The future of UI with AI

For years I’ve been a proponent of a UX-first approach to design.

The user experience matters immensely. What we’re making, its purpose, the metrics we must move, the way the design will impact our users—all of these factors are important.

The user interface (UI) is a complimentary—while requiring an entirely different mindset—piece to the UX.

It’s easy to get roped up into the UI and lose sight of the whole point of the project.

So I don’t.

Instead I focus heavily on the UX side of things. My iPad Mini always at the ready to sketch out a flow and figure out how the pieces fit. Then, in an ideal world, I’ll go straight to code and feel out the concept on an actual device, and not in a Figma prototype.

This approach has helped shape my career and I’m grateful for it.

And yet, focusing so heavily on UX has limited some of my ability to work on certain teams. It’s not that I can’t, or won’t. But rather, that I see the tradeoff as rarely worth it. I don’t want to spend a significant portion of my life cleaning up tiny design artifacts.

If you’re part of a startup pushing to survive, it matters far more to understand the impact of someone clicking a button and making the sale—than to spending inordinate amounts of time on the shape and definition of the button.

One of my favorite projects in recent years involved countless hours focused heavily on UI. I loved it. We made some of the best work of my career—myself and another designer. But the startup ran into issues with cash and had to let most of the team go. We made a beautiful UI, but missed the boat on moving the business metrics that mattered.

Now is that an argument for inconsistent design? No. Well crafted buttons in a design system can speed up time to deployment and increase sanity among the entire team. All those details are helpful, if they’re not a distraction.

Now, with the shift toward AI we’re close to a vision I first saw about eight years ago—it was a demo (I can’t remember which company was doing it) where a designer was sketching out UX elements on pencil and paper and scanning those in to turn the scribbles into code friendly components.

That vision is here. Almost.

The AI tools are getting quite good at this, where I could take my whole design process and shift much of the time to tinkering with concepts and taking them straight to code—frontend code at least.

And I see this as a good thing.

Being able to understand a problem, work with the team to narrow it down to a few features, sketch out designs, take then straight to code, then test live with users—This sounds amazing and I can’t wait for it.

This kind of stuff has me more stoked than every before and is reinvigorating me in my career. In a way it feels a justification of the direction I chose.

Now, If you’re a designer who has spent years of your life focused on the UI side of things—diving into the intricate details and caring deeply about every little part of the system—I encourage you not to despair. That time and care you put into understanding the emotional connections between the pieces and the users, how all the parts fit together, that stuff matters.

The eye you’ve developed can help you in ensuring that the output of AI—because surely that’s where we’re headed—can be improved and buttoned up.

Over the last few months I’ve adopted auto layout in Figma wholesale. I can’t build out a project without it. This almost sounds opposite of my UX first approach with sketching. That’s because sometimes you still have to feel things out and see how they fit together, and an organized Figma project still has its benefits—at least for now.