Rankin AI Makes the Final Six at IT Arena 2026 & GEO Insights

IT Arena 2026 Startup Competition: Rankin AI Made It to the Final

Lviv, September 2026

Rankin AI reached the final of the IT Arena 2026 Startup Competition, held in Lviv from September 25 to 27. Of the 189 startups that applied, 40 made the semifinals, and 6 reached the final stage, competing for a $60,000 prize fund provided by the Ukrainian Startup Fund.

Our CEO and co-founder, Nazar Dziadyk, pitched Rankin AI on the Business Stage alongside fellow finalists Alatyr Group, FNAV Systems, Sirocco Energy, Versi Bionics, and Zmiyar. The final was dominated by hardware, defense tech, and energy, which is exactly what Ukraine needs right now, and we're proud to have stood on that stage with teams building it. Rankin AI was the only SaaS and the only marketing product in the final.

We also ran a stand at the event, where Iryna Smuk (Head of Product), Anastasiia Hodiak (Marketing Manager), and Yaryna Voitseshchuk (Product Designer) spent the day talking with founders, executives, marketers, and brand teams. Those conversations proved as valuable as the competition itself.

The market knows GEO now, yet doesn't know how it works

A year ago, most conversations about AI visibility started with explaining what GEO (generative engine optimization) even is. That's no longer the case. Almost everyone who stopped by knew SEO well, and most had heard of GEO and could roughly define it.

Where people got stuck was the next layer. How do you actually track whether a brand shows up in AI answers? Where do the prompts come from? And once you see the numbers, what do you do to change them?

The education gap has moved. It's no longer "what is GEO." It's "how is it measured, and what do I do next."

How AI visibility tracking actually works

Since most visitors had no mental model of the mechanics, this is the explanation we walked through again and again:

1. Prompts go in. A brand adds its own prompts, or uses ones we suggest based on the information it shares about itself.
2. We run them the way a real person would. Not through an API, which often returns different answers from what a user sees in the app, but by emulating a real query inside the AI tool.
3. Country is a setting. You can track answers for specific markets, with 50+ countries available.
4. We collect answers from all AI models we cover and analyze them.
5. We extract what matters: whether the brand is mentioned, which competitors appear and how often, which sources are cited, and which attributes are used to describe the brand.
6. We check for hallucinations. If a brand adds facts about itself, we compare them with what models say, and every mismatch can be traced back to the prompt that produced it.
7. You see your metrics, gaps, and opportunities. All of this comes together in one place: your visibility and position across models and markets, where competitors win prompts you don't, which cited sources are worth targeting, and which content and positioning gaps to close first. Instead of raw answers, you get a clear picture of where you stand and what to work on next.

A common myth: "there's a database of real prompts"

Many visitors assume that somewhere there's a feed of what people actually type into ChatGPT. There isn't. ChatGPT and other AI assistants are private applications, and what a user types stays between them and the model.

That's why reliable tracking relies on prompts the brand chooses, run through real-user emulation. And while some ChatGPT traffic does show up in analytics thanks to UTM tags, that picture is incomplete by design: AI often mentions a brand without linking to it at all. Traffic data undercounts AI presence, and that gap is exactly what visibility tracking exists to close.

Tracking is table stakes. The real question is "what now?"

The most repeated question at our stand wasn't about tracking. Visitors treated tracking as a given. What they wanted to know was: I can see I'm not mentioned. What do I do about it?

As Nazar Dziadyk put it in his pitch:

Our answer is sources. AI models build their answers from the sources they cite, so the list of most-cited sources for your prompts is effectively a ready-made work plan: which articles to get included in, where to pursue link building, PR and brand placements, and where to improve how you're described. The higher your brand appears in those cited articles, the better.

For in-house teams and agencies, this turns AI visibility from a dashboard into a to-do list.

Big brands have a different problem: how they're described

For large brands, being mentioned isn't the issue. Being described correctly is. We heard the same story several times: "We're the best in our category, independent research says so, but when you ask AI, a competitor wins."

The reason is simple: models follow the sources, not the truth. If the pages AI relies on describe a competitor more favorably, the answer will too.

Rankin AI addresses this through attribute tracking, which shows which words AI uses to describe your brand, where you win or lose against competitors, and how often. From there, Sources and Content Intelligence help close the gap: inspecting how your brand is described on the pages AI cites, placing brand mentions with the right positioning on high-impact sources, and creating content based on tracked gaps.

Where Rankin AI is heading

When the jury asked how we differ from competitors, our answer was straightforward: most tools in this space try to measure everything. We're going deep in one direction, sources, because that's where brands can actually influence AI answers and measure the impact of specific actions.

Part of this already ships and our clients use it today. Next up is making the sources workflow simple enough for any team to act on, with or without a dedicated link-building function, plus deeper attribute tracking that lets brands trace where a description comes from and change it.

As Nazar describes the direction:

Everyone else measures everything. We're building the part that changes the outcome.

Want to see how your brand appears in AI answers? Start free or book a demo.