The organisational chart that served your grandfather’s corporation — and probably your own, for most of your career — is disappearing. Not slowly. Not tentatively. Whole tiers are being erased, people are being asked to absorb their manager’s responsibilities, and reporting lines that took decades to calcify are being drawn again with a lighter hand.
This isn’t a prediction. It’s already happening. Forty-one percent of major organisations have already removed at least one management layer in the past three years. Twenty percent more are planning to do so by using AI to augment the people they have. And the pace of delayering is three times what it was five years ago. Something structural has shifted.
The question is whether it’s strategy, or just a good year for cost-cutting dressed up as digital transformation.
Why the pace has quickened
The removal of middle management layers is not new. Every management consultant since the 1980s has preached the gospel of flattened hierarchies: faster decision-making, closer connection between strategy and delivery, fewer people extracting rent from the middle. The theory is sound. The practice has always been harder.
What’s changed is not the idea. It’s the tools. When you have AI that can coordinate information flow, surface risks, track execution, and alert you to problems without requiring a manager to attend five meetings and write a status report, the case for the layer beneath it suddenly weakens. If the machine can do some of what a middle manager does, and the frontline team can absorb some of what a middle manager does, and the technology can surface the rest — then why is the middle manager there?
Middle management was always partly a translation layer between humans. Now that we have machines that can translate, the business case for the human in the middle has fractured.
The irony, of course, is that most organisations aren’t approaching this as a strategic redesign. They’re approaching it as a cost-saving exercise that happens to be enabled by technology. And that distinction matters far more than the spreadsheet suggests.
What middle managers actually do
This is where most flattening projects hit a wall, though not before they’ve already broken several things. Middle managers on an org chart have a title and a budget. Middle managers in practice have four jobs that rarely make it into the job description.
First, they translate strategy into work. A board decision about entering a new market or shifting a product roadmap doesn’t automatically turn into a daily task for a team member. Someone has to interpret it, figure out what it means for this team, remove obstacles, and explain the why when the work doesn’t align with what people expected. That translation — which is partly interpretation, partly politics, and partly genuine leadership — is usually invisible until it’s gone.
Second, they hold organisational memory. They know why a decision was made three years ago. They know which relationships are strained and which are solid. They know the pattern in how the executive team tends to reverse course. When you remove them, you remove people who have watched the organisation long enough to notice patterns. You replace them with a flatter structure and a shorter institutional attention span.
Third, they develop people. A manager’s job is partly to make today’s work happen and partly to decide which of their team members will be capable of doing tomorrow’s work, and to give those people the chance to grow toward it. Remove the manager, increase the span of control, and that coaching function either becomes someone’s “spare” responsibility — and therefore neglected — or it vanishes entirely.
Fourth, they absorb complexity. When a decision creates a conflict, when a client relationship frays, when someone is struggling personally and it’s affecting their work, or when two teams need to coordinate on something ambiguous — that’s where a capable middle manager makes their living. They are the people who absorb the friction so it doesn’t jam up the frontline or the executive suite. Remove that function and the friction doesn’t disappear. It just gets pushed to one of those two places.
Why most flattening fails
The most common mistake is a very human one: organisations flatten their org chart without redesigning the work. They announce that three reporting layers are now two, the span of control has widened, and people can now “be more empowered” — as though empowerment were something that happened automatically when you had fewer people above you.
What actually happens is chaos dressed up as flexibility. The work that the middle manager was doing doesn’t get reassigned. It gets abandoned. And it doesn’t stay abandoned. It pools at the bottom — frontline people suddenly spending twenty percent of their time in meetings trying to clarify what they should be working on — and it pools at the top, where senior leaders spend more time arbitrating operational questions because the translation layer has vanished.
47% of organisations that flattened their structure without redesigning decision rights reported a measurable decline in execution speed within the first twelve months. The cost of the flattening, when you account for the chaos and rework, often exceeded the savings from the headcount reduction. Gartner, 2025
The technology story makes this worse, not better. When an organisation announces it’s flattening because AI will now do some of what managers did, there’s an implicit promise: the AI will handle the coordination, the escalation, the information synthesis. In practice, AI is good at some of these things — escalation, information synthesis — and useless at others. It cannot develop people. It cannot translate strategy in a way that accounts for culture and context. It cannot absorb the complexity of human relationships and politics.
What boards should actually ask
Before you flatten, before you reduce headcount, before you announce the new leaner and more agile future, ask three questions that are easy to overlook in the momentum of change.
First: have you redesigned the work? Not the org chart — the work. Have you actually mapped what each role requires, removed the work that doesn’t need to happen, consolidated the work that overlaps, and explicitly reassigned the work that remains? Or are you just hoping it will sort itself out? If you can’t answer that question with specifics, you’re not ready to flatten.
Second: have you upskilled the people who’ll absorb the responsibility? When you remove a manager, you’re asking their team and their boss to do more. Have those people been trained? Do they have the tools? Have you explicitly given them permission to say no to work that can’t fit into their existing capacity? Or are you betting that increased productivity from “doing more with less” is just a cultural mindset? If you can’t point to a specific development plan for the people above and below the missing layer, you’re building failure into the system.
Third: have you thought about what middle managers actually do versus what the org chart says they do? This is the question that reveals whether you’re doing strategy or cost-cutting. If you’ve actually looked at the calendar, the emails, the decisions, and you’ve decided that the function is redundant, then you have a case. If you’ve just looked at the org chart and decided that the layer is expensive, you’re about to find out how expensive the chaos is.
This is where a Chief People Officer earns their seat at the table. Flattening is not an HR project. It’s an organisational design project, and it has to lead. Most CPOs are brought in at the end to manage the layoffs, figure out the severance, and handle the morale problem that emerges six months after launch when people are drowning.
The CPO who matters is the one who sits with the board before the decision is made, maps the actual work, asks the hard questions, and either says “this won’t work the way you’ve designed it” or “yes, we can do this, and here’s how.” The technology is going to accelerate flattening — that’s a given. Whether it accelerates progress, or just accelerates problems, depends on whether you’re thinking like a strategist or just like someone with a cost target to meet.
Domi Alzapiedi is a Chief People Officer in banking, focused on the intersection of people strategy, organisational design, and commercial performance. She writes about the questions that keep leadership teams honest.