A well-designed team includes less experienced people developing mastery.
If AI takes over the work that junior people used to do, where do senior people come from?
Let me tell you a story.
Years ago, I was hired by a very large company — you’d know the name — to document a mainframe system. This system ran their entire fulfillment and logistics operation. Every order, every shipment, every warehouse movement flowed through this one program. The problem? Nobody was left at the company fully understood it. The builders had retired. The people they’d trained had moved on. What remained was one guy keeping the lights on, and a growing dread about what would happen when something broke that nobody could fix.
They didn’t lose that knowledge overnight. The mainframe hummed along fine while the company stopped investing in the people who understood it. Then one day they realized: the last person who truly knows how this works is about to hand in their badge.
That’s exactly what’s about to happen with AI.
If AI handles the tasks that entry-level workers used to cut their teeth on, the research, the first drafts, the data cleanup, those folks never develop the judgment that comes from doing that work. They never build the pattern recognition. They never make the small mistakes that teach you how the whole system fits together.
Fast forward.
You need someone who can do the 15% that AI can’t touch. The innovation, the novel problems, the strategic judgment, the innovation. But your pipeline is dry. Nobody came up through the ranks because there were no ranks to come up through. You’ve got a mainframe problem all over again. Except this time it’s your entire knowledge workforce.
So what do you do?
You stop thinking about talent development as an individual career ladder and start thinking about it as a team design problem. A well-designed team includes less experienced people developing mastery, not through busywork, but by contributing to high-stakes work alongside experienced colleagues and AI tools that carry the load they can’t handle yet. The team operates at the expert level. The individuals inside it are growing toward it.
That’s not a training program. That’s an architecture decision.
And most organizations aren’t making it yet.