Twenty-two years in the markets taught me how businesses really make money. Now I help small and mid-sized companies put AI to work — on their terms.

I’m Sven Duve, founder of Neural Step. I spent twenty-two years in the markets — first on the stock exchange, then trading energy and commodities across Europe. The work was risk management: not single bets, but reading how many moving exposures offset and compound across a portfolio, and steering the whole position. Energy taught me to see the abstraction underneath it — that a complex value chain is really a set of cash flows driven by a few variables like price, time, and rate of change, and once you see it that way, it can be managed deliberately rather than left to chance.

That instinct — stripping something tangled down to the few things that actually move it — is what led me into mathematics, and then into AI. It helped that I could see AI reshaping my own industry from the inside, and I’d rather help build that future than watch it happen. So I went back to study properly: a degree in mathematics and statistics, then a master’s in artificial intelligence. I don’t just talk about this technology — I can build it.

Today I run Neural Step, working with small and mid-sized businesses on purpose. That’s where AI can make the biggest, fastest difference — and where being there in person, working alongside your team, counts for more than the size of the vendor. In practice, the interesting part is rarely bolting a chatbot onto what you already do; it’s rethinking the workflow itself — letting AI run the steps it can handle on its own while people stay in the loop where their judgement matters. Sometimes that means customer and internal questions answered around the clock; sometimes it means a whole multi-step process redesigned end to end. The common thread isn’t a fixed list of use cases, it’s any workflow worth rethinking — and everything is built to fit the systems you already run, with EU data rules in mind from the start.

My approach is simple: start from a real business problem, build something that pays for itself, and make sure your people work better with AI instead of being replaced by it. Trading taught me to treat return on investment as a real, measured number, not a line on a slide — and that’s no different from how business works, the more so as markets grow more transparent and products more commoditised. The result is a way of working that’s close, practical, and judged on outcomes.