I joined Juicebox back in November. For more than 10 years, I’d been trained around a pretty well-defined idea of how design should be done. And to be fair, it worked. It taught me the value of iteration, research, getting the interaction right before getting consumed by all the visual details.
But that idea of how design needed to happen was changing – and changing fast. At Juicebox, engineering was moving in days, not weeks. The product was constantly evolving. If design was going to keep up, the way I worked had to evolve with it.
That shift accelerated with AI. I stopped thinking about AI as another design tool and started treating it as a layer across my entire process. I could throw a half-formed idea at it and explore ten directions before committing to one. I could use it to challenge assumptions, think through edge cases, generate copy, prototype interactions, or take something that existed only in my head and make it tangible quickly. When an idea takes minutes instead of hours to explore, you become much more willing to throw it away. And when you’re willing to throw away more ideas, you usually end up somewhere better.
Today we are a team of 3 designers, and that mentality has become part of how we work together. We’re spending less time thinking about where design ends and engineering begins. Our designers are getting closer to the code, using AI to build and prototype ideas rather than only representing them on a canvas. Sometimes that still means Figma. Other times it means jumping directly into a working prototype.
The question isn’t “Where should I design this?” It’s “What’s the fastest way to understand whether this idea is good?” That changes the relationship designers have with engineering too. We’re no longer just handing over a picture of what something should become, we can participate much more directly in figuring out how it should actually work.
But speed only matters if the quality comes with it. One of the things we care deeply about at Juicebox is getting the details right before we ship. The last bit of polish isn’t something we want to put into a backlog and promise ourselves we’ll come back to later. It’s part of building the feature. AI gives us leverage on the parts of the process that used to consume a disproportionate amount of time, which means we can spend more of that time on judgment: Does this interaction actually feel good? Is this the clearest way to communicate something? What happens in the weird edge case? Is there a simpler solution? That final layer still requires taste, craft, and a willingness to obsess over details. If anything, moving faster has raised our expectations for what we ship.
AI is changing what we expect a designer to be capable of. Designing at Juicebox increasingly means being comfortable moving between thinking, designing, writing, prototyping, building, and polishing – sometimes all in the same afternoon. Figma isn’t going away, and craft certainly isn’t going away. But the canvas is getting bigger. AI has given us more leverage, more ownership, and a much shorter distance between an idea and a real product. The goal isn’t to use AI so we can do less design. It’s to use that leverage to explore more ideas, get closer to the product, sweat more of the details, and ultimately ship better work.