He automated away part of his own job, then built an escape hatch

Dario Codipietro built Fiuto, at first, for himself. He was working as a lead designer inside a lean AI product, and while design and development had become dramatically faster, one part of the job stubbornly had not. User testing was still painful, mostly manual, and time-consuming enough that it was the easiest step to quietly skip. So he built a tool that solved his own problem, and then wondered whether anyone else had it too.

They do. Codipietro's argument is that there simply are no good self-serve, AI-powered user research tools. The category is heavily gated: paid AI, paid tools, your data locked down. The incumbents are slow and do not plug into modern agentic workflows, which turns research into a luxury reserved for teams with plenty of time, budget and dedicated staff. And if you are building with AI, none of that fits, because you need research at the pace of delivery, which is now measured in days, not months.

What Fiuto does

Fiuto is an agentic user research platform that can build, launch and run studies, with AI and with real users, and then hand back insight on the feedback. It can act as an expert research assistant and return something useful in minutes, from nothing more than a URL. One example he points to is AI testers: by no means perfect, he says, but good enough to pre-validate a product before it goes anywhere near real users. His edge, he believes, is speed, delivering features at a pace competitors cannot match.

It is a very lonely, Don Quixote-esque endeavour. People think I'm mad most of the time, and sometimes I doubt I'm not.

The part nobody warns you about

Codipietro is honest that he underestimated how hard doing everything alone would be. He knew it would be difficult in the abstract; the reality, the sheer number of areas you have to hold in your head at once, was much heavier than expected. Had he known in advance, he admits, he might have taken a more cautious route. One concrete casualty was scope: he cut an entire upstream product-intelligence section, the part that pulled insights directly from connectors like GitHub, Linear, Notion and analytics tools, and narrowed Fiuto down to downstream tests and results. The lesson, he says, was to focus on one thing at a time.

There is an urgency underneath it all. Codipietro watched his own tools reduce the need for a dedicated design team, which raised an uncomfortable question about the value of his own traditional work. Fiuto is partly an escape hatch: a way to start creating his own products rather than helping other people build theirs, before, as he puts it, the next startup boom is over. It is bootstrapped, and his ambition is grounded. Not a billion-dollar company, but enough paying users to make a decent living and, eventually, the freedom to build the next thing.

This profile is drawn from Dario Codipietro's own account of building Fiuto, an early-stage bootstrapped venture. Product details are as described by the founder.