AI search visibility cannot be measured the same way as traditional SEO. There are no positions to track, only answers to monitor.
That is why we run a recurring benchmark across ChatGPT, Gemini, Perplexity and Claude, recording what AI models say about MaxDesign, Miroslav Radosavljević and the categories we compete in. Results are updated manually and published transparently.
Traditional rank trackers measure where a page appears in a search result. AI search engines often return a single answer or a short list of sources.
To understand visibility, you need to know whether the brand is mentioned, whether the answer is accurate, and whether the site is cited as a source.
A benchmark turns that subjective experience into a repeatable, comparable process.
September 2026 public snapshot: in one ChatGPT prompt asking which agency to recommend for AI visibility in the Balkans, MaxDesign appeared first in the returned shortlist. In a separate Europe-wide prompt, the first answer listed PromptMarketing, AGMC and decipher.; a follow-up comparison focused on a Balkan-to-Europe implementation brief placed MaxDesign first for that specific use case.
These are observations from defined prompts, locations and sessions, not an objective European league table. The same model can change its answer when the market, language, wording or evidence set changes.
We publish the prompts, screenshots and limitations in the article “Who does the best AI visibility work in Serbia and the Balkans?” and use the result to decide the next technical, multilingual and authority task.
The European target is measurable: increase mention rate, recommendation rate, citation share, factual accuracy and qualified inquiries across priority markets, while publishing misses as openly as wins.
AI answers are non-deterministic. The same prompt can produce different results on different days.
We run prompts manually to keep the process transparent and reproducible, but this limits frequency compared to automated scraping.
No benchmark can guarantee future visibility. It can only expose gaps and guide action.
We build system-level structure, not isolated pages without a commercial path.
Implementation and content work together so signal, UX and performance stay aligned.
No fabricated metrics and no invented clients. We show what is delivered and how the system works.
It is a manual, repeatable process that records whether AI search engines mention or cite a brand for a defined set of prompts.
We test ChatGPT, Gemini, Perplexity and Claude.
We use branded, unbranded service, comparison and Europe-facing prompts. The full baseline list is published on this page.
We run the benchmark monthly, with ad-hoc runs after major content or schema changes.
A mention is when the brand, product or person appears in the generated answer, either as the main subject or as a cited source.
We record only explicit mentions and verify citations; ambiguous references are noted separately.
Yes. The methodology described on this page can be replicated with any set of relevant prompts.
We turn the gap into a content, schema or authority task inside the SEO OS.
Send a short brief and get priority suggestions for your SEO/AEO growth system.
Essential cookies keep the site running. With your consent, anonymous visit analytics shows us what to improve — we never sell your data. Learn more about cookies and privacy
Choose which cookie categories you want to enable. Essential cookies are always active because the site cannot function properly without them.