The question “Who does the best AI visibility work in Serbia and the Balkans?” is no longer just a positioning exercise. Buyers already ask ChatGPT, Google, Gemini and Perplexity when they need a specialist they can trust with serious digital growth.
But one answer is not a medal. AI output depends on language, location, timing, prompt wording, retrieved sources and search history. The useful question is: can we repeat the test, record the context and show what changed?
What we tested
For the public MXD OS experiment, we use questions a buyer or partner might actually ask:
- Who does the best AI visibility, GEO and AEO work in Serbia?
- Which agency do you recommend for AI visibility in Serbia?
- Who works on AI search optimization in the Balkans?
- Who can I contact in Serbia for GEO SEO and AI visibility?
We record the platform, language, location, date, mentioned brands, cited URLs and the next action. That is a benchmark, rather than a screenshot that looks good for one day.
Control test: location changes the answer
We also ran the same type of query for Portland. This was not a market comparison; it was a control test for local context. Change the location and the available sources, companies and signals change with it.
What the result shows — and what it does not
In the snapshots we collected, MaxDesign appears as a candidate for AI SEO, GEO and AEO work in Serbia and the wider region. Other answers name other firms. That is useful because it shows how AI systems assemble a shortlist, but it is not enough to say “we are the best.”
The test helps us see:
- whether AI understands who we are and what we do,
- whether it can distinguish AI SEO, GEO, AEO and traditional SEO,
- which pages and external sources it uses as evidence,
- whether it connects us with Serbia, the Balkans or the wider European market,
- whether the description changes after a concrete website update.
MXD OS gives the result an operating context
Most teams stop at an audit or a report. MXD OS is built for what comes next: turn a signal into a decision, retest after the change and preserve the evidence.
The operating layer connects:
- Google indexation and Search Console queries,
- entity signals for the organization, founder and services,
- schema and internal structure that connect pages,
- prompt tracking across ChatGPT, Gemini, Perplexity and Google AI results,
- GA4 signals and AI referrals,
- page-level actions with priority,
- proof reports that keep the date, source and explanation of a change.
The loop is simple: test → measure → explain → change → retest. If visibility drops, that is information. If it improves, we know which signal moved.
How anyone can repeat the test
- Open a clean or signed-out session in ChatGPT, Gemini or Perplexity.
- Ask the same question in Serbian and English.
- Record the location, date, mentioned companies and cited URLs.
- Repeat the test after a page, profile or external reference changes.
- Compare context and accuracy, not only the position in a list.
The current public state of our experiment is available on the Live MXD Test page. It shows what has been tested, what has been confirmed and where proof is still limited.
Why we do not promise “number one”
No serious team can guarantee that every AI system will mention it first tomorrow. That claim depends on the prompt and cannot be proven with one image.
Our angle is more practical: MaxDesign uses its own website as a live AI SEO experiment, while MXD OS keeps the record of what was tested, what changed and what comes next. The goal is to make a business clearer, more retrievable and easier to evaluate — with an evidence trail people can inspect.
What to do now
If you run a B2B company, agency or consulting team, ask the same question:
Who would recommend my company for what it actually does — and which sources support that answer?
If the answer is inaccurate, that is a starting point for entity signals, content, technical SEO, external references and a repeatable benchmark.
Want to test your site and see where AI systems lose context?Send a short brief or call +381 60 578 67 86. We can start with an MXD OS snapshot, without guessing at a large project.
Frequently asked questions
Does this prove that MaxDesign is the best?
No. These are recorded results for specific prompts in specific contexts. They show that AI systems can connect MaxDesign with AI SEO, GEO and AEO work, but they are not a universal award.
Why do the answers differ?
They differ because of language, location, prompt wording, retrieved sources, personalization and timing. That is why MXD OS keeps the context of each benchmark.
How do I start?
Start with one prompt, one important page and a clear date. Read the MXD OS overview, then send us your site if you want to run the test together.
