You Can't Buy the Judgment You Stopped Growing

This year the numbers got hard to ignore. Major firms are sharply cutting their MBA and entry-level hiring, some by as much as a third to nearly half. Across the industry, the work that used to fund the bottom of the pyramid, the analysis, the research, the first drafts, is being handed to AI. The headlines call it efficiency. I think it is something more serious, and most leadership teams have not named it yet.

Look closer at what is happening to entry-level work and you find a quiet contradiction. The same roles that are shrinking are also being seniorized. Employers now expect a new graduate to arrive with judgment that used to take years to build: the ability to frame a problem, interpret a model's output, question a result that looks wrong, and hold a client relationship. We are asking juniors to start senior. At the same time, we are removing the junior years that made senior judgment possible.

Here is the part the efficiency story misses. Judgment is the one thing AI cannot hand you. It is not a fact to be retrieved or a task to be automated. It is earned slowly, through repetition: the hundredth model you reviewed, the analysis that fell apart in a meeting, the client who taught you what the data did not say. Those reps lived inside exactly the repetitive, data-heavy work that AI is now absorbing. Strip out the grunt work and you also strip out the training ground.

So the question every leader should be asking is not how much AI can do. It is this: where does the next generation of senior judgment come from once the bottom of the pyramid is gone?

It helps to be honest about what clients were ever paying for. It was never the leverage of junior hours. That was the delivery mechanism, the way the work got done and the way young people learned. What clients paid for was judgment: the senior who knew which question mattered, which answer to trust, and who would stand behind the recommendation. AI changes the delivery mechanism. It does not change what is valuable. If anything, it makes judgment scarcer and more valuable, right at the moment we are dismantling the system that produced it.

This is a strategy problem, not an HR footnote. A firm that cuts its junior class to ride the AI efficiency wave is borrowing against its own future. It will look more profitable for a few years, then discover it has no one ready to become the senior who validates the AI, frames the problem, and signs the work. You cannot buy the judgment you stopped growing. By the time you need it, the people who would have developed it were never hired, or never given the reps.

The leaders who win the next decade will treat this as the real design question. They will use AI to remove drudgery, yes, but they will deliberately rebuild how judgment gets grown now that the old apprenticeship path is breaking. That might mean using AI itself as a teaching tool, putting juniors on harder problems sooner with a senior reviewing the reasoning rather than the typing. It might mean smaller, more intentional cohorts instead of none. It will certainly mean keeping an accountable human in the loop, not only because the work is safer that way, but because that loop is where the next expert is made.

I built my own firm on a version of this conviction: AI for scale, and a senior human accountable for every output. I did not do it only to protect quality. I did it because judgment is the asset, and an asset has to be cultivated on purpose. The technology is the easy part. Growing the people who can be trusted to wield it is the work.

The pyramid is not collapsing because judgment stopped mattering. It is collapsing because the old way of building judgment is breaking, and we have not yet built the new one. That is the task in front of every leadership team. Deploy the AI. But do not, in the same breath, cut off the way your future leaders are made. You cannot buy the judgment you stopped growing. You can only grow it, on purpose, starting now.