Uhlík vs Křemík · Czech podcast about artificial intelligence

Why Czech Companies Struggle with AI More in Their Heads Than in Their Technology

Episode 3 · 17 February 2026 · 1:15:05 · Guest: Lukáš Benzl

Lukáš Benzl, director of the Czech Association of Artificial Intelligence, talked with hosts David Salomon and Adam Hanka in the third episode of the Uhlík vs Křemík podcast about why AI adoption in Czechia often stalls. The conversation also covered the European AI Act, fears of a bubble and the effort to bring an AI gigafactory to Czechia.

Technology is not the main problem

Benzl distinguishes two levels of AI adoption. The first is personal: everyone wastes time on emails, presentations and content and can get help even without a deeper strategy. According to him, however, this level often lacks business value, because ten minutes saved a day does not earn a company much money. The second level is larger projects across entire departments or processes, where the return on investment can be calculated.

According to Benzl, the biggest obstacle is company culture. The association's data do show things changing for the better, but in reality, he says, more than half of companies do not have a culture suited to adopting AI. "AI is not a technological discipline, it is an organizational discipline." Czech companies, he says, are also not used to experimenting, and hardly anyone gets praised for a week of trying something that leads nowhere. That is why the association wins over people who are afraid of AI in a simple way: it asks them what annoys them most at work and shows that it can be solved.

Adam Hanka, who leads data and AI at Creative Dock, named the two most common mistakes even before the guest arrived. People do not know how to prompt properly and do not have their data ready. Professional prompts, he says, often run to three, four or even ten pages of text. Benzl sees it even more deeply: for him, prompting is mainly the art of describing a problem well and communicating with the other side.

Regulation is necessary, but still vague

Benzl has a clear position on the European AI Act. Regulation is necessary, he says, because AI can do genuinely bad things. "None of us would want cameras in public spaces to film us, monitor our emotions, evaluate our behavior and, based on that, decide whether a bank gives us a mortgage or not." This is exactly what the European Union prohibits.

At the same time, he admits that the regulation is vague and that how it will work in practice will only become clear from the codes of practice. The association proposed simplifications, for example simpler documentation for primary research, and also dealt with AI literacy, which was extremely vague in the regulation. According to Benzl, the European Commission is proposing a dual labeling of content: a red icon for content created entirely by AI and an orange one for assisted generation. Whoever publishes content under their own name is responsible for it.

Where Czechia can get ahead of the world

Benzl sees areas where Czechia could excel. The first is adoption itself: by his estimate, roughly a third of people in Czechia use some AI service daily, while in some countries, such as Singapore or, if he is not mistaken, Norway, 50 to 60 percent of people already use AI at work in public offices. He also names healthcare, the defense industry and public administration. The Ministry of Health has set up a committee on artificial intelligence, where Matěj Misař, founder of the startup Carebot, represents the association.

A major topic is the AI gigafactory, a large data center with 100,000 of the most advanced chips. According to Benzl, the Czech government would have to release about 15 billion crowns for it, České Radiokomunikace is to be the private investor and part of the money would come from the European Union. The association has supplied data on how many companies would want to use the center, and the business case looks positive so far. Benzl would see the mere submission of the application as a strong signal that Czechia is serious about AI.

Not a dot-com bubble, according to Benzl

Asked whether a bubble is forming around AI, Benzl replies that it is not a dot-com type bubble. Money, he says, is not flowing into small companies without value, and large companies have well thought-out monetization models, such as advertising in chatbots. Some stocks are overvalued, however, and the market will see a correction.

On the job market, Benzl does not see a clear winner between specialists and generalists. He warns, though, that locking yourself into a narrow specialization and ignoring the world around you is dangerous. According to the latest reports from the Labour Office, even qualified programmers in Prague are now being laid off, something nobody would have expected before.

The full conversation with Lukáš Benzl is available in the Uhlík vs Křemík podcast.

The episode is in Czech. Czech page with chapters, quotes and the full transcript