Roman Kapp at Gcore - Episode 163 - The Route to Networking

Building, Breaking and Reinventing with AI: A Conversation with Roman Kapp - AI Implementation Lead at Gcore

This week on the Route to Networking podcast, Ben Davies sits down with Roman Kapp, AI Implementation Lead at Gcore and a technologist with a uniquely human approach to innovation. Roman has spent more than fifteen years moving through psychology, corporate learning, entrepreneurship and now enterprise AI. His career is a constant cycle of building, experimenting, learning and starting again.

He joined us fresh from Cyprus, where it was twenty-five degrees the day before, only to land in Austria at minus one. Laughing, he said, “Yeah, it was twenty-five yesterday, and now it is minus one. Big change.” The contrast set the stage for a relaxed, open conversation.

Leaving the Comfortable Corporate World

Roman began his career in corporate learning, expecting to educate people and improve how organisations function. Instead, he discovered something very different.

“They say it is education, but actually it was just entertainment. Fun is important, but it should not be the only thing,” he recalls.

More than anything, he felt constrained. “I spent fifty per cent of my time writing documents instead of doing something valuable. There was not enough space for my ideas.”

The financial crash of 2008 ended up being an opportunity. He suddenly had time to think, read and analyse what wasn’t working. After devouring business books, experimenting with ideas and recognising his need for creative freedom, he made the leap into entrepreneurship.

“I fell in love with my idea, you know, like all entrepreneurs. So, I decided to leave this comfortable corporate world and launch my own business.”

That business ran for more than a decade and laid the foundation for how he approaches innovation today.

The Power of Trial and Error

Roman speaks openly about failure, iteration and experimentation. For him, it’s the only real path to success.

“You just try and test it. Error, try again, error, try again. One day, you finally find the solution.”

His major breakthrough came from an onboarding chatbot he built in 2017. Before that, many projects fell flat. But once he launched this one publicly, the reaction was instant.

“Thousands of people started playing with the chatbot. I said Wow, that is what the market needs.”

That experience cemented his belief that ideas must be tested quickly and in real conditions.

A Psychology Background That Still Shapes His Work

Roman studied psychology, but not therapy or counselling. His focus was scientific: designing experiments, analysing data and understanding how people behave.

“It is not about mindset. It is about how to construct experiments and analyse data. That is what helped me.”

These skills became invaluable as he moved into AI, where experimenting, measuring, and refining models are daily disciplines.

How AI Is Changing and Where It’s Going

Roman has seen AI evolve rapidly, but he is cautious about the hype.

“Every two weeks, there is something incredible. But it is impossible that we see game changers every two weeks. The world does not work that way.”

Instead, he sees deeper trends shaping the future:

Smaller models matter more

“Small language models you can run on your smartphone. That is the future.”

Blockchain as infrastructure

He emphasises the practical side of blockchain.
“Blockchain is exactly supposed to support agents transferring money and signing things autonomously.”

Robotics becoming affordable

Roman predicts household robots will eventually cost the same as a car, then drop to 5,000 or 10,000 euros, unlocking everyday use.

Bringing AI to Every Desk at Gcore

At Gcore, Roman’s mission is simple: make AI a daily tool for everyone, not just developers.

“We bring AI to each desk, to each employee, to increase quality and speed.”

But he stresses that tools aren’t the real challenge.

“The issue is not the tools. The issue is how to use them.”

Most organisations, he explains, lack structured knowledge. Information is scattered, outdated or locked away in different systems. That means one of the biggest tasks in AI adoption is actually cleaning and organising knowledge so AI can use it meaningfully.

“We spend a lot of time cleaning, structuring and putting knowledge together. With proper knowledge, results are much better.”

He also sees an interesting pattern: many juniors adopt AI faster than senior developers, simply because they are more willing to explore.

“Developers often adopt AI slower. They spend all their time on their tasks and have no time to test new tools.”

Roman himself integrates AI deeply into his daily workflow. Instead of typing messages manually, he speaks to an AI that rewrites his words in his exact writing style. “Even I cannot see the difference,” he laughs.

Context Engineering Is the New Skill

Roman believes prompt engineering is becoming less important as AI evolves.

“It is no longer about prompts. It is about context engineering.”

For him, AI performance depends on the quality of the information you give it. Clear, structured context leads to clearer results.

This shift is redefining how organisations integrate AI into their processes, moving from clever prompting to thoughtful knowledge design.

Leadership: Keep Things Simple and Keep Delivering

Roman’s leadership philosophy is built on clarity, practicality and delivery.

“I just want to deliver, deliver, deliver. I like this feeling when people use what I am building.”

But work should also be fun. He shares a story about a colleague who once worked somewhere that banned AI tools.

“He did his boring job in the day and played with AI at night. Now he does valuable work ten times faster and plays with AI at the same time. When the fun stuff is your job, that is a miracle.”

Advice for Future AI Professionals

Roman’s advice to newcomers is refreshingly simple.

“Do not be afraid. Try it. Watch videos. Try again.”

But he also encourages aspiring AI professionals to build foundational skills before claiming expertise.

He recently interviewed a candidate who described themselves as “AI native” but lacked core understanding. His advice:

“Spend a little more time. Know better what you are supposed to do and the tools you use.”

His golden rule?

“Find a simple problem, ask AI how to solve it and build something. Build, ship, build, ship.”

 The Quick-Fire Round 

We then finish off this insightful conversation with the all-important quick-fire round! Roman’s answers reveal his personality beautifully. He shares:

• the childhood moment that sparked his love for technology
• the idea he believes will reshape society
• the book that influenced his entrepreneurial mindset
• the skill he believes every AI newcomer should learn

But the details are reserved for those who tune in. No spoilers here.

 Listen to the Full Episode

Packed with honest reflections, practical advice and a grounded view of the AI landscape, this episode is a must-listen for anyone interested in technology, product development or the future of work. Whether you're a developer, leader or someone curious about AI, Roman’s perspective is both refreshing and insightful.

Connect with Roman Kapp on LinkedIn here.