AI and Contact Center Jobs: Checking Gartner's $80 Billion Prediction Against the 2026 Data

What the U.S. employment data actually shows

The U.S. Bureau of Labor Statistics doesn't track "conversational AI cost savings," but it does track customer service employment directly, and its most recent Occupational Outlook Handbook data gives the clearest independent read on the labor side of Gartner's prediction. As of 2025, about 2.67 million people work as customer service representatives in the U.S., earning a median of $44,770 a year. The BLS projects that number will decline 5% -- roughly 141,800 jobs -- from 2025 to 2035, and its stated reasoning names automation directly: "there is expected to be less demand for customer service representatives, especially in retail trade, as their tasks continue to be automated," with self-service systems, mobile apps and social media absorbing simple interactions that used to require a person.

That's a real, government-measured decline attributable in part to automation -- but it's a gradual erosion over a decade, not a cliff. The same BLS data projects 289,500 job openings a year on average through 2035, mostly from workers retiring or transferring to other occupations, meaning the customer service labor market keeps churning even as its overall size shrinks. Nothing in the BLS data isolates AI specifically as opposed to older forms of self-service automation, and nothing in it puts a dollar figure on cost savings -- which is a first indication that Gartner's $80 billion is not something government labor statistics were ever going to directly confirm or refute.

Source: U.S. Bureau of Labor Statistics, Occupational Outlook Handbook.

Klarna: the full arc, hype to reversal to hybrid

Klarna's AI customer service story is the most-cited case in the industry precisely because it doesn't fit a simple narrative in either direction. In February 2024, the buy-now-pay-later company said its OpenAI-powered assistant did the equivalent work of 700 full-time agents, handling 2.3 million chats -- roughly two-thirds of all customer service inquiries -- in its first month, with an average resolution time under two minutes.

By May 2025, the story reversed. CEO Sebastian Siemiatkowski told reporters that "cost unfortunately seems to have been a too predominant evaluation factor when organizing this," adding that the result was "lower quality," and Klarna began recruiting human agents again, running a pilot with flexible, remote positions.

By February 2026, Klarna had landed on a hybrid model rather than a full retreat: AI still handles routine, standardized queries, but human agents -- recruited directly from Klarna's own customer base in what the company calls an "Uber-style" flexible-shift model -- now serve as a premium tier. As Siemiatkowski put it: "If AI can do customer service, it means it's going to be the cheap customer service... the future of VIP experience will be the human connection." Even with the rehiring, Klarna's total customer service headcount fell from about 7,000 before the 2024 AI push to roughly 3,000 by February 2026, with the company projecting it could shrink below 2,000 by 2030.

Read together, Klarna's numbers show a real net headcount reduction of well over 50% -- consistent with the direction of Gartner's prediction -- but arrived at only after a public quality failure forced a correction, undercutting the idea that AI-driven cost-cutting in customer service has been a clean, straight line anywhere it's been tried at scale.

Sources: CX Dive, May 9, 2025; CX Dive, February 20, 2026.

So did the $80 billion happen?

Nobody has published a reconciled answer, and this report can't manufacture one that doesn't exist -- no analyst firm, Gartner included, appears to have gone back in 2026 and measured actual global contact center labor cost savings against the 2022 forecast. What can be said, based only on sources that can be checked directly:

• The mechanism Gartner described is real and documented. Salesforce and Klarna both made large, AI-attributed cuts to customer service headcount, on the record, from company leadership.
• The automation-rate piece of the prediction looks conservative in hindsight. Gartner predicted 1 in 10 agent interactions automated by 2026; Salesforce's own November 2025 State of Service report put AI's share of customer service cases at about 30% already, projected to hit 50% by 2027 -- well above Gartner's 2022 estimate, at least among the more AI-forward organizations Salesforce surveyed (see our contact center AI adoption report for the full survey breakdown).
• The macro employment data shows erosion, not a cliff. BLS projects a 5% U.S. decline in customer service jobs over a full decade, with the labor market still generating hundreds of thousands of openings a year from turnover -- a picture that's consistent with real cost pressure building gradually across an entire industry, not a discrete $80 billion event landing in a single calendar year.
• Where AI-driven cuts happened fastest, quality problems showed up too. Klarna's very public reversal is a reminder that headcount reductions attributed to AI aren't automatically a clean win; the company itself said cost had been overweighted relative to quality.

The honest summary: the direction of Gartner's 2022 prediction was right, and some individual companies plausibly hit or exceeded its implied savings rate on their own books. Whether the aggregate $80 billion figure is accurate for the industry as a whole in 2026 remains, as far as this research could verify, unaudited.