I've spent nearly two decades building digital products — at Adidas, Nike, Bowflex, and a handful of startups in between. I've shipped connected hardware, subscription apps, and platforms used by millions of people.
Recently, I decided to rebuild my personal website from scratch using AI. Not because I needed a website — I had one. But because I wanted to pressure-test something I've been thinking about: that working with AI isn't just a productivity hack. It's actually making me a sharper product leader.
Here's what I mean.
The Best Brief You'll Ever Write Is for an AI
When you work with an AI to build something real, you find out very quickly how precise your thinking actually is. Every vague requirement gets exposed. Every assumption gets tested.
I started this project by creating what I called a CLAUDE.md file — essentially a product brief that described everything about my site: the design philosophy, color palette, content structure, component hierarchy, and the rationale behind each decision. It's the kind of document I've written hundreds of times for engineering teams.
The difference? An AI doesn't nod along in a meeting and then interpret things its own way. It builds exactly what you describe. Which means if your description isn't clear, you see the result immediately. There's no three-week sprint buffer between a bad requirement and a bad outcome. The feedback loop is minutes.
That compression has changed how I think about writing requirements for human teams too. Tighter. More specific. More intentional.
Know When to Pivot — Even From Your Own Plan
Early in the process, I gathered reference sites I admired — dark premium aesthetics, bold typography, interactive hover effects, parallax depth. I analyzed what I liked about each one and worked with Claude to synthesize three distinct design concepts.
They were technically sound. And I didn't love any of them.
The old instinct would have been to keep iterating on concepts, burning time trying to make one of them work. Instead, I did what I'd tell any product team to do: I looked at what was already working. My existing Framer prototype had the right design. What it didn't have was the right foundation.
So we pivoted. Rebuild the Framer design in Next.js — same look, better infrastructure, room to grow. The AI wasn't the creative director. I was. The AI was the execution layer.
Knowing when the plan is wrong and having the discipline to change it — that's not a technical skill. That's product leadership.
Ship in Increments, Learn at Every Step
I built the site the way I'd ship any product: in phases, not all at once. Hero section first. Then the values section. Then the bio card. Each piece was a cycle of prompt, preview, tweak, commit, deploy. The full loop from idea to live on the internet took about two minutes.
Some things worked immediately. Some didn't — a scrolling animation that worked in one direction and completely broke in the other, layout choices that looked great on desktop and fell apart on mobile. Normal product development, just at a pace I've never experienced before.
What I found was that working this way — fast, iterative, with an AI that responds in seconds — trains you to make decisions quickly. You stop over-deliberating because the cost of trying something is nearly zero. Ship it, look at it, adjust. The same "Ship, Learn, Repeat" philosophy I've always believed in, but with the friction almost completely removed.
Articulating Problems Is the Actual Skill
Here's the thing that might surprise other product leaders: the hardest part of working with AI isn't the technology. It's the communication.
When I needed to redesign a section that felt too static, I couldn't just say "make it better." I had to articulate what wasn't working — "this vertical stack is hard to scan on desktop, I want side-by-side cards with hover interactions and staggered scroll animations." The more precisely I described the problem, the better the solution.
That's the muscle this builds. Not coding. Not prompt engineering. Problem articulation. The ability to look at something, diagnose what's off, and describe the desired outcome clearly enough that someone — or something — can execute on it.
That skill transfers to every conversation I have with designers, engineers, and stakeholders. It's made me more deliberate, more specific, and more effective.
This Is How Product Leaders Stay Cutting Edge
The gap between strategy and execution has never been smaller. Product leaders who understand how to leverage AI don't just move faster — they think differently. They prototype instead of just wireframing. They test ideas instead of just debating them. They ship instead of just planning.
My site is live. It's deployed on Vercel with automatic builds from GitHub. It has a blog, an about page, and a design system I can evolve. And the real value isn't the website — it's that I now have hands-on fluency with the same AI-assisted development workflow I'm bringing to my team at Matter.
This isn't about replacing developers. It's about product leaders getting closer to the build, understanding the craft more deeply, and making better decisions because of it.
The tools are here. The question is whether you're going to use them to lead differently — or keep doing things the old way and wonder why someone else shipped faster.