Verizon CEO: AI will take over 'a large percentage' of customer service | CX Dive

TEXT START: AI will replace "a large percentage" of Verizon customer service, CEO Dan Schulman said at Bloomberg Tech conference in San Francisco Thursday.

THE DISSECTION

This is a CEO accidentally publishing the death certificate of his own workforce while framing it as strategic differentiation. Schulman is doing what executives do: narrating a technology transition as if it were a product improvement cycle. The article treats this as normal business news while embedding the data point that should terrify every working person in America: AI agents are already scoring 1,280 basis points higher on customer satisfaction than the humans they replaced. That number is not a projection. That is a completed experiment.

THE CORE FALLACY

The text assumes this is a recomposition of labor — "complex stuff is a combination of an agent and a human working hand in hand" — when the actual trajectory is a successive elimination. The framework presented is that AI handles the "rote" work and humans handle the "complex" work. But this ignores the mechanical incentive structure: as AI capabilities improve (which they will, relentlessly), the boundary of "complex" shifts continuously toward the human. Ball even states this honestly: "the number of agents, human agents working with customers, is going to go down more over time." The article lets this sit there like it's a neutral observation rather than a catastrophe.

HIDDEN ASSUMPTIONS

1. "Disruption" is a soft word for structural collapse. The text treats job displacement as a side effect of product improvement, not a fundamental rupture in the wage-consumption circuit.
2. AI specialist roles will absorb the displaced. This is the canonical copium. Ball's framing — "people teaching AI how it can look up these simple things" — assumes a 1:1 or even 0.5:1 replacement ratio of human trainers per eliminated agent. The math does not support this. One human can train AI systems that replace dozens of humans. The jobs destroyed will vastly outnumber the jobs created. This is not a prediction; it is arithmetic.
3. "Brutally hard" remaining roles are a sufficient labor market. Ball acknowledges the remaining customer-facing roles will be "brutally hard." He implies these will be the human jobs that persist. But "brutally hard" and emotionally demanding work that survives AI does not automatically generate mass employment. It generates a residual boutique workforce.
4. Customer satisfaction is the only metric. The 1,280 basis point improvement is measured in satisfaction scores. But who captures the economic value of that satisfaction? The shareholders. The consumption circuit doesn't care if customers are satisfied — it cares whether the people displaced can still consume. They cannot.

SOCIAL FUNCTION

This article is transition management theater. It normalizes mass displacement by embedding it in a narrative about customer experience improvement, incremental progress, and new job categories. It is designed to make the reader nod along as if this is simply how business works, not as if it is the precise mechanism by which the post-WWII economic order severs the mass employment link. The Forrester quote about halving customer service jobs by 2030 is the one moment the article flirts with honesty, and it is immediately defused by the AI specialist framing.

THE VERDICT

This article is a case study in how the Discontinuity Thesis operates in plain sight and is reported as business news rather than civilizational diagnosis. Verizon has already run the experiment. The AI agents won. The remaining question is not whether the human customer service workforce will be gutted — it will be — but whether anyone in the political class is paying attention to the consumption implications of what Dan Schulman just announced at a Bloomberg conference to polite nodding heads.