The impact of generative AI in the workplace has understandably attracted as much hype as reality. This is perhaps compounded by the often narrow analyses of the impact the technology is having. For instance, a recent study from Stanford’s Erik Brynjolfsson looks at the productivity gains when workers use generative AI.

The study documents how customer service workers were paired with an AI assistant and became not only quicker at their jobs but also more productive. These gains were far from uniform, however, with lower-skilled workers generally benefiting far more than higher-skilled workers.

Reshaping work

For decades, computers have reshaped the economy by handling tasks with greater speed, accuracy, and reliability than humans. Yet, despite major advances, some workplace activities—writing emails, analyzing data, or making presentations—have been harder to automate.

Traditional computing works best with clear rules, making it well-suited for tasks like bookkeeping, data entry, and assembly-line work. As a result, automation has reduced demand for workers in these roles, pushing wages down. At the same time, it has increased the value of skills such as programming and data analysis, widening the pay gap between different types of workers.

Generative AI is now shifting the balance again. Researchers studied its impact in customer service by tracking 5,172 agents at a Fortune 500 software firm. They found that AI assistance boosted productivity, but its benefits were not evenly spread.

Greater productivity

On average, AI-assisted agents resolved 15% more customer issues per hour. The gains were even bigger for less experienced workers, who handled 30% more cases. AI also helped new employees learn faster—those with two months on the job performed as well as those with six. More experienced workers, however, saw little benefit, and in some cases, their work quality declined slightly.

AI also affected workplace behavior. Agents who followed AI’s recommendations closely tended to improve the most, and over time, they relied on the system more. The technology even improved employees’ English fluency, particularly among international workers.

Perhaps most surprising, AI seemed to change how customers behaved. Customer service work is stressful, with employees often dealing with angry or impatient callers. But AI-assisted agents faced fewer hostile interactions. Customers were less likely to question their competence or demand a supervisor.

Generative AI, then, is not just a tool for efficiency. It may also make some jobs less frustrating.