High score, zero mentions: the Click Made case, July 2026

Below, we describe the third measurement clickmade.pl and how to read a result that looks better at first glance than it really is. We described the earlier stages in the articles A graphic design studio in AI chatbots: the Click Made case, May 2026 and Off-site signals: what Click Made was missing in AI.

What has changed since 2 July?

The website now has what the previous report said was missing. Click Made has separate service pages for visual identity and print design, an expanded page about the author, structured data and improved local pages. The 27 July report lists the new service pages among what is on clickmade.pl and counts over 87 articles on the blog. Off-site, we are creating directory listings for the company.

What did the 27 July measurement show?

The visibility score rose from 0% to 47%, and the citation rate from 0% to 8%. This is the first measurement in which clickmade.pl appeared as a source in a model’s answer. The split between models changed too: Gemini and Mistral mentioned the brand for every question, while ChatGPT, Claude and Perplexity still left it out.

There was a big difference between these two models, though. Mistral put Click Made on average in position 1.8 on the list, i.e. right at the top. Gemini mentioned the studio on average in position 14.4, far beyond what a customer reads. The same 100% in both models therefore meant two completely different situations.

Why doesn’t 47% mean that AI recommends the brand?

Because the visibility score is a composite measure, while a mention answers a simple question: is the brand in the answer or not. The score takes into account how many models mention the brand and in which position. Two models out of five that mentioned the studio for every question produced a high average, but most models still did not mention it. The report shows this in the mentions column: for no question did a majority of models mention the brand.

This is a common trap when reading reports. One number goes up, and a customer asking ChatGPT still doesn’t hear about the company. That is why we always look at three things separately: mention, citation and recommendation. How to read these results in the panel is described in the article AI Visibility: how to read your scan results.

Where did the first citations come from?

Most likely from the new and strengthened service pages. A citation means the model gave the page as the source of its answer, so it had to find a specific passage on it that answers the question. Pages with the scope of a service, the process and the price give it such a passage more easily than a general description or a blog post. We describe the metric in the entry citation rate.

8% is not much, but it is a qualitative change: for the first time a model treated clickmade.pl as a source. The report also listed the domains the models cited alongside it, mainly websites about design and publication layout. It is a map of places worth appearing in.

Why did the recommendation not change?

Because the barrier was the same as in May. The report still pointed to a lack of independent rankings, comparisons and publications about the brand. The website and directory listings helped the models find and understand the studio, but did not give them a reason to recommend it over others. All four AI engines checked knew the brand, and one recommended it.

The report also noted that, in this category, models are keen to recommend studios with a strong presence on other websites and in review directories. This confirms the analysis set out in the article Off-site signals: what Click Made was missing in AI: the next step lies off-site. More on this in the article Knows the brand, but doesn’t recommend it.

What does the third measurement tell us?

• The composite score and mentions have to be read together. 47% with no mentions in most models is still a weak result.
• Each model reacts at its own pace. Two out of five reacted after a few weeks, three have not yet.
• Citations appear more quickly than recommendations, as they depend on the website itself, whilst recommendations depend on external sources.
• Always check the position on the list: 100% visibility in 14th place doesn’t give much.

Methodological caveats

Between measurements we refined the set of questions, and the number of models checked changed from three in May to five in July. That is why we compare the direction of change, not exact differences in percentage points. Why results can differ even with the same questions is explained on the page Methodology and in the entry frozen query set.

How does Aelo help you interpret these results?

The Aelo report shows the visibility score, mentions, citations and recommendation separately, broken down by model and position on the list. This makes it clear whether growth is real or comes from a single model. The measurement is done by Visibility in AI, the work plan is drawn up in Tasks, and the differences between models are explained by the Advisors. We describe what the first scan looks like in the article Your first measurement in Aelo: what happens and when the results are ready.

Small businesses will find more on the page Aelo for small businesses, personal brands on the page Aelo for personal brands, and agencies that measure their clients, on the page Aelo for agencies. You can take your first measurement on the Free account, without a card.

Try it for free

FAQ

Sources

• Aelo, visibility reports
• Aelo, a methodology for measuring visibility in AI: https://aeloapp.io/wiedza/metodologia/