If you’re looking for the first real casualty of AI… don’t overthink it. It’s call centers.
Why?
Because the call center model is a perfect storm of inefficiency:
• High cost, low value-add: Average U.S. call center agent wage is ~$38k. Add training, turnover (30–45% annually), and management overhead… the true cost per seat is easily $60k+.
• Customer experience is broken: Long waits, scripted answers, inconsistent quality. Net Promoter Scores for many call center-heavy industries hover at embarrassing levels.
• AI’s edge case is good enough: You don’t need an LLM that passes the Turing test. You need one that can handle 80% of the routine “where’s my order?” and “I need to reschedule” calls. That’s already here.
The ROI math is brutal: if AI can eliminate just half of inbound volume, you’re looking at a 30–40% cost reduction in year one. Scale that across telecom, airlines, healthcare, insurance… you’re into tens of billions in efficiency gains.
Everyone talks about AI disrupting “knowledge workers” but the first big bloodbath will be in call centers. The jobs are repetitive, the unit economics are ugly, and the customer outcomes are poor.
The real question: who moves first?
Because once one competitor slashes their call center costs and reinvests in price or product… the rest will have no choice.