When the Bottom Rung Disappears: Training University Finance Leaders in the Age of AI

Jensen Huang put it bluntly: you won't lose your job to AI, you'll lose it to someone using AI. I agree with him, and the framing is important. The disruption is not really about a machine replacing a person. It is about a widening gap between professionals who have made these tools part of how they work and those who haven't. In higher education finance, that gap is about to collide with a structural problem we have been building toward for years.

For my entire career, the path into financial leadership ran through the bottom rung. You started as the staff accountant doing the reconciliations. You found the variance, traced it back to the transaction, and learned through hundreds of small repetitions how the numbers actually move and what they meant. That work produced an output, but its real value was the operational understanding and judgment it produced. By the time you were running a close, you understood how it worked, not just what it produced.

AI is very good at the bottom rung. The high-volume mechanical work that has always defined the entry tier such as reconciliations can be automated or accelerated now. In a profession short on hands, that is welcome relief. But the reconciliation that a machine now handles in seconds was also the thing that taught a twenty-four-year-old why the reconciliation is important. Remove the rung, and you have not solved the problem of how someone climbs the ladder.

This is the second cut, not the first

Here is what makes higher education different, and why I am more worried than a generic "AI changes work" piece would suggest. We in higher education have already cut this pipeline once.

I have written before about the loss of the analytical middle layer — the directors, analysts, IR staff, and budget professionals who turn strategy into executable work. That loss was not theoretical. Over the last 10-15 years, the student enrollment cliff forced institutions to moderate expenses, and for the universities that were paying attention, that work has already happened. The cuts came out of the middle. The prepared institutions are leaner now, and many of them are leaner precisely in the analytical tier that converts decisions into outcomes.

So the sequence matters. First, the enrollment cliff hollowed out the middle layer. Now, AI is disrupting the bottom rung that used to feed people into that same layer. We are looking at a profession that thinned its analytical capacity for budget reasons, and is now watching the organizational learning that replenished it get automated away. Two cuts, a decade apart, landing on the same part of the organization from opposite directions.

Where AI becomes part of the solution

And yet this is also where I find reason for optimism, because the same technology doing the disrupting may be the best tool we have to add back that lost analytical capacity.

If AI can absorb the mechanical work, it can also compress the learning curve that work used to provide. A first-year professional who no longer spends a year building schedules by hand can, with the right tools and the right expectations, be moved up the ladder faster — given structured exposure to why the schedule is built that way, the controls it supports, the judgment calls buried inside it. The organizational learning does not have to be preserved in its old form. It has to be rebuilt deliberately, and AI can assist in that rebuilding. Upskilling that used to take five years of repetition might take two if we are intentional about it.

I saw the early version of this last week. At JP Morgan's Higher Education Summit, the message on talent was direct: they are hiring fewer analyst positions because AI can do that work, and instead bringing on associates with two to five years of experience whom they then upskill. JP Morgan can do this. They have the tools, the training infrastructure, and the name brand to attract people who already have a few years of professional experience. JP Morgan will come out of this stronger.

The canary, and the part I have not resolved

But not every institution is JP Morgan, and that is the part I keep worrying about.

JP Morgan's move works because someone else invested in the ‘on-the-job’ training. They are hiring the finished product — people who got their bottom-rung learning somewhere else, before the rung disappeared. That is a perfectly rational decision for them. It is also a decision that only works if some other institution is still running the training ground.

For higher education, this is the canary in the coal mine. The universities with capacity will use AI to upskill and pull in experienced talent, the way JP Morgan does. The universities without that capacity — the ones already hollowed out by the first round of cuts — become the training ground: they bear the cost of forming people, and then lose them to the institutions that can pay for finished talent. AI rebuilds the analytical layer, but it rebuilds it unevenly. The gap between the institutions that can reach the solution and the ones that cannot may not close the problem at all. It may just relocate it.

I am optimistic about what AI can do for a short-staffed profession that already cut too deep once. For the first time, we have a tool that could rebuild analytical capacity faster than the old apprenticeship ever did. But optimism about the tool cannot make us careless about who gets to use it. The bottom rung was solving two problems at once: it got the work done, and it made the next generation of leaders. AI solves the first problem everywhere. It solves the second only where institutions have the capacity to put it to work — and in higher education, that capacity is exactly what the enrollment cliff already hollowed out.

The institutions that win the next decade will be the ones that adopt AI fastest and use it to rebuild the analytical layer twice over: the layer they cut for budget, and the apprenticeship they are about to lose to automation. The tools are available to everyone. The capacity to turn them into leaders is not, and that gap is the one I would be watching.