Call center stocks are not being priced for today's earnings alone. Investors are marking them down for the moment AI makes a human-heavy service model look too expensive.
The selloff in customer-service outsourcing now has a current trigger, not just an old anxiety. Barron's reported on June 30, 2026 that Concentrix shares fell 24% in premarket trading after the company missed second-quarter expectations by a penny and gave weak guidance for the rest of the year. Teleperformance was dragged down 11.5% as well. When one company reports $2.46 billion in quarterly revenue, up 1.9% from a year earlier, and the market still punishes the whole sector, you should pay attention to what investors are really selling.
They are selling confidence in the old labor model.
The Klarna announcement in February 2024 was the moment the thesis became easy to understand. The Swedish fintech said its OpenAI-powered assistant had handled 2.3 million conversations in its first month and was doing the equivalent work of 700 full-time customer service agents. The Times reported that Klarna expected a $40 million profit improvement from the assistant that year. Le Monde later noted that Teleperformance shares fell 35% in the week after Klarna made the claim. No Teleperformance contract had to disappear for the market to get the message. A client had just put a number on how many human seats software could replace.
The short sellers have not gone away. The Financial Times reported this month that hedge funds are betting against the shares and debt of call center and customer experience outsourcing companies including Teleperformance, Concentrix and TTEC. Teleperformance had become one of the most shorted stocks in Europe, with short positions rising to 17.2%, according to the FT's account of Breakout Point data. Kasper Elmgreen, chief investment officer for fixed income and equities at Nordea Asset Management, called the sector a clean AI disruption case. That's blunt, and frankly, it's hard to argue with the setup: these companies sell human time at scale, and AI vendors are telling clients they can buy less of it.
Concentrix shows you how fast that repricing can arrive. The company bought Webhelp in a $4.8 billion deal announced in 2023, according to Reuters reporting at the time, expanding its exposure to the very customer operations investors are now questioning. Its latest quarter was not a disaster on the top line. Revenue still grew. But Barron's reported that third-quarter adjusted earnings guidance of $2.65 to $2.77 a share came in below the $3.08 analyst consensus, and full-year revenue guidance also disappointed. A business can still be large, profitable on an adjusted basis, and in trouble with the market if investors think its future contracts will be smaller, cheaper or more automated.
The money is not leaving customer service. It is changing sides. PolyAI, the University of Cambridge spinout founded by Nikola Mrkšić and others, makes voice AI systems that answer customer calls for businesses. The Times reported earlier this year that PolyAI raised $86 million in December at a $750 million valuation and works with clients including Whitbread and PG&E. That is the part incumbent BPO operators should find uncomfortable. Clients are not only asking existing vendors to add AI to the call floor. They are also funding and buying from companies built around removing the call floor from the center of the product.
Cresta is a different bet, and that may make it more dangerous for incumbents rather than less. The company sells AI that helps human agents respond faster and more accurately instead of replacing them outright. It raised $125 million in a Series D in 2024 with backers including Accenture, Andreessen Horowitz, Sequoia and J.P. Morgan. Accenture's presence matters because it already sits inside large enterprise transformation budgets. If AI copilots let one agent handle more work, the client still saves money. The BPO vendor still has to explain why the same old headcount pricing deserves the same old multiple.
There is a limit to the bear case. Complex claims, regulated financial support and multilingual technical troubleshooting are not all going autonomous in the next 18 months. Anyone who has been trapped in a bad phone bot knows the difference between answering a simple billing question and solving a messy customer problem. Teleperformance and Concentrix are not standing still either; both describe themselves as providers of AI-enabled customer experience rather than plain call center operators. That shift may work in some accounts. But shareholders are being asked to trust management teams to protect pricing while their clients are being shown credible ways to cut labor.
Here's the thing: investors do not need AI to finish the job before they reprice the stocks. They only need enough evidence that the next renewal cycle will be harder than the last one. Klarna supplied the simple example. Concentrix supplied the fresh earnings scare. The FT's short-interest data supplied the institutional verdict. If you run a labor-intensive service business, this is the sequence to watch. The multiple falls first. The revenue argument comes later.
Also read: The AI notetaker sitting in your Zoom call may be your next legal liability • SAP reorganizes its executive board around AI as investor patience runs thin • OKX bets the agentic economy needs its own payment rails before anyone else builds them
