How to Protect Your Leadership Pipeline From AI Cuts

And the executives cheering the loudest are usually the ones who should go first.

Let me be upfront about something before I start: I like AI. I am genuinely excited about what it can do. I am not one of those people sitting in the corner predicting doom while the world moves forward without them. AI is real, it is useful, and the companies that figure out how to use it well are going to run circles around the ones that don’t.

That said, I have seen too many poor strategies disguised by flashy technology to miss when disaster is brewing. Right now, many companies are choosing what looks good on paper, but those decisions will cost them everything in five years.

They are eliminating middle management and calling it a transformation.

Korn Ferry just released data showing that 82% of boards and CEOs plan to cut up to 20% of their workforces over the next three years due to AI. A big chunk of that is made up of middle management and entry-level roles. The pitch is efficiency. Flatter org. Faster decisions. Less overhead.

I agree with part of that pitch. I disagree with the assumption that companies can eliminate the layer where leadership is built and still expect to develop leaders.

The Fat Triangle Is a Real Problem

I want to be honest because this is not a defense of bloated org charts. If you have ever sat inside a company where every two teams has a manager, every two managers report to a senior manager, every two senior managers report to a director, every two directors report to a senior director, and two senior directors report to one VP, you know exactly how much of that structure is about coordination and how much of it is about politics, territory, and people who needed a title to feel important.

That structure is a tax. It slows decisions down. It creates layers of people whose primary job is to translate information from one level to the next without losing too much in the process, like a game of telephone where every round costs you six weeks and three rounds of revision.

So yes, strip that. Cut the fat. Push authority down. Build a flatter organization where the people closest to the work have the power to make decisions about the work. I am fully on board with that.

The problem is that most companies are not doing that. They are cutting people and calling it flat. Those are two completely different things.

A flat organization gives employees real authority and trust in their work. In contrast, many companies today build “thin” organizations with fewer people and fewer managers, yet the same workload. This is simply staff cuts disguised as strategy.

Here Is What Actually Gets Cut When You Cut Middle Management

Middle managers are annoying. I get it. I have been annoyed by plenty of them. Some of them exist purely to attend meetings about meetings and to protect their calendars as if they were national monuments.

Yet this layer, despite its flaws, is essential. It is where leadership is developed through real experience.

The first-time manager is not ready. Nobody is ready. You get handed a team, a new title, a pile of expectations, and approximately zero preparation for the fact that everything you were good at before just became less relevant. You have to figure out how to have hard conversations. How to coach someone who is struggling. How to give credit without losing authority. How to protect your team from organizational nonsense while still delivering results. How to push back up the chain when the direction is wrong.

That is not a course. You cannot learn it in a seminar. You learn it by doing it, failing at it, having someone pull you aside and tell you what you missed, and then getting another chance to do it better.

That is the pipeline. That mess of trial and error, of first-time managers figuring it out with a little bit of cover from someone who has already been through it, is where your next leaders come from.

Cut that layer, and you do not just save salary. You cut the apprenticeship. You skip the crucible. And then five years from now, you sit in a room with a VP role to fill, and you look around at your senior individual contributors who are brilliant at what they do and have never managed a single difficult conversation in their careers, and you wonder where all the leaders went.

They didn’t go anywhere. You just stopped making them.

The Real Villain in This Story Is Not AI

Here is the take that will make some people uncomfortable.

I am not particularly worried about AI replacing middle managers. I am worried about the CTO.

Not because CTOs are bad people. Some of them are excellent. But look at the pattern. The average tenure of a CTO in a large organization is 2-3 years. They come in, inherit a technology landscape built by someone else under constraints they did not have, immediately identify everything that is wrong with it, and launch a transformation.

New architecture. New platform. New tools. New vendors. Reorganized teams. Rebranded strategy with a name that sounds like it came out of a McKinsey deck.

Some of that is legitimate. Organizations do get stuck. Technology does get stale. Transformation is sometimes exactly what is needed.

But watch what happens. Two years in, the transformation is halfway done. The organization is disrupted. The teams that were building things are now rebuilding things in a new way. The old vendors are out. The new vendors are learning the business. And the CTO has a headhunter calling about a CEO role at a mid-size company, seeking someone who has driven transformation at a larger company.

They leave. The transformation is handed to someone who did not design it, does not fully believe in it, and now has to finish a race they did not sign up for. Or it gets quietly wound down and replaced with the next CTO’s version of the same idea.

Meanwhile, the actual builders, the engineers, the product people, the architects who have been there through three CTOs and three transformations, they are exhausted. They have seen this movie. They know how it ends. The best ones leave. The ones who stay learn to wait it out.

That is where I think AI changes the game, and where I am genuinely pro-replacement. The functions that a rotating, politically motivated executive class has traditionally owned, setting technology direction, evaluating vendors, assessing build versus buy, designing architecture, a well-governed AI layer informed by people who actually know the business, can do a better version of that work with more continuity and less disruption.

I am not saying fire the CTO. I am saying the argument for AI replacing leadership functions is much stronger at the top of the org than in the middle, and yet that is not the conversation anyone is having.

Balance Is Not a Buzzword. It Is the Actual Answer.

The companies that get this right are going to figure out something that sounds obvious but is apparently hard to execute: you can have a flat organization and a leadership pipeline at the same time, but only if you are intentional about it.

Flat doesn’t mean leaderless. It means fewer unnecessary layers and more distributed authority. You can achieve this without losing the managers who develop people.

What you cut is the fat: the layers that exist to manage layers, the directors who manage managers who manage people who do work. What you keep is the connective tissue: the experienced leaders who are actively coaching, developing, and building the next generation while also delivering results.

The difference sounds subtle. It is not. One produces a lean, capable organization that gets stronger over time. The other produces a lean organization that gets weaker every year because it keeps hiring senior leaders from outside, while its own people hit a ceiling they can never overcome.

I have seen both versions. The second one looks great in the short-term earnings call and reads terribly in the five-year talent review.

What This Means If You Are Leading Right Now

If you are a leader in an organization undergoing this kind of restructuring, you have a decision to make.

You can participate in the reduction theater: cut the headcount, hit the efficiency metric, take the credit, and be somewhere else when the pipeline runs dry. Plenty of people make that choice. It has a short-term return.

Or you can be the person in the room asking the question nobody wants to ask: Who develops the next layer of leaders if we cut the layer where leadership development happens? How do we build a flat org that still has a bench? What is the actual plan here, not the slide, the plan?

That question will not make you popular in the moment. It never does. But the organizations that come out of this AI transition with real capability rather than just reduced cost are going to be the ones where someone asked for it.

Be that someone.

The Bottom Line

AI is not the problem. AI is a tool. The problem is the same problem it has always been: organizations optimizing for what they can measure right now at the expense of what they need to build for later.

Cutting headcount has a line item. Building a leadership pipeline does not. So headcount gets cut, the pipeline gets ignored, and five years from now, the same organizations paying consultants to explain why they cannot find enough senior leaders will be paying consultants to explain that, too.

You do not build a great organization by removing the layer where great leaders are made. You build it by being honest about which layers add value and which just add politics, and then making the cuts in the right places, not the convenient ones.

That takes judgment. It takes leaders who are willing to protect the pipeline even when the spreadsheet says not to.

It turns out that is not something AI can do yet.