🎧 Listen to this article

Narrated by Talon · The Noble House

In September 2025, Fiverr CEO Micha Kaufman announced a 30% workforce reduction — approximately 250 employees — in a letter that was explicit about the reason: the company was restructuring management layers to become "leaner, faster, and AI-first." Kaufman had warned his employees as early as June 2025: "AI is coming for your jobs. It's coming for my job too. This is a wake-up call." The interesting part of the September cut wasn't the layoffs. It was the target: fewer management layers, not fewer engineers or designers. The coordination class.

Workday eliminated roughly 1,750 jobs in February 2025 — about 8.5% of its workforce. CEO Carl Eschenbach cited AI investment explicitly in the restructuring announcement, per Reuters and Fortune coverage. Amazon executed multiple rounds of what internal communications called "efficiency restructuring" in 2025, targeting middle-management positions. The pattern across these organizations is specific: AI is not primarily replacing workers who make things or workers who decide things. It is replacing workers who coordinate things.

Why the Coordination Layer Existed

The modern large organization is fundamentally a coordination machine. Peter Drucker recognized this studying General Motors in the 1940s: the company's competitive advantage wasn't better cars. It was better coordination between engineering, manufacturing, distribution, and sales than its competitors could manage. That coordination required humans — layers of people whose primary function was translating strategy into tactics and tactics into execution, routing information between departments, managing handoffs between specialists.

Ronald Coase won the Nobel Prize in 1991 for explaining why firms exist: transaction costs. It's cheaper to coordinate work inside a firm than to negotiate market contracts for every task. The size of firms expanded as the complexity of their coordination tasks grew, and the coordination was handled by human managers.

AI just dropped the cost of information processing, translation, and routing — which is most of what coordination actually involves — close to zero.

Corporate hierarchy coordination layer middle management being automated by AI visualization
The Coasian firm exists to reduce transaction costs through internal coordination. When AI reduces the cost of information routing, translation, and synthesis, the economic rationale for multiple coordination layers weakens. The organizational structure follows the economics.

What Middle Managers Actually Do

The romanticized version of middle management is strategic insight and people leadership — coaching teams, making nuanced judgment calls, translating organizational vision into operational reality. This work is real and valuable and exists at every management layer.

The accurate description of how most middle management time is spent is different: synthesizing customer feedback from support tickets, sales calls, and surveys; translating that synthesis into product requirements; attending weekly status meetings and updating Jira; preparing quarterly business reviews from data that exists in dashboards; writing emails that route information between teams who don't talk to each other; scheduling and facilitating meetings that exist to resolve information asymmetries that could have been resolved asynchronously.

This is information processing. It requires judgment applied to domain knowledge. It also requires collecting information from multiple sources, synthesizing it into a summary, translating the summary into a different format, and routing it to the right recipients. Each of those sub-tasks is something AI does well and has been doing well since 2023.

What Replaces the Coordination Function

The coordination function doesn't disappear — it gets restructured. Three patterns are emerging:

Direct span expansion. A senior manager who previously oversaw eight reports because each required significant coordination overhead can now oversee twenty with AI-assisted synthesis of their output, automated routing of routine decisions, and AI-generated status tracking. The hierarchy flattens, not through elimination of management, but through each manager handling more direct reports at lower coordination cost.

Player-coaches. The emerging pattern in AI-native companies is management roles that are also individual contributors — people who ship work and coordinate a team, rather than people who only coordinate. This is possible when coordination overhead drops. The "pure manager" role that existed in large organizations because coordination was genuinely a full-time job at scale becomes a smaller proportion of the workforce when AI handles the mechanical parts of coordination.

Protocol-based coordination. Instead of human coordinators routing information, organizations design explicit protocols that AI systems execute: weekly status updates synthesized from all team members' input, automatic escalation of blocked items, AI-generated cross-team summaries. The coordination work becomes system design rather than daily execution. This requires upfront investment in protocol design but reduces ongoing human coordination cost significantly.

New organizational structure post-AI coordination layer flat hierarchy player-coach protocol visualization
Three organizational adaptations replacing traditional coordination layers: expanded span of control through AI-assisted synthesis, player-coach hybrid roles, and protocol-based automated coordination. Each trades upfront design cost for reduced ongoing coordination overhead.

The Human Part That Doesn't Compress

The Fiverr and Workday restructurings are specific about what they're cutting: coordination layers. They're not eliminating human judgment on consequential decisions. They're not eliminating people leadership — coaching, developing, evaluating performance. They're not eliminating the relational work that builds organizational trust and culture.

These are also the parts of management that are hardest to specify, hardest to automate, and most directly responsible for whether an organization's people are capable of executing strategy. Removing coordination overhead creates opportunity to do this work better. Whether organizations use that opportunity or simply reduce headcount without redesigning what management provides is the operational question that will separate good and bad outcomes from AI-driven restructuring.

Sources: Fiverr 30% layoff announcement, Micha Kaufman letter, September 2025 (IT Pro, Calcalist, Programs.com); Fiverr CEO "AI is coming for your jobs" email, June 2025 (Yahoo Finance); Workday 1,750-job reduction, February 2025 (Reuters, Fortune, Washington Post); Amazon efficiency restructuring, 2025 (CBS News); Ronald Coase, "The Nature of the Firm," 1937; Peter Drucker, Concept of the Corporation, 1946

Sources