For years, the process was the same. A founder had an idea. They wrote a brief, maybe a deck. They came to a design agency. The agency ran discovery, mapped user flows, built wireframes, iterated on visuals. Weeks later, the founder got a set of polished screens. That was the product of design.
The brief is dead. Long live the prototype.
Design
06
.
03
.
2026
Szymon Michalczyk
That sequence made sense in a world where building was expensive and slow. You had to think everything through before writing a single line of code, because code was the bottleneck. Design existed upstream. Its job was to reduce risk before development started. That world is disappearing.
The starting point moved
The last few founders who came to us didn't bring briefs. They brought working products. Not wireframes. Not mood boards. Functioning prototypes built with AI tools. Onboarding flows, dashboards, settings pages. Rough, but real. Something users could click through and react to. It sounds like a small thing, but it rewires the whole conversation.
When a founder already has something running, you're not imagining a product together. You're looking at one. The question isn't "what should this be?" anymore. It's "what's wrong with what we have?" That's a different skill. And a more honest starting point.
The old model had a hidden problem
Here's something nobody in the design industry likes to admit. The traditional process, where weeks of discovery lead to beautiful screens in Figma, had a built-in conflict. The designer's job was to explore and refine. The founder's job was to ship and learn. These two timelines rarely matched.
Founders would sit through weeks of process, waiting to get something they could put in front of users. By the time the screens were done, the market had moved. The assumptions had changed. The beautiful work needed to be redone. AI tools killed that waiting period. A founder can now go from idea to clickable prototype in a weekend. The feedback loop that used to take weeks now takes hours. And that's not going back.
So what's left for design?
If founders can prototype on their own, and if code is no longer the bottleneck, what does design actually do? I think the answer is systems.
A prototype built fast with AI tools works. But it doesn't scale. The spacing is inconsistent. The typography has no hierarchy. Every screen looks slightly different because there was no shared logic underneath. The product feels like it was built by five people who never talked to each other. Because, in a sense, it was. Each prompt was a separate conversation with no memory of the last one.
This is where design craft still matters. Not in making things pretty. In making things hold together. The gap between a prototype and a product isn't polish. It's coherence. That invisible logic that makes every new screen feel like it belongs, without anyone having to think about it.
The deliverable changed. Quietly.
Something else happened that I didn't expect. Founders are no longer asking only for screens or components. They're asking for rules. Explicit instructions that their AI coding tools can follow so that new features stay visually consistent with the system.
Think about what that means. The output of design work is becoming less about how things look and more about how things should behave. Design decisions encoded into something a machine can follow. It's not about one tool or another. It's that when your engineering team builds with AI, the design system has to be readable by those AI tools, not just by the people on the team.
The real change is about timing
If I zoom out, here's what I think actually happened. Design didn't become less important. It moved in the timeline. It used to live at the beginning. Before code, before users, before anyone knew what would actually work. Now it shows up after. After the first version exists and users have touched it. After the founder already knows what sticks and what doesn't.
I'd argue that's a healthier place for design. Closer to reality. Solving problems that actually showed up instead of problems someone predicted in a workshop. And there's an irony here. By moving later in the process, design ends up more valuable. Because now it's working with real products, real usage, real constraints. Not assumptions.
The brief is dead. The prototype is the new brief.
_
Wherever you are in that timeline, we can help. Building an AI-powered MVP from scratch. Bringing structure to a prototype that's already live. Or running a classic design process from the ground up, because we get that larger companies and enterprises still prefer it that way, and for good reason.
