For twenty years the chart was a relay race. The relay just stopped making sense.

I was in a staffing conversation not long ago, the kind where you map a project to a team. We did it the way everyone has always done it. One researcher to scope the problem. A couple of designers for the screens. A prototyper. Then it goes to engineering. Clean lanes, clean handoffs, boxes and arrows.

Halfway through, one of my designers stopped me. “Why are we still drawing it this way? I already built the thing we’re about to spec.”

He had. Over a weekend, with AI in the loop, he’d gone from a rough problem to a working, clickable prototype that answered half the research questions we were about to assign to someone else. The plan on the whiteboard described a process that his actual work had already skipped three steps of.

I have run design teams for twenty-five years. That was the moment the old org chart stopped describing reality for me. Not in theory. On a whiteboard, with a plan that was already wrong.

First, why the relay model was good

Let me defend the thing I’m about to argue against, because it was not stupid. It was the correct answer to a real constraint.

When execution was slow and expensive, specialization paid off. A researcher who did nothing but research got very good at it. A designer who lived in the tool shipped cleaner, faster. Handoffs added friction, but the friction was cheaper than the alternative of everyone being mediocre at everything. The relay chart was how you bought quality at scale when every stage took real time. For twenty years, it earned its place.

I built teams on that model. At one company I grew a practice from a handful of people into a layered org — directors, seniors, mid-levels, juniors, researchers — each lane staffed by someone who owned it. It worked because the lanes were real. Each stage was slow enough, and distinct enough, to be somebody’s whole job.

What broke

The relay assumes one thing: that execution takes long enough to be worth handing off. Collapse that assumption and the whole structure stops making sense.

That is exactly what happened. When a designer can go from problem to working prototype in an afternoon, the handoff is no longer a time-saver. It is a time cost. You spend more hours packaging the work for the next lane than it would take to carry it forward yourself. The baton-pass, the thing that used to buy you quality, now buys you delay.

The lanes don’t hold anymore. Designers prototype in code. Researchers ship copy. Engineers make design calls. Nobody agrees on the “correct” process, and there may not be one.

This is uncomfortable, and I want to be honest that it is not pure upside. Specialization produced depth, and some of that depth is genuinely at risk. But pretending the relay still works because it used to is not a strategy. It is nostalgia with an org chart.

The shape replacing it

The teams pulling ahead are not the biggest. They are the most adaptable. Fewer lanes, wider range, and people who build their own tools instead of waiting for a lane to open.

The new unit is a small group that can run the whole loop — sense the problem, shape a solution, make it real, and judge whether it is right — without passing a baton at every turn. The roles inside it blur on purpose. The generalist who can research and prototype in one motion. The design engineer who lives between the file and the codebase. The person whose actual job is taste: curating and killing, not producing. These are not junior roles. They are the hardest roles to hire for, because range plus judgment is rarer than depth in any single lane.

What to do about it if you lead

Stop org-charting by handoff stage. That instinct is the residue of the old constraint, and it will quietly recreate the relay inside whatever new structure you draw.

Build small teams that can run the whole loop. Protect the space for people to tinker, because the ones building their own tools are the ones discovering where the new lanes actually are. And change what you reward. If you still promote on volume of output, you are rewarding the thing that just became free. Reward judgment — the calls, the kills, the direction — because that is the part that did not get cheaper.

My designer was right to stop me at the whiteboard. The plan was describing a team that no longer existed. The job now is not to defend the chart. It is to see the shape that’s actually forming, and build for that.

The relay served its era. This is a different era. The org that wins it will not be the one with the most lanes. It will be the one that needed the fewest.

Joe Cahill has spent 25+ years leading design across enterprise and fintech. He writes about design, judgment, and leadership in the AI era.

Related Articles