The Future Of Management May Be 3 Employees, 19 AI Agents And One Big Accountability Problem

TEXT START: Over the past several months, we have been interviewing CIOs and other senior executives about how AI is changing management.

THE DISSECTION

The article identifies a real first-stage phenomenon: AI is turning reporting, memory, monitoring, status aggregation and follow-through into persistent infrastructure. Management is being unbundled.

But it treats redistribution of managerial responsibilities as the central event. Under the Discontinuity Thesis, the central event is more severe: a smaller number of humans, controlling AI systems, produce more output with fewer economically necessary workers. “Three employees and 19 agents” is not a new equilibrium of human employment. It is a ratio of concentrated productive leverage.

THE CORE FALLACY

The article mistakes new supervisory functions for durable replacement employment. Directing agents, tuning workflows, reviewing outputs and assigning accountability are real during the transition. They are also largely control-layer labor created by immature systems, governance requirements and liability exposure.

As agents improve and workflows stabilize, the same force eliminating information-collection managers will compress agent managers. A human may remain legally accountable without performing much economically necessary labor. Accountability can be assigned; it cannot restore mass productive participation.

The article also assumes that freed capacity will become coaching, creativity, mentoring and service improvement. That is a managerial choice under institutional lag, not a structural guarantee. Competition can convert the same capacity into larger spans, thinner staffing, higher output or greater returns to AI owners.

P1, P2 and P3 remain untouched: cognitive automation dominates; institutions cannot preserve stable human-only economic domains at scale; and the majority lose access to economically necessary labor. Management redesign is one mechanism of that collapse, not an escape from it.

HIDDEN ASSUMPTIONS

• Productivity gains will be reinvested in richer human jobs rather than headcount reduction or intensified leverage.
• Every important workflow will require a permanent human manager.
• “Human in the loop” means substantive productive work rather than a sign-off and liability shell.
• Coaching, mentoring and strategic judgment will remain scarce human functions instead of becoming partially automated.
• Agent training will remain labor-intensive rather than becoming front-loaded, standardized or automated.
• Accountability structure is equivalent to employment structure.
• Workers can migrate into orchestration and control roles faster than other workers and AI systems compete those roles away.
• Large-enterprise examples generalize to the wider labor market.
• Existing managerial titles and hierarchies will survive long enough to absorb the transition.
• Measuring outcomes rather than outputs changes who owns the productive system. It does not. Ownership still determines who captures the gains.

SOCIAL FUNCTION

Primary classification: transition management and ideological anesthetic.

Secondary classifications: partial truth, prestige signaling and elite self-exoneration.

The article is operationally competent. It accurately describes the cockpit rearrangement: old information-heavy management layers weaken, while agent supervision, workflow governance and accountability controls emerge.

Its ideological function is to present displacement as thoughtful organizational redesign. It directs attention toward job bundles, spans of control and human oversight while leaving ownership of AI capital, distribution of gains and the destruction of mass bargaining power largely offstage.

The accountability problem is real. It is also useful camouflage. A human owner can remain attached to the decision while the productive labor beneath that decision is automated away.

THE VERDICT

This is a competent description of management being compressed, elevated and partially converted into infrastructure. It correctly sees that the old manager bundle is breaking apart and that accountability remains a human and legal bottleneck.

It mistakes that bottleneck for a labor-market rescue. The future implied by its own evidence is fewer humans controlling more agents, with a narrow class of Sovereigns and high-value Servitors above a widening population made economically redundant. The 3-to-19 example is not a blueprint for broad prosperity. It is an early warning of leverage concentration.

Management is not saved. It is being stripped of routine labor, loaded with temporary governance work, and then positioned for further compression as the systems mature. Post-WWII wage-consumption capitalism loses its human circuit as this pattern scales.