Giving customer service agents AI help boosted resolved tickets per hour by 15% — but only for some agents.

This explores what AI does to the productivity of individual customer service workers, as opposed to economy-wide figures, and whether gains at the agent level carry through to the business. The clearest evidence comes from a study of 5,172 support agents at a Fortune 500 company. AI assistance raised issues resolved per hour by about 15% on average. The average hides the real story. Less experienced agents gained the most, getting both faster and better. The most experienced agents saw only small speed gains and a slight drop in quality Does AI assistance help less experienced workers most?. In practice, the AI seems to pass along what top performers already know to newer agents. For people who already know the job, it can get in the way.

The gain also depends on what the agent is doing. The productivity improvements in studies like this one were measured on tasks inside workers' existing skills. When people used AI to learn something new, the gains disappeared and their learning suffered When does AI actually boost worker productivity?. A related finding complicates the meaning of "faster." AI doesn't always shrink the time a task takes. It can move that time from doing the work to writing prompts and checking what the AI produced Does AI really save time, or just change how we spend it?. A large survey of active AI users found that almost 40% of the time saved went into fixing errors and verifying outputs Where does AI's time savings actually go in practice?. In support work, a quick but wrong answer creates a second ticket.

The weakest point is the handoff between AI and a human. A randomized field experiment on Alibaba's Taobao marketplace found that when a human agent took over after the AI failed, service quality recovered if the problem was technical. It did not recover if the customer was already frustrated. Timing was decisive: early handoffs worked and late ones didn't When human workers take over from AI, does it actually fix the problem?. The productivity of an AI-plus-human support team therefore depends heavily on when the AI is designed to step aside.

The agent-level gains often don't show up in the company's accounts. Klarna said its AI agent did the work of 853 employees and saved $60 million, yet documented customer service costs rose by about $8 million. That gap is a reason to treat vendor-reported savings with suspicion until someone audits them Do AI customer service agents actually reduce total support costs?. Two explanations from the broader corpus fit this pattern. AI speeds up the middle step of carrying out the work while deciding and delivering stay the same size or grow Does AI really compress all layers of knowledge work equally?. And gains at the level of single tasks only add up when organizations change how decisions are coordinated around them Does AI adoption require rethinking how organizations coordinate?. Executives also tend to see larger gains than the numbers show Do AI productivity gains feel larger than they actually measure?.

One takeaway is that customer service may be the clearest case of AI working as a training tool for new staff. The gains are largest where expertise is thinnest, which matters more for teams with high turnover than for teams of veterans. The open question in this corpus is what happens to that advantage once AI handles the routine tickets entirely and humans see only the hard, emotional cases that AI-assisted novices handle least well.