Edip Kilinc is the founder of Webwavers. Together with his team, he develops websites, brands, and digital presences for ambitious companies – with over 250 completed projects and accolades including the German Design Award 2026.
A position in Google Search is clear-cut: third place is third place. A generated answer is not. It turns out slightly differently with every query, sometimes listing three sources and sometimes none, and it does not necessarily link to what it is based on.
In addition, a large portion of these queries end without a click. The user gets the answer and leaves. None of this appears in Google Search Console, nor does it show up in website analytics. The visibility exists, but it leaves no trace in the tools that most companies use.
As a result, anyone who wants to know how they stand in AI systems must actively look. Passive tracking does not provide the picture.
A position in Google Search is clear-cut: third place is third place. A generated answer is not. It turns out slightly differently with every query, sometimes listing three sources and sometimes none, and it does not necessarily link to what it is based on.
In addition, a large portion of these queries end without a click. The user gets the answer and leaves. None of this appears in Google Search Console, nor does it show up in website analytics. The visibility exists, but it leaves no trace in the tools that most companies use.
As a result, anyone who wants to know how they stand in AI systems must actively look. Passive tracking does not provide the picture.
Mention rate. In what percentage of answers to a defined list of questions does the brand name appear? This is the basis. A mention without a link is valuable because it builds trust and often leads to a later brand search.
Citation rate. In what percentage of answers is your own website cited as a source or linked? This is the stricter standard and the one that brings visitors.
Share of Voice. How often does your own brand occur, measured against all mentioned providers? Two mentions out of three mentioned providers is different from two out of twenty.
A fourth metric is sentiment: Is the brand mentioned as an example, as a recommendation, or as a side note? This is difficult to capture automatically, but is instantly recognizable when reading the answers.
No tools are needed for an initial measurement. What is needed is a list of questions and discipline.
Step 1: Formulate twenty questions that a customer would actually ask. Not keywords, but sentences. A mix of category questions, comparison questions, pricing questions, and problem questions.
Step 2: Enter each question into at least three systems. Important: without being logged into an account or in a new window, so that no personalization plays a role.
Step 3: For each answer, note three things – does your own brand appear, is your own site named as a source, which competitors show up.
Step 4: Enter this into a spreadsheet and calculate the three rates.
The result is a snapshot, not a statistically clean survey. To answer the question of whether you appear at all, it is perfectly sufficient.
Meanwhile, there are providers who automate exactly this measurement: they regularly send a defined prompt list to multiple systems and evaluate the responses. This saves time and provides a progression instead of a snapshot.
One should be aware of three limitations.
The prompt list decides everything. A tool only measures what you let it measure. A poorly chosen list produces a nice curve that has nothing to do with reality.
The responses fluctuate. The same system delivers different results to the same question on two different days. Individual values are therefore worthless; only trends over weeks are meaningful.
No tool covers everything. The systems differ greatly in where they draw their information from. Whoever measures only one, measures only a section.
A measurement without consequence is just a report. It becomes interesting when you ask why a competitor is mentioned and you are not.
In most cases, it is due to one of four points: There is no content on the topic. There is some, but it does not answer the question directly, only after a five-paragraph introduction. The content is there, but not technically correctly marked up. Or the brand is not mentioned anywhere else on other sites, so no model has a reason to know it.
The fourth point is the one most commonly underestimated. Generative systems rely heavily on what others write about a brand – directories, specialist articles, comparison lists, forums. A brand that is not written about anywhere remains invisible, even with a perfect website of its own.
Monthly is sufficient in most cases. The underlying models and their sources do not change daily, and measures take time to have an effect.
A fixed procedure makes sense: the same list of questions, the same systems, the same day of the month. Only in this way are the values comparable. Anyone who expands the list in between should continue the old one in parallel, otherwise the series will be interrupted.
One-off additional measurements are worthwhile after major publications, after a mention in a specialist medium, or after a technical transition.
The most important appointment is the first one. Those who measure today can say in six months whether something has happened. Those who do not measure can never prove it – neither to themselves nor to others.
