The Prototype Is the Pitch Now

There's a moment in every project I used to dread: the handover. You've spent days perfecting a flow in Figma, and then you present it, a static picture of an idea. The client nods, but they're squinting at a screenshot of something that doesn't move, doesn't respond, doesn't feel like anything yet. Everyone fills in the rest themselves. And that's exactly where good ideas get misunderstood.

That part of my job has changed. I want to be honest about how, without dressing it up as a revolution when it's really something quieter and more useful.

Validation, in the span of an afternoon

The biggest shift for me is validation. How fast can I prove out an idea, an animation, an interaction, a whole flow , by actually building it instead of describing it? With a tool like Cursor, the gap between "I think this works" and "here, try it" has shrunk to almost nothing.

We don't hand a client a static file anymore. We send a real link. They can see it, test it, poke at it, do whatever they want with it. That changes the whole conversation. You're no longer defending a picture; you're watching someone use the thing. The feedback that comes back is real, because they interacted with something real.

Not a small upgrade. It's the difference between "I imagine this feels smooth" and "it does."

The morphing navbar, and the day I stopped just prompting

Here's when it clicked. I was building a morphing navbar, it starts as a rectangle, then opens into a square. Easy to describe, annoying to get right. And I couldn't get it to work. I kept prompting, it kept missing.

Then I realized the tool wasn't the problem. The logic was wrong, my logic. Once I understood why the animation kept breaking, I told Cursor exactly how it should behave: the timing, the easing, the precise specs of the motion. It solved it.

That reframed everything for me. I wasn't asking the AI to guess what I wanted. I was directing it, giving notes the way you would to an animator, because I finally understood the mechanics well enough to correct it. Same with scroll triggers, in-viewport animations, all of it. The work moved from "describe and hope" to "diagnose and specify."

Strange, good feeling. I don't really know how to code. But I've picked up the basics, the logic, how things connect, why an animation fires when it does. I've started thinking of myself as a kind of "full stack designer." Not a developer, but someone who understands enough of the tech to build, correct, and ship , not just draw.

Both sides are learning

People ask whether this pushes designers to think more technically, or whether AI is finally speaking a language designers already use, timing, easing, spatial logic , that we never had a direct way to execute. My honest answer: both.

We're learning the technical side. And the AI, in its own way, is getting better at design taste. The two are meeting in the middle, and the result isn't designers getting replaced. It's that we get room back.

Which is the part that actually excites me. This leaves far more space for the work that matters, creativity, interaction, polish. The craft. Instead of nudging 8px of padding in Figma for the hundredth time, I'm building something that moves and responds and can be tested by a real person the same day.

What I'm taking from this

The tools didn't make me less of a designer. If anything, they raised the bar on what "done" looks like. A polished, working, testable prototype in a fraction of the time isn't a shortcut, it's a better starting point for the real conversation with a client.

To any designer eyeing this shift: don't outsource your judgment to the prompt. The navbar didn't get fixed because I typed something clever. It got fixed because I understood the logic and then directed the tool precisely. That's still design work. It just ships faster now, and it ships something you can click.