The Manager Deletion Experiment Is Failing — Here's What the Data Shows

The narrative was tidy: bloated management layers, hired recklessly in 2021–2022, needed to go. Meta, Google, and Amazon gave it a name — "flattening the org," "reducing bureaucracy" — and every mid-market company with a consultant on retainer followed suit. The problem is that the narrative confused overhead with infrastructure, and organizations are now paying the price on multiple fronts simultaneously.

O'Reilly Radar (2025-05-01) makes the structural argument plainly: managers are not a cost center sitting atop productive workers. They are the connective tissue that translates strategy into execution, surfaces information upward, and absorbs organizational ambiguity before it reaches individual contributors. When you cut 30% of your management layer to hit a headcount target, you don't remove bureaucracy — you redistribute its load onto people who have neither the authority nor the context to carry it.

The timing of this structural error is particularly bad. Organizations are simultaneously asking their remaining managers — whoever survived the cuts — to oversee AI adoption, manage the performance of employees using tools that didn't exist 18 months ago, and maintain culture across hybrid arrangements. According to HR Dive (2025-05-01), peak office utilization has now reached 80%, topping pre-pandemic levels — and CBRE credits managers specifically for driving those gains by doing a better job matching space with what employees actually need. In other words, the return-to-office wins that executives are celebrating were delivered by the very layer they spent two years thinning.

Now layer in the AI skills problem. HR Dive (2025-05-01) reports that CompTIA Vice President of Research Seth Robinson is explicit: companies that assume employees will bring AI skills into their jobs will set themselves up to fail. The skills gap persists even as personal AI use widens — which means there's growing distance between what employees do at home with AI tools and what they can do with them inside a governed, accountable work context. Bridging that gap requires someone with positional authority and operational context. That someone is a manager. But the manager is gone, stretched across twice as many direct reports, or — critically — untrained on the fundamentals of their own role.

That last point deserves more attention than it gets. Personnel Today (2025-05-01) reported that most managers receive no training on how to handle probationary periods — one of the most basic, legally consequential, and culturally defining functions a manager performs. If organizations aren't investing in manager competence for something that routine, they are certainly not investing in it for AI governance, skills development, or the ambiguous performance conversations that agentic tools are about to make far more frequent.

The entry-level pipeline makes this worse. HR Dive (2025-05-01) found that nearly half of organizations now ask a senior worker plus AI to do the work that several entry-level employees would have handled. This is not efficiency — it is the systematic removal of the developmental rungs that turn junior employees into senior ones. You cannot hire experienced managers in five years if you stopped producing them today. The org chart doesn't just reflect current capability; it shapes future supply.

The question executives should be asking is not "how many managers do we need?" but "what is the cost of the infrastructure we've dismantled?" Those costs are showing up in three places at once: slower AI adoption because there's no one accountable for capability-building at the team level; weakened return-to-office results in organizations that cut too deep; and a shrinking pipeline of people who know how to manage because they were never managed themselves.

MIT Sloan Management Review (2025-05-01) frames the leadership challenge in the AI era as fundamentally one of thinking — not tooling. The blind spot isn't about which platforms to deploy. It's about whether organizations have maintained the human infrastructure capable of making consequential decisions in ambiguous conditions. Deleting that infrastructure to meet a short-term headcount target was always a bet. The results are coming in.

The real question for your next board conversation: if managers are infrastructure, what's the replacement cost of the infrastructure you cut — and does anyone in your organization know how to calculate it?