AI prototyping changes the rules of the game between designers and developers

Every product team knows this pain: a designer brings beautiful static screens, the team gathers for a review — and it turns out that half the user scenarios simply weren't thought through. What happens if there's a loading error? How will the interface behave with empty data? Where will that button click lead? An exhausting back-and-forth in comments and chats begins, and instead of handing the task to development, the team gets bogged down in clarifications. The product group at Selectel responsible for dedicated servers decided to break this vicious cycle using AI tools for prototyping — and their experience deserves attention.

The problem described by Selectel's senior product manager Alia is familiar to the industry. The classic waterfall approach to interface design — when mockups are first drawn, then agreed upon, and then handed to developers — generates systemic time losses. Static pictures in Figma are by definition incapable of conveying interaction dynamics. They show the ideal scenario, but say nothing about edge cases, error states, screen transitions, and responses to user actions. As a result, bugs in interface logic are discovered only after the code has been written, and rework at this stage costs many times more.

It is precisely into this gap between static design and working code that AI tools are now actively being embedded. The Selectel team turned to Figma Make — a relatively new service that uses generative artificial intelligence to turn mockups into interactive prototypes. Instead of manually scripting every interaction and every transition, the designer states the intent, and AI builds out the logic. This fundamentally changes the pace of iteration: what used to require days of painstaking prototyping work now takes hours.

But it's not just about speed. The key advantage of AI prototypes is that they force the team to face real use scenarios much earlier in the development cycle. When you're not looking at a static picture but a working interactive mockup, questions like 'what happens if data doesn't load' arise naturally — because the prototype literally demands answers to them. Testers get the chance to check user scenarios before the first line of production code is written. Developers see not abstract screens but a concrete model of interface behavior, which radically reduces the number of interpretations and misunderstandings.

Selectel's experience fits into a broader trend. Over the past year and a half, the market for tools at the intersection of design and development has experienced a real boom. Vercel released v0 — an AI interface generator based on text description. Dozens of startups have emerged, offering to turn screenshots and sketches into working code. Figma itself is actively integrating AI features into its platform. All these tools solve the same fundamental problem: the gap between what the designer sees and what the developer understands. Artificial intelligence acts as a translator between two worlds, creating a common artifact — a working prototype that both sides can evaluate unambiguously.

It's important to emphasize the limitations of the approach. AI prototypes don't replace either designers or developers. Generated code is rarely production-ready, and automatically constructed interaction logic may not account for the specifics of a particular product. We're talking about a tool for validating ideas, not about replacing professional development. The Selectel team, based on the description, uses these services in exactly this way — as a way to quickly test a hypothesis and ensure that all scenarios are accounted for before investing resources in full-scale implementation.

Nevertheless, the direction is set. The boundary between design and development will continue to blur, and AI tools will take on increasingly more of the routine work of turning visual ideas into working interfaces. For product teams that still live in a cycle of endless approvals of static mockups, this is a signal to reconsider their processes. Not because AI will magically solve all communication problems, but because it provides a tool that makes these problems visible at the earliest stage — when it's still cheap to fix them.