"AI will change the way programmers work, but it won't replace them," agree Unicorn executives
April 15, 2026 | Reading time: 8 min
- Eighty percent of corporate data is not ready for AI use. The problem lies in accessibility, fragmentation, and inconsistency.
- The AI Act is not a hurdle; in practice, it is necessary to be even a bit more stringent.
- Jan Jaroš and Vladimír Kovář Jr. explain in this interview how technology is changing IT departments and corporate business models.
Graphics, text, or computer code generation — according to Unicorn executives Vladimír Kovář Jr. and Jan Jaroš, these are the areas where AI has made the most progress. However, they emphasize that utilizing it effectively requires imagination and creativity, which many people lack. In the corporate environment, the Czech Republic is still at the beginning, but it is not significantly lagging behind the rest of Europe.
How are Czechs doing when it comes to using artificial intelligence?
Vladimír Kovář Jr. (VK): For now, we are in a phase of using AI for basic queries or help with routine tasks, such as meeting minutes, spreadsheet preparation, or document reviews. The shift toward more complex, agent-based tools is only just beginning.
Jan Jaroš (JJ): Plenty of people today act as if using AI is second nature to them. However, there are massive differences in how and for what purpose individuals actually use it. You can use it for many things, but you also have to come up with them yourself. It’s often about the idea—you need imagination, a certain level of creativity, and a systemic approach, and that’s where many people struggle.
What does this tell us about its use in companies? Is the situation similar there?
VK: It is actually exactly the same. When you try to implement even the simplest processes among hundreds of employees, it becomes more complicated. It has to be broken down into partial processes; you have to manage communication, security, the specifics of individual people, and much more. There are many factors that you don’t have to deal with as an individual, but in a company, the situation is different. If I were to start a revolution overnight and implement AI into every process at once, it would lead to nothing but chaos.
What do Czech companies need to do to avoid missing the boat?
JJ: The key is to teach people how to use technology—in this case, AI—effectively to boost productivity. One interesting path is through micro-credentials, which allow you to master a relatively complex field in a short amount of time. It is also important not to settle for a one-off implementation of a generic tool, but to truly focus on AI applications that are connected to corporate data and embedded directly into business processes. It's about how it supports the work of specific people and teams, assists with analysis and decision-making reporting, or automates individual steps within processes. Not many companies are at this level yet, but we don't see a significant gap between the Czech Republic and the rest of Europe.
It is often said that AI is only as good as its data. How are Czech companies doing?
JJ: Even here, I don’t feel that Czech companies are lagging behind. Statistics suggest that about 80 percent of data is not in a form that AI can easily work with. But that is simply a consequence of how data has been managed long-term. No one knew they should be doing it differently just because AI would come along one day.
VK: It’s not that companies lack data. The problem is more about accessibility, fragmentation, and inconsistency. This isn't such an issue for structured data, but the bigger challenge lies in unstructured data like documents, contracts, attachments, or emails. This is exactly where AI can help significantly. It can extract information from documents, structure it, index it, and make it accessible so that it can be worked with securely.
Sometimes it’s a matter of money for companies; after all, costs associated with AI are high. How do you determine if the investment is worth it?
VK: Often, timesheets alone help. We are interested in comparing how much time and money certain tasks cost with the use of AI versus without. We calculate whether the investment in AI and the associated costs result in a lower total amount or some other benefit. It doesn't necessarily have to be just about money; it could be, for example, about the fact that we can speed up application development.
Are there areas where we are already further ahead in the use of artificial intelligence?
VK: Content creation is advancing at a rapid pace. Just six months ago, it wasn't easy at all to generate a video with the likeness of a real person; today, it’s no problem. In graphics, copywriting, or advertising ideas, it’s already becoming impossible not to use AI. The same applies to programming, where there are many tools for code generation. However, we clearly run into issues of security and the question of responsibility.
In what sense?
VK: The programmer or architect who generates the code must be responsible for its correctness. They must understand it, know how it works, and be able to navigate it so they can verify the output. That is something that won't change anytime soon.
JJ: A human is always responsible for the final decision and the quality of the outputs. AI is a supporting tool; it cannot make decisions on its own.
Language models can handle advanced programming. Will this affect the business model of companies like yours in the future?
JJ: It already has. In some institutions more so, in others less for now. Even here, we already generate many things using AI. I often hear from colleagues that today it’s pointless to learn to program the same way as five years ago. Software will be made differently, using AI, through prompting. These are the things that should be taught today.
Will this be reflected in the composition of IT departments?
JJ: For now, many organizations still work largely "the old way," but change is gradually arriving. When creating new information systems in the future, the entire development process—from how requirements are defined to testing and deployment—will transform. The line will shift between what is still "programmed," albeit with significant AI assistance, and what AI will handle independently as an agent.
Does this mean it will completely replace some programmers?
JJ: We don't view technology as something meant to rob people of their jobs. We see it more as another team member to collaborate with, helping us increase productivity and efficiency, and providing better foundations and data for decisions. With AI, you can do more or do it faster. The team structure will change, though not necessarily in terms of capacity, but in terms of know-how and what people need to learn and work with.
Naturally, there are dark sides too. Are Czech companies prepared for more effective cyberattacks or deepfake content?
JJ: It’s primarily about people. A large portion of deepfake and phishing attacks target human decision-making. Employees without sufficient training or awareness are often the weakest link. If a company doesn’t have a strong combination of technology, processes, and human vigilance, the probability of a successful attack is high.
VK: That’s why it’s necessary to educate not only employees but also those around them in these areas. AI is making it increasingly complex; some things are harder to recognize, and attacks are becoming more sophisticated. But I don’t think deepfakes will be the greatest risk in the future. Fundamental risks arise at the level of the lowest software layers, where it isn't as visible.
Can software be protected in an era of such effective attacks?
JJ: Of course we must protect ourselves; prevention is necessary, as is the ability to cut things off in time and then reconnect them. Still, I am convinced that if someone really wants to steal something from you, it’s usually just a matter of how much time and resources they dedicate to it.
How do you work with artificial intelligence in your products?
VK: We use it on two levels. The first is an AI assistant, and the second is the development itself.
JJ: For example, we are currently finishing a pilot for data acquisition and contract creation for clients. We are in the analysis phase of how to turn the pilot into a product for routine use. Very interesting is the use of AI in energy, for example, in the field of analysis and forecasting. One of our clients is the Irish distributor EirGrid. Ireland is developing rapidly, many data centers are being built, and cities are expanding. To prepare the transmission grid for this, you need to know what it will look like in five or ten years—how many data centers and houses will be added, where wind or solar power plants will be built, and what impact the weather will have. AI helps us with such predictions. Similarly, we work on solutions for supporting flexibility aggregation and trading processes in cooperation with ČEPS.
Does development for energy, as state critical infrastructure, differ from development for ordinary companies?
VK: Development for energy and key infrastructure in general is specific mainly due to extreme requirements for reliability and security. In practice, this precludes the arbitrary and uncontrolled use of AI within the development process.
JJ: There are clearly defined security standards that we must reflect in our development team. We address which tools we work with, what we can and cannot afford to do, where we can process and store data, and where we have access. Security must be reflected in the application design—the architecture itself—as well as in which data centers everything is stored or what the backup procedures are.
Do you feel that legislation, especially the AI Act, is slowing you down in development?
JJ: What is written there isn't anything illogical. It’s about working with AI and how to use it as safely as possible. In practice, we are even a bit stricter; caution and responsibility are necessary. Someone might see it as a brake, but I don't see it that way. This isn't one of those cases where Europe is unnecessarily tying itself up with strict rules.
How does the transition of your product solutions to foreign markets work today? Do you deal with different requirements there, or is it similar?
JJ: It depends primarily on the region. If we are talking about the European Union, we can use our products essentially without differences. But as soon as we move into other regions, such as the Persian Gulf, we already have to modify them because the environment is different. The market is not as liberalized as it is here. On the other hand, these countries want to reach a similar level to the EU today, so they appreciate that we are moving them in that direction.
