I lead design, and I still build. That’s the job now.

For most of my career, the industry had one story about design leadership. You start with your hands on the work. You draw, you prototype, you push pixels. Then, if you are good, you get promoted away from all of it. The reward for being great at the craft was permission to stop doing it.

I never bought that story. I have led design for 15 years, and I am still in Figma. I still write the copy. I still open the front-end code and change it. That is not a hobby I keep on the side. It is the center of how I lead.

For a long time, staying hands-on was a personal preference. You could argue it either way. Then AI arrived and turned a preference into an advantage. The gap between a leader who can build and one who only delegates is no longer small. It is the difference between shipping this afternoon and waiting two weeks.

The advice that aged badly

The classic career ladder told designers to trade making for managing. Move up, and your job becomes meetings, roadmaps, and reviews. You stop touching the thing and start talking about the thing. People called this growth.

I understand where it came from. A team needs someone to set direction, protect focus, and clear the path. That work is real and it matters. But somewhere along the way we decided that doing the work and leading the work were opposites. You had to pick one.

That was always a little wrong. Now it is dangerous. When the cost of building drops to almost nothing, the leader who can build moves at a speed that a leader who only writes tickets cannot match. The old advice tells you to give up the exact skill that just became your edge.

AI widened the gap between deciding and building

Here is the shift that changed everything for me. AI did not just make building faster. It widened the gap between people who can go from a decision to a working thing and people who can only describe the thing and hand it off.

Think about what a decision used to cost. You decide a flow should work a certain way. You write it up. You explain it in a meeting. You wait for it to enter a sprint. You review what came back. Days pass, sometimes weeks, between the idea and the proof. Most of that time is not thinking. It is waiting and translating.

Now I can have the idea and build a working version of it before lunch. Not a sketch. A real thing I can click. When deciding and building live in the same head and the same afternoon, the whole loop collapses. You learn faster because you see the result of your own call almost instantly.

That is the part the org chart misses. The value of a hands-on leader in the AI era is not that they save a junior person some work. It is that they close the distance between a good decision and a real, testable result. AI made that distance the most expensive thing in most companies. A leader who can build erases it.

I ran this experiment inside a multi-billion-dollar company

I did not learn this from a blog post. I learned it leading senior UX product design inside a multi-billion-dollar design org, running a real AI-first transformation across B2B and B2C eCommerce. That is a big, careful environment. The kind of place where the safe move is to manage from a distance and let the process carry you.

I went the other way. I built an AI agent for a workflow that used to eat the team alive. It was a recurring task that took roughly 320 hours every time it came around. After the agent, it took about one hour. Same output, better consistency, a tiny sliver of the time.

I want to be clear about what made that possible. It was not just access to AI. Plenty of people have access. It was that I understood the workflow deeply enough to redesign it, and I was willing to build the thing myself instead of writing a brief and hoping it survived three handoffs. The craft and the leadership were the same act.

Then I built a whole company to prove it

After the agent, I wanted to know what happens when you build a business this way from day one. So I founded Story Genie. It is an AI-native company that makes personalized hardcover children’s books. Agents write the story and illustrate it. Humans supervise the quality and approve the work before anything reaches a customer.

The customer experience is the part I am proudest of. You see the entire finished book before you pay a cent. The whole thing builds in about 60 seconds. No deposit on a maybe. No waiting to find out if it turned out well. You look at the real result, then you decide.

I did not direct that from a whiteboard. I built it end to end. The flow, the copy, the front-end, the way the agents hand off to human review. I made the product decisions and then I made the product. When the preview felt slow or the copy felt off, I did not file a request. I fixed it that night.

If you want the longer version of how it came together, I wrote up the full Story Genie case study with the decisions and the tradeoffs behind it. The short version is this. A small team that builds can outrun a large team that only delegates, and the gap is getting wider every month.

The craft keeps the leadership honest

Here is the quieter reason I stay hands-on. It keeps me honest. When you build the thing, you cannot hide behind a slide. You feel exactly how hard it is, where it breaks, and which of your clever ideas fall apart on contact with reality.

A leader who has not touched the work in years starts making confident calls about things they no longer understand. They underestimate the hard parts and overestimate the easy ones. They ask for the impossible and wave off the things that actually matter. Not out of malice. Out of distance.

Staying close to the craft is a feedback loop on my own judgment. Every time I build, I am checking my instincts against what is real. That is what makes the strategy any good. Direction set by someone who knows the cost of each decision is worth far more than direction set by someone guessing.

This is not a story about heroics

I want to head off a misread. None of this means a leader should do all the work alone or refuse to delegate. That is just a bottleneck with a fancy title. Building yourself does not replace a team. It sharpens how you lead one.

When I sit down and build a piece of it, I learn the real shape of the problem. Then I can brief the team with precision instead of vibes. I know what is hard, so I protect their time on the right things. I know what is easy, so I do not turn a one-hour task into a two-week saga. The hands-on work makes the delegation smarter, not unnecessary.

I have led plenty of work that I did not build line by line. Kinjo, with two iOS learning apps and a design system. Supply Drop and its predictive subscription commerce. The Shopify rebuild for The Mysterious Package Company. Client work at Surf Studios, and my time at Enverus. On all of it, the closer I stayed to the actual making, the better my calls got.

What this means if you lead design now

I think a lot of design leaders feel a pull right now and are not sure what to do with it. They sense that the old ladder is cracking. They want to be near the work again but worry it looks like a step down. It is not. It is the opposite. The people who will lead well in this era are the designers who build, and they are about to have an unfair advantage.

So my advice is simple. Do not let anyone promote you out of your edge. Keep your hands on the keys. Learn the AI tools the way you once learned your design tools, deeply and daily, until building a working version of your idea feels as natural as sketching it. The leaders who do this will set direction at a speed nobody else can match.

This is also why I love working with founders and teams in a hands-on way, not from a distance. It is the same belief in practice. The most useful thing I can offer is not a deck about what you should build. It is sitting down and building enough of it with you that the direction becomes obvious. That is the heart of how I think about fractional design leadership.

The job changed. Deciding and building used to be two separate roles you handed between people. AI put them back in the same chair. The leaders who notice that, and act on it, are going to leave the ones who waited far behind.

How I actually keep my hands in the work

People assume that if a leader is still building, they must be hovering over their team or hoarding the fun work. That is the fast way to become a bottleneck. The trick is being deliberate about which work I touch and which work I hand off. I do not build to feel useful. I build to learn something the team cannot learn any other way, then I get out of the way.

My rule is simple. I build the riskiest unknown first, and I delegate the known. If we have never tried a flow before and nobody can tell me whether it will feel right or even work, that is mine for an afternoon. If the path is clear and the team can execute it well, that is theirs. I am not the best person to build the tenth version of a pattern we already understand. I am the right person to find out whether version one is worth building at all.

At Story Genie this shows up constantly. When we wanted to show parents a full book before they paid, I did not write a brief and wait. I went into Figma and the front-end and built a rough version of the preview myself. About sixty seconds, real pages, real feel. Once I knew it worked and knew how it should feel in your hands, I could hand the team a working thing instead of a paragraph. That is a much better starting point than a wish.

Time-boxing is what keeps this from eating my week. I give myself an afternoon, sometimes a morning. The goal is never a finished feature. The goal is to kill the biggest question. Does this interaction feel good. Can the agent actually do this. Is this idea as strong as it sounded in the meeting. When the box runs out, I stop, and what I learned goes to the team. A rough thing that answers the question beats a polished thing that took a week and proved nothing.

Building the riskiest part first de-risks everyone. The cost of a bad assumption is highest at the start, because the whole team builds on top of it. When I prove or break the core idea in an afternoon, my designers and engineers are not pouring days into something that was never going to hold. They inherit a tested foundation, not a hopeful one. That is the real reason I stay hands-on. It is not about craft for its own sake. It is about spending the team’s time on the right things.

The line I will not cross is taking ownership away from my designers. My rough build is a question, not a spec. I hand it over loose on purpose, and I tell them plainly that I expect them to make it better, change it, or throw it out. They own the real thing. I just go first into the scary part so they do not have to start from a blank page and a guess. When that handoff is clear, building myself makes the team faster instead of stepping on their work.

Leadership is not the reward for leaving the craft. In the AI era, the craft is what makes the leadership worth following.

I still build because it makes everything else I do sharper. The strategy, the team, the product, all of it gets better when I have my hands in the work. That is not me clinging to my old job. That is me doing the new one.