CEE AI Challenger Roundtable

AI Works When It Reaches the End User

On 15 June 2026, leaders from public administration, business, and European institutions gathered in Prague for the CEE AI Challenger roundtable, hosted by AAVIT and Digital Poland. The discussion centred on a single practical question: how to turn AI investment into real adoption across Czechia and the wider region.

Petr Bednařík, CEO of DataSentics, a Bull company, brought a clear argument to the table. The hard part of AI is not building it. It is getting AI to work in the real world, every day, so that it actually moves the business, not just impresses in a test.

The gap between experiment and production

Most of the current debate focuses on capacity: more compute, more funding, more startups. These matter, but they are the foundation, not the result. Getting AI into reliable production is still surprisingly hard. Many organizations run promising pilots, yet far fewer manage to integrate AI into complex processes where it actually drives efficiency and profit. Until that step is solved, investment does not convert into value. It just sits in the experiment phase.

Not all end users are the same

This is where the focus shifts. Funding startups to invent new technologies is necessary but not sufficient. The decisive move is to help end users actually deploy AI in practice. And the end user is not one group. Large enterprises and small and mid-sized companies operate very differently, adopt at a different pace, and need different kinds of support. A strategy that treats them as one market will reach neither. For the region, this is not abstract. CEE countries still lag behind the EU average in AI use among small and mid-sized firms, and that gap is exactly where adoption support should be aimed.

Where investment goes decides where value stays

If the goal is adoption, then the question of where investment flows becomes central. Support should not only fund the creation of new technology. It should be directed so that it builds a strong local AI ecosystem, with know-how and products anchored in the region.

Today most investment goes into infrastructure: compute, GPUs, data centres. That foundation matters, but the value and the profit are generated one layer up, in the software, the platforms, and the use cases built on top. Capacity alone is not enough. Without owning that upper layer, the value flows to those who do, most of them outside the EU. Adoption done well keeps more of that value where the investment was made.

Regulation as an enabler

Regulation was one of the four levers on the table, and the timing gave it weight, with the Council of the EU due to take a position on the Digital Omnibus the following week. Regulation should act as an enabler, not a brake. The goal is simpler and more predictable rules that let firms innovate, grow, and deploy, rather than a debate that stalls on the cost of compliance.

The real question

The question is no longer whether the region will use AI. It is whether Czechia and CEE will become active builders of the AI future or remain its users. That will not be decided by infrastructure alone. It will be decided by whether AI reaches the end user, runs in production, and creates value that stays in the region.