from the answer which kinds of software exist and only clicks when she needs prices or integrations.
What our data shows
Our own measurements and studies, with sources. Where a number applies to all industries, we say so.
85%of provider searches: ChatGPT names specific providers
2.48xhigher chance of a Perplexity citation when a page answers the question directly
−40%click-through rate for questions AI answers completely
35%of searches: none of four AI services names a provider
For questions about providers, AI almost always names names. The only open question is whether yours is among them. In our data test, a citation was most strongly linked to whether a page answers the question directly.
Sources: All values apply to all industries: 85% and 35% from our provider analysis with 873 searches in four AI services, October 2026. 2.48x from our data test with 522 searches, a correlation, not proof. −40% from our AI Mode study, summer 2025 vs summer 2026. Provider analysis · GEO data test · AI Mode study
Which questions lead to names
Not every question ends with a recommendation. When someone asks for providers, ChatGPT almost always names names, for knowledge questions rarely. The first two rows matter most for you.
Aspect
Example from your industry
ChatGPT names providers
Provider search
"which software offers the best automated invoice processing with intelligent receipt recognition?"
85%
Price question
"fleet software cost"
55%
Comparison
"audit proof archiving software comparison"
39%
Knowledge question
"what is an ai tool"
22%
Source of percentages: our provider analysis with 873 searches, October 2026, all industries. The example questions come from our analyses and show how users in this industry ask, translated from German.
The questions decision makers really ask
From our analyses: this is how decision makers ask when they look for offers like yours, translated from German. Next to them: what your website needs for them.
Provider search"which software offers the best automated invoice processing with intelligent receipt recognition?" · "which invoice processing platform integrates with our accounting and erp systems?" · "looking for a crm platform focused on data security for users in austria." Requirements in one sentence: function, integration and data protection. Pages per use case answer that.
Price question"fleet software cost" · "fleet management software cost" · "contract management software cost" Cost: show the pricing model and examples openly, otherwise AI names others.
Comparison"audit proof archiving software comparison" · "where can i request a demo for r&d software" · "what is an ai tool" Comparison and demo: comparison pages and a clear path to a demo.
Source: our own analyses of search queries, mostly from Austria, June 2025 to October 2026. Only questions that came up several times independently, without company, brand or personal names.
How we work
Four steps, each with a result you can see.
01Fini
Collect your customers' questions
Which questions do decision makers ask AI? From search data, customer voices and your expertise.
Result: question list by priority
02Murmel
Measure who gets named
Each question several times in ChatGPT, Perplexity, Gemini and Google AI, without questions containing your name.
Result: baseline per question and system
03Ulli
Answers on your website
Pages that answer the questions directly, with clear details: use cases, team size, prices, integrations and data protection.
Result: pages that can be cited
04Aron
Observe and refine
Measure the same questions again regularly and see what changes.
Result: a trend, not a snapshot
Which vendors benefit
Depending on the product, different people ask different questions. We measure the ones that fit your audience.
B2B SaaS
Business software
Long selection processes with many people involved, each asking different questions.
DevTools
Developer tools
Technical questions about languages, APIs and setup.
Vertical
Industry software
Questions about software for a specific industry, often with few names in the answers.
Data and AI
Data and AI platforms
Questions about data protection, hosting and use in the EU.
Atziry platform
See how AI talks about you
The platform shows how ChatGPT, Gemini and Claude describe your brand, which providers get named instead of you and where you can catch up fastest.
A real analysis across ChatGPT, Gemini and Claude. Brands replaced.
Per systemChatGPT, Perplexity, Gemini and Google AI separately, because they often cite different sources
Without your namewe measure questions the way your customers ask them, not questions about your brand
With reasonsfor every suggestion you see which data it is based on
What you get
Aron
Measure first, then act.
Question listthe questions decision makers ask AI, sorted by priority
Measurement per systemwho gets named for which question, you or others
Clear detailsuse cases, team size, prices, integrations and data protection on your website in a way Google and AI can read
Contentpages that answer specific questions directly
Trendregular measurement in the platform, clearly explained
FAQ: AI visibility for SaaS companiesIncluding the questions for which AI cites other providers today.
How do decision makers really ask about software?
Often as a long, precise sentence with requirements: function, integration with existing systems and data protection, for example "which invoice processing platform integrates with our accounting and erp systems?". Pages per use case and per integration answer exactly that.
How does our software get into AI recommendations?
When AI can clearly assign your product to a use case and finds good sources about it: pages per use case, honest comparisons, mentions on trade sites. We first measure for which questions you get named today. AI visibility audit
Which pages should we build first?
Pages per use case and comparisons that answer a selection question directly. In our data test, a citation was most strongly linked to whether a page answers the question directly. GEO data test
Do we need review platforms?
They can help, because AI often finds sources there. We have no own evidence that they are the strongest lever for software. We check which sources actually get cited for your questions.
AI comparisons describe our product with wrong features. What helps?
Wrong details usually come from outdated or conflicting sources, such as old pricing pages, comparison portals or review sites. What helps is a current page per feature and per comparison with clear details, corrected entries on the portals the AI cites, and a follow-up measurement of whether the new details are picked up.
How do we see this in sales?
Mentions per question and system in the platform, plus visitors and inquiries from AI sources under their own tag. That way you can follow what reaches the pipeline.
Does it work in German and English?
Yes. Many software questions are asked in English. We measure and write separately for each language.
What does it cost?
It depends on the number of questions, locations and languages and on whether we also write content. In the first call we clarify the scope, then you get a fixed quote.
