
Who Gets to Say No to a European Fintech?
The regulator writes the rules, but your banking partner decides what you can actually launch. My notes from our conversation with Victor Tuson Palau on payments, partners, and shipping faster with AI in a regulated business. Companion notes to Human in the Loop HITL
Ask most founders who can stop a fintech launch and they'll say the regulator. That's half the answer. The regulator sets the rules. The bank that provides your accounts or payment infrastructure decides which products and which customers it will support. That second "no" is the one that tends to surprise people.
That distinction was the thread through HITL #32. Oscar and I sat down with Victor Tuson Palau, founder and CTO advisor at Tuson Advisory, for a guest episode: one expert and one long conversation about how money actually moves in Europe, with no news rundown.
The second "no"
The part that stuck with me most is what changes when a fintech operates under a partner's licence. Your product roadmap is no longer just yours. The partner has its own risk limits, and those limits shape the business: which merchants you can onboard, which customers you can serve, what evidence you need to collect before you do.
The builder version of this is familiar. Every platform dependency is a quiet co-owner of your roadmap. In fintech the co-owner is a bank, and the bank's appetite for risk is a product constraint whether or not it appears in your spec. If you find out about it after you've set a launch date, you found out too late.
Europe is one market on paper
We started with Europe as a fintech market and why fintech can still move slowly there. A big part of that is local payment habits. How people prefer to pay, and what they trust, varies from country to country. A product that feels obvious in one market can feel strange in the next, and trust is earned locally.
Instant isn't always usable
Then we followed the money. A payment can be instant on the rails and still not be instant for the customer. Victor's example: a broker can receive an instant transfer before it shows up in the customer's trading balance. The money moved. The customer still can't use it.
I like this because it's the gap between a system metric and a user experience. "Settled in seconds" is a fact about infrastructure. "I can trade now" is what the customer cares about. We also covered verification of payee and how it fits into fraud prevention, which lives in the same space between moving money and trusting where it's going.
Stablecoins and FX risk
We spent time on stablecoins and cross-border payments, and on the less glamorous problem underneath them: currency risk. For a small business making a cross-border sale, FX movement and working capital can eat into the margin before the money ever lands. That's a real, boring problem, and real, boring problems are usually where the good businesses are.
We also talked about regulation as three things at once: a cost, a moat, and a product constraint. It slows you down, it slows your competitors down too, and it decides which products you're allowed to build in the first place.
The Ferrari and its brakes
The second half moved to engineering. How do fintech teams approve software changes? Where does the human belong in an AI workflow? What should an AI coding agent actually be allowed to do?
Victor's answer came through an analogy I'll be stealing: a Ferrari and its brakes. Good brakes are what let you drive fast. In a regulated team, the controls over what reaches production aren't the thing slowing AI-assisted development down. They're what let you ship code faster with AI while keeping control of what goes live.
That matches what Oscar and I see with clients. The teams that get the most out of coding agents aren't the ones with the fewest guardrails. They're the ones whose review and release process is clear enough that speeding up the writing doesn't mean losing track of what ships.
Agents with bank accounts, and where to start
We closed on what banking for AI agents could look like, which fintech bets Victor would avoid, and where founders should start. You'll want to hear those in his own words, so I'll leave them for the episode.
Try this before you set a launch date
Write down the customers you want to serve, which partner has to support that decision, and what that partner needs to see. If you're opening up to a new category of merchants, ask whether your provider supports it and what evidence your onboarding process needs to collect.
Then tell me: who could stop your next launch, and when would you find out?
You can find Victor at Tuson Advisory.
Human in the Loop — Watch on YouTube · Listen on Spotify · Listen on Apple Podcasts
Every episode — full show notes, sources, and the archive — lives at podcast.vallyseed.com. The show is produced by VallySeed, where Oscar and I help organizations design, build, and deploy AI systems that create measurable competitive advantage.

