The Enterprise AI Boom Is Breaking Down the Call Center Model

In an interview on The Logan Bartlett Show in September 2025, Salesforce CEO Marc Benioff disclosed that the company's customer support workforce had shrunk from 9,000 to approximately 5,000 employees since the beginning of 2025.

AI agents now handle 50% of customer interactions. Support costs fell 17%. The business case was clean enough that Benioff described it matter-of-factly, not as a difficult organizational decision but as a natural consequence of deploying technology that works.

What he did not say, and what rarely gets said in these announcements, is where that work went before AI took it.

Most of it was handled by outsourced call centers, in the US, Philippines, and India, staffed by people whose entire job description was the structured, repetitive, high-volume customer service work that AI now handles for a fraction of the cost.

The AI customer service agent is not just replacing internal headcount. It is replacing the outsourcing model that enterprise customer service was built on.

The Economics That Made This Inevitable

A live agent call costs between $6 and $12 per interaction, according to Forrester Research. An automated AI interaction costs as little as $0.25. That is a cost differential of up to 48x.

For enterprises running millions of customer interactions per month, the math makes AI adoption a financial imperative rather than a strategic option.

Gartner had projected that conversational AI will reduce contact center agent labor costs by $80 billion globally in 2026. The savings come not from mass layoffs but from AI absorbing the 60% to 70% of inbound calls that follow structured, repetitive patterns. Exactly the work that outsourced call centers were built to handle at scale.

Klarna replaced the work of 700 customer service employees with AI agents. Cloudflare announced in May 2026 plans to lay off 1,100 workers, with CEO Matthew Prince describing the cuts not as cost reduction but as "defining how a world-class, high-growth company operates and creates value in the agentic AI era."

According to Programs.com's tracker, nearly 80,000 employees have been impacted by AI-driven layoffs in 2026 so far.

More than 60% of those cuts occurred at companies with over 100,000 employees, the same large enterprises that historically generated the outsourced call center volume that BPOs depend on.

The pattern is consistent across sectors. Enterprises are not cutting call center contracts and replacing them with nothing. They are cutting them and replacing them with AI agents that cost less, operate at all hours, and do not require the infrastructure of a third-party provider.

What the Data Shows, and What It Hides

Gartner's data introduces an important nuance that complicates a simple displacement narrative. A survey of more than 320 customer service and support leaders found that most US contact centers are pursuing workforce redesign over mass layoffs.

85% are expanding human agent responsibilities. Only 31% have implemented or plan AI-related layoffs through Q1 2027. "They are far more likely to be pursuing workforce redesign rather than role elimination," said Kathy Ross, VP Analyst at Gartner's customer service and support practice.

But the hiring freeze tells a different story. 50% of service leaders surveyed by Gartner said they have or plan to pause hiring within the next 18 months.

The workforce is not being cut dramatically today, it is being allowed to shrink through attrition while AI absorbs the volume growth that would previously have required new hires.

The outsourced call center that relied on enterprise contract renewals and volume growth to stay viable is discovering that neither is coming.

Anthropic's own labor market analysis, released in March 2026, found customer service representatives are among the most exposed roles in the AI economy. US Bureau of Labor Statistics data showed a 14% productivity increase in customer service in 2025.

A roughly 14% drop in job-finding rates for young workers entering highly exposed customer service occupations compared to 2022 levels suggests AI adoption is suppressing workforce entry before it produces visible layoffs.

The Offshore Consequence

The call center outsourcing industry that grew in India and the Philippines over the past two decades was built on the same repetitive, high-volume customer service work that AI is now automating inside US enterprises.

When a US enterprise deploys an AI agent to handle its customer service volume, it does not just reduce its own headcount. It reduces the outsourced volume it was sending offshore.

India's top IT firms added only 17 net employees in the first nine months of fiscal 2026, down from thousands of hires the year before, a near-total collapse in entry-level demand, according to Outsource Accelerator.

In the Philippines, a quality analyst was among more than 70 QA roles replaced by AI at his call center, after spending months helping train the systems that eventually displaced him. "I helped improve the work of the AI and now the AI replaced my job," he said.

Both countries added outsourcing jobs in absolute terms in 2025, over 80,000 in the Philippines and 120,000 in India. But those numbers reflect contracts that were in the pipeline before enterprise AI deployment accelerated. The pipeline is narrowing.

The Work That Remains

AI agents can handle 70 to 80% of customer queries today, with vendors targeting 90 to 95%. The remaining work including complex disputes, emotionally charged complaints, regulatory edge cases, relationship-driven conversations, still requires human judgment. That is where the role of human agents is evolving rather than disappearing.

The enterprises deploying AI are bifurcating customer service: AI for volume, humans for complexity. The question for outsourced call centers is which side of that divide their contracts were written for.

The outsourced call center was built to handle volume at low cost. AI handles the same volume at a fraction of the cost, with no contract, no headcount, and no offshore infrastructure. The business case for outsourcing that work no longer closes the way it used to.