There are companies right now building themselves without middle management. Three roles: the ones who build, the ones who own a problem, and the ones who coach. That’s it. The coordination layer is gone.
And the premise holds.
Think about what actually goes on in a normal company today. Meetings where information is moved between people who already have access to the same information. Status updates summarizing things that are already in the system. Managers translating between layers. Coordinators keeping coordinators coordinated.
That is not strategy. It’s routing. And routing is exactly what AI gets good at first.
So why do most companies still look like their 2015 selves?
Is it technical debt? No, that’s just a cost.
Is it legacy systems? Part of it, but not a wall.
Is it the organization?
Now we’re getting closer.
If you were starting a company today, with the tools and models that exist, would you build four layers of management? Would you hire people whose primary function was moving information between other people? Would you hold weekly meetings where everyone tells someone what they’re working on when that someone could have read it?
You wouldn’t. Nobody would.
What blocks it is not the technology. It’s the people whose role consists of coordinating, and who therefore have a deep, personal interest in coordination remaining necessary.
It’s not malice. It’s structural inertia.
And it doesn’t just apply to PMs. It applies to project managers. Program managers. Scrum masters in the wrong sense of the word. Team leads who mostly run 1:1s and relay messages upward. The entire layer of people whose job, at its core, is making sure other people know what’s going on.
That work is the first thing AI takes. Not in five years. Now.
The paradox is that we already know it. Everyone has been in the meeting that could have been an email. Everyone has sat through the status round thinking “why am I here.” Everyone has written the report nobody reads.
We know it. And we do it anyway. Because someone has to run the meeting. Someone has to write the report. Someone has a job title that depends on it.
But that argument no longer holds. Not when the system can do it. And the system can do it.
The question is not whether companies get flatter. It’s who dares to do it first.
And what you do when it happens.
If your job is coordinating, you have a problem.
If your job is understanding something deeply, making decisions nobody else dares to make, or building something new, you are more relevant than ever. Because now it’s the only thing you do.
The company that starts over is coming. The only question is whether you’re one of the people building it, or one of the people standing in the way.