AI Change Leaders are not issued feelings…

Every serious AI-driven consolidation project starts the same way. Systems that multiplied for a decade get folded into one platform. Data that lived in several different formats gets reconciled into one source of truth. Manual handoffs that used to take days start happening in minutes. All of that is the easy part, or at least the part that shows up cleanly in a project plan.

The hard part shows up later, in the question nobody puts in the project plan: what happens to the people running the organization once the systems change?

I have run enough of these reviews now to recognize the pattern before I see it in the data. Somewhere in the organization, a section has accumulated more supervisory structure than its actual workload justifies: several people at the same level, spending more time coordinating each other than producing output. Somewhere else, a senior leadership position is about to be filled into an office that does not yet have the mid-level bench to support it. Neither situation is anyone’s fault. Both are what happens when growth is incremental and nobody is required to look at the whole system at once. AI-driven consolidation forces exactly that look, because the technology only works if the workflow underneath it makes sense.

Once I see it, I have to write it down. Once it is written down, it stops being an abstract inefficiency and starts being a set of specific recommendations about specific offices, staffed by specific people who did not do anything wrong and are about to lose something anyway: a position, a level, proximity to the leaders they have spent years building trust with, influence that took a long time to earn and will not take long to lose.

That is the cost side of the ledger, and it is real. I do not think leaders who wave it away are being honest, and I do not think leaders who let it stop them are doing their jobs.

There is a second effect that gets less attention, because it does not generate the same discomfort. The same review that costs some people access also tends to remove barriers that never had a good justification in the first place: layers of approval that existed to protect someone’s proximity to a senior leader rather than to add judgment to a decision, informal rules about who gets to brief the C-suite or the Staff that had more to do with tenure and turf than with who actually understood the problem. When those layers come out, capable mid-level leaders get direct access they should have had years earlier. That access is one of the actual points of the modernization, as real a return as any efficiency metric in the business case.

Both of these things are true in the same project, and often in the same conversation. A recommendation that costs one good person their position is frequently the same recommendation that gives several other good people a seat they were unfairly denied. There is no version of this work that delivers the second outcome without risking the first.

I do not think the discomfort ever fully goes away, and I am suspicious of anyone who tells me it should. Experience leaves the feeling untouched. It sharpens the standard instead.

The standard I have settled on is simple to state and hard to live inside: the organization and its mission come first, ahead of any individual’s position, including my own, and without turning the individuals affected into line items along the way. Living inside that standard means being honest in the brief about what a recommendation will cost and who it will cost. It means building in fair transition support wherever the process allows it. It means refusing to soften a finding just because the person it affects is someone I respect. The moment a recommendation starts protecting a relationship instead of serving the mission, the job has quietly become a different, smaller one.

AI can be very good at revealing how an organization actually functions, as opposed to how it looks on the org chart everyone agreed to years ago. What to do with that truth is still entirely on the people leading the change, and that does not get easier just because the technology got better.