FutuStories - Olli Kilpi

You’ve talked about the value of rapid experimentation in your work. What makes it so important for you?

If I didn’t spend time testing new things, I wouldn’t have the skills ready when projects demand something new. Quick experiments help me build a library of tested ideas, compared to slow, heavy processes where you might only test one. With AI, I can try many things fast and connect the pieces later.

Can you share an example of how you’ve used AI in your design work?

Recently, I skipped the Figma prototype and handed the developers an AI-generated code snippet instead. It was small but functional, and it worked. It felt like a real contribution, because it wasn’t just a design someone interprets later; it was something that became part of the actual product.

How does working this way change your view of design?

Tools like Figma are great for collaboration, but they’re always a step removed from the real product. The product isn’t the file — it’s what users actually interact with. AI lets me work closer to that core, focusing on usability and interaction rather than surface styling. It feels like doing design where it matters most.

You have a background in industrial design. Do you see a connection to what you’re doing now?

Absolutely. In industrial design, you build real prototypes, things people can touch and use. A chair has to be something you can actually sit on. I feel like I’m doing the digital version of that now. My “material” is code instead of clay. It’s like digital industrial design, hands in code instead of plaster.

If you could envision your dream project five years from now, what would it look like?

I’d like to work with fast iteration and testing with real users. If development becomes faster thanks to language models, we’ll probably see A/B testing become much more common, more “hands in the code.”

Five years is a long time. I hope I’ll be designing for something other than phones and traditional computers; maybe a “home robot” that you train by talking to it sensibly!

In my dream project, I’d work with developers right from the start, not tied to any specific platform or legacy code. We’d focus on solving a real problem, and during the concept phase, already test functional coded prototypes that we can easily scrap or rebuild. AI can really help there.

Once we’ve found a good solution, we’d implement it so it’s maintainable and provides usage data for continuous improvement. Maybe ownership is what I’m hoping for — the feeling of truly owning the product. And paradoxically, AI might actually increase that sense of ownership, even if it first seems like it would take it away. AI allows such fast development that the product ends up reflecting its creators more than ever before.

What keeps you inspired outside of work?

Skateboarding, building Lego with my kids, drawing, photography, 3D modeling, animation, and generally making things.