A few weeks ago, I opened a design file for the first time in three years.
Not a mockup review. Not feedback in a comment thread. An actual blank canvas, waiting for me to make something. I was building a marketing slide—an intranet announcement for a product our team had just rolled out—and I sat there for a second thinking, "okay, where do I even start."
Then my hands just... started. Layout first. Hierarchy second. Color last. Three years of not designing, and it came back like nothing had happened.
Riding a bike, except the bike learned some new tricks
Here's the thing nobody tells you about stepping away from a craft: the muscle memory doesn't disappear. But it also doesn't come back unchanged. I wasn't designing the way I did three years ago. I was designing with three years of practice leadership stacked on top of it—three years spent in sales conversations, coaching other product people, and figuring out how to scale product skills across a team instead of just my own work.
I knew why the layout mattered before I drew a single box. I knew which visual decisions would need to be defended in a stakeholder review, and I made those calls faster because I'd sat on the other side of that table. I wasn't just making something look good. I was making something that could ship, get explained, and hold up under scrutiny.
That's not something design school teaches you. That's what three years of leadership and twelve years of straight design teaches you.
What PM gave back to design
Product management didn't erase my design instincts. It sharpened what I pointed them at. I got better at asking "what problem is this actually solving" before I got attached to a solution. I got better at knowing when "good enough" is the right call and when it isn't. I got better at translating a design decision into a sentence an executive would actually read.
Going back to design didn't feel like starting over. It felt like getting handed the same tool with a better grip.
What AI actually changed for me
Here's where it gets more interesting. Lately I've been leaning on AI across every part of my job, not just design. I've used Claude to build out components like a timeline or export function, then handed the HTML straight to engineering. I've had it write a Python script to export a mapping exercise for an AI-powered testing suite. I've even brainstormed my way through a gnarly Bitbucket timeout issue with it, then walked engineering through the fix myself.
None of that replaced my judgment. It replaced the friction between having an idea and having something to show for it.
Design used to be the part of my job that took the longest to spin up. Now it's often the fastest. Which means the thing that actually slows me down isn't execution—it's knowing enough about design, product, and code to tell if what came back is any good.
Three years ago, stepping into a director role meant stepping away from hands-on design. A year ago, I started actually making the move into product management—something I'd been circling for a while. Turns out the real advantage was never picking one path. It was building enough range to move between both.
The bike didn't change. I just finally learned the terrain.
Ready to turn insights into impact?
Great AI systems come from teams that actually like building them together. If you're curious what that looks like in practice, come find out.