Return-to-office mandates grab headlines, but the quieter story is a structural squeeze on the middle. Openings for middle management have fallen, managers themselves are disengaging, and AI is moving from inbox novelty to org-chart knife. Employers in the United States posted 42 percent fewer middle-management roles at the end of 2024 than in spring 2022, a shift executives associate with a coming wave of AI-enabled flattening (Deloitte, 2025). Gartner forecasts that by the end of 2026, one in five organizations will use AI to flatten structure, and in those firms more than half of middle-management posts could go away. That is a forecast, not a fait accompli, but it frames the stakes clearly (Deloitte, 2025; Fortune, 2025).
The human signal is no healthier than the hiring data. Global employee engagement fell to 20 percent in 2025, with Gallup attributing about 9 percent of global GDP, roughly 10 trillion dollars, to lost productivity. Manager engagement slid hardest, dropping from 27 percent to 22 percent in a single year and down nine points since 2022. Yet the same report shows a design gap. In best-practice organizations, 79 percent of managers are engaged, almost four times the global average, a reminder that structure and role design still matter more than mood or macro cycle (Gallup, 2025).
Why the traditional middle is failing
Managers spend too little time on the one thing no algorithm can credibly do alone: grow people. Deloitte finds that managers devote nearly 40 percent of their week to firefighting and administrative work, and only 13 percent to developing their teams. More than a third report feeling unprepared for the people side of the job (Deloitte, 2025). If AI absorbs the reporting, scheduling, and status-herding that dominate the calendar, the value left for humans is coaching, context, and judgment. Many current role designs do not set managers up for that, which helps explain the engagement slide among managers Gallup documented in 2025.
Complexity compounds the problem. In a global survey of more than 2,500 leaders, McKinsey reported that two-thirds see their organizations as overly complex and inefficient, and only half feel ready to adapt to external shocks. Since the pandemic, about 90 percent of organizations have adopted some form of hybrid work, which adds coordination overhead to already dense structures if fundamentals are unclear (McKinsey, 2023). Middle managers often become the buffer that absorbs that friction. When the buffer is thin, work degrades. When the buffer is thick, costs and cycle times swell.
Proof you can scale with almost no middle
Self-management is not a fantasy. It is a design choice with prerequisites. The Dutch home-care provider Buurtzorg operates more than 900 self-managing teams of up to 12 nurses, serving roughly 14,700 employees as of January 2021. The company runs with two directors, about 50 back-office staff, and no middle managers. Instead, around 21 internal coaches support roughly 950 teams, about one coach for more than 45 teams, without formal authority or accountability for team productivity (Martela and Nandram, 2025). Analyses from the mid-2010s put Buurtzorg’s overhead at about 8 percent compared with a 25 percent sector average, but those are dated figures and should not be read as current. The same case notes that Buurtzorg’s billable hourly rate was higher than peers, €54.47 versus €48.74, an important reminder that overhead and total cost are not the same thing (Martela and Nandram, 2025).
The boundary conditions matter. The model performs best when interdependence between teams is low, frontline professionals are highly competent and motivated, and service quality can be verified directly by outcomes, all of which fit community nursing. Coaches support teams with reflection and conflict navigation, not directives. That is not a fit for every industry, but it is clear evidence that large-scale operations can run with minimal hierarchy under the right constraints (Martela and Nandram, 2025).
What the holacracy fight obscured
A decade of debate about holacracy produced more heat than light. The largest experiment at the time, Zappos, made self-management famous and polarizing. The smart read of that period was to avoid caricature, neither celebrating a bossless utopia nor dismissing the work as naïve. The substantive contribution was a balanced lens on when and how self-organizing structures might work, not a blanket endorsement of any one template (Bernstein et al., 2016). Zappos was, in 2016, the largest company to implement holacracy fully, but the record since shows that models evolve and that labels are less useful than the underlying mechanics of roles, decision rights, and accountability (Bernstein et al., 2016).
Redesign, do not defend, the role
The argument for keeping a thick middle has always been that someone must coordinate, translate strategy, grow talent, and maintain standards. That remains true. The question is whether you need a managerial layer designed for a paperwork era to do it.
Four practical moves follow from the evidence:
-
Strip out the administrative load at the source. If managers spend nearly 40 percent of their time on firefighting and admin, their calendars are the first design surface. Standardize routine processes, push approvals to clear rules, and apply AI where rules are stable. Gartner’s forecast that 20 percent of organizations will use AI to flatten structures by 2026 is not destiny, but it signals where productivity is likely to come from. The aim is not fewer humans as a goal, it is more human time on development and decisions that move the business (Deloitte, 2025).
-
Recast managers as coaches and product owners. Measure them by the health and output of their teams, not by span count or email velocity. Deloitte’s finding that only 13 percent of time goes to development sets a baseline. In best-practice settings, manager engagement is 79 percent. That gap is design, not fate (Deloitte, 2025; Gallup, 2025).
-
Where work allows, form small, self-managing units with explicit P&L or mission outcomes, and provide coaching without authority. Buurtzorg shows you can scale with minimal layers when teams can own the customer outcome end to end and when coaches are advisors, not bosses (Martela and Nandram, 2025). McKinsey highlights Haier as a beacon organization for self-evolving structures, another sign that alternative designs can perform at scale when roles and market interfaces are clear (McKinsey, 2023).
-
Be honest about fit. High interdependence, volatile interfaces, or low skill maturity argue for stronger coordination and clearer hierarchy. Buurtzorg’s boundary conditions are not optional. Ignore them and you recreate the worst parts of middle management, only with vaguer titles (Martela and Nandram, 2025).
The middle manager is not vanishing everywhere. The role is being unbundled. Some tasks will migrate to shared platforms and AI, some will move to the edge in autonomous teams, and what remains in the middle will look more like coaching and product stewardship. Hiring trends and forecasts point to a leaner layer in the near term, and the engagement data says design is the lever leaders control. The organizations that thrive will not defend the old middle. They will specify what judgment, development, and coordination they truly need, then build only that.
Sources
- Deloitte (2025). 2025 Global Human Capital Trends. Finds managers spend nearly 40% on admin and only 13% on developing people, reports a 42% drop in U.S. middle-management postings since 2022, and cites a Gartner forecast that 20% of organizations will use AI to flatten by 2026. https://www.deloitte.com/us/en/insights/topics/talent/human-capital-trends/2025.html
- Gallup (2025). State of the Global Workplace 2025. Reports global engagement at 20% with about $10 trillion lost productivity, and a sharp decline in manager engagement to 22% while best-practice organizations reach 79%. https://www.gallup.com/workplace/659610/state-global-workplace.aspx
- Martela and Nandram (2025). Buurtzorg: scaling up an organization with hundreds of self-managing teams but no middle managers. Documents 900+ self-managing teams, 14,700 employees (Jan 2021), two directors, 21 coaches without authority, dated overhead estimates, and higher hourly costs. https://link.springer.com/article/10.1007/s41469-024-00184-y
- McKinsey (2023). The State of Organizations 2023. Two-thirds of leaders see their organizations as overly complex and inefficient, only half feel ready for shocks, and hybrid work is widespread post-pandemic; highlights Haier as a beacon. https://www.mckinsey.com/capabilities/people-and-organizational-performance/our-insights/the-state-of-organizations-2023
- Fortune (2025). Surviving the great flattening: the coming extinction of the middle manager. Cites a Gartner forecast that by end-2026, 20% of organizations will use AI to flatten and could eliminate more than half of middle-management roles in those firms. https://fortune.com/2025/09/19/surviving-great-flattening-coming-extinction-of-middle-manager-layoffs/
- Bernstein, Bunch, Canner and Lee (2016). Beyond the Holacracy Hype. Urges a balanced assessment of self-management experiments and notes Zappos as the largest holacracy adopter at the time. https://hbr.org/2016/07/beyond-the-holacracy-hype