TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A study of 5,172 customer-support agents found that an AI assistant increased issues resolved per hour by an average of 15%, offering a measurable example of AI entering service work. The result sits against a long economic trend: services have taken a growing share of US employment, while many service tasks have been difficult to make more productive. How broadly AI will change costs, jobs and service quality remains uncertain.
A study of 5,172 customer-support agents found that an AI assistant increased the number of issues resolved per hour by 15% on average, according to an article published by Thorsten Meyer AI on November 9, 2026. The finding offers a concrete example of AI being used in services, where productivity gains have often been harder to achieve than in manufacturing.
The study’s reported result is an average across the agents examined; the source does not provide the study’s authors, publication date, research design or details about how the AI assistant was used. The figure should not be read as evidence that every support team will see the same gain, or that the increase came from fewer workers or lower costs. The source frames the result as one example of AI’s potential in a sector that employs a large share of the workforce.
That sector’s growth predates the decline in factory employment. The article’s calculations from US Bureau of Labor Statistics payroll data put service-providing industries at 62% of nonfarm payroll jobs in 1939, 72% in 1979 and 86% in 2025. The measures exclude farm workers and the self-employed. Separately, the source cites International Labour Organization estimates that services accounted for 79.6% of all US workers in 2025, a broader workforce measure.
The economic explanation at the center of the article is Baumol’s cost disease: some work can be made more productive through technology and scale, while tasks centered on human attention or live performance are harder to speed up without changing the service. The article argues that AI is now being applied directly to some of these tasks, including customer support, but does not establish that the technology can remove the underlying limits across services.
Why So Much AI Ended Up in Services
AI is being applied to work that has long been difficult to make more productive. One customer-support study offers a measurable example, but its broader effects remain uncertain.
A long shift toward services
Service industries have taken a growing share of US nonfarm payroll jobs over time.
Why service productivity is hard
Baumol’s cost disease describes why some work is harder to speed up without changing the service itself.
- A string quartet still needs four musicians to perform a 45-minute piece.
- Some services depend on human attention, judgment or presence.
- Manufacturing has found ways to produce more with less labor per item.
- AI assistants can help support agents handle more issues per hour.
What the 15% result does—and does not—show
The reported measure is productivity, not proof of lower costs, better service or fewer jobs.
| Question | What the source says |
|---|---|
| More issues resolved? | ✓ Yes 15% more per hour on average |
| Lower prices or costs? | ~ Unknown Not established by the result |
| Better customer experience? | ~ Unknown Customer outcomes are not reported |
| Fewer jobs or changed pay? | ~ Unknown Workforce and wage effects are not reported |
Evidence still needed
The underlying study details are not provided, so the result’s reach and staying power are unclear.
- How the study measured productivity and what tasks were included.
- Whether the effect lasted over time and applies beyond customer support.
- What happened to service quality, customer satisfaction, workers and operating costs.
AI Meets Service-Sector Productivity
The reported 15% average increase matters because services account for most US employment, yet productivity improvements have not been evenly distributed across the economy. If AI can help workers handle more routine tasks without reducing the quality of service, businesses could gain capacity and customers could potentially benefit from faster responses. The cited result is a productivity measure, however, not proof of lower prices or better customer experiences.
There are also consequences for workers. An assistant that helps each agent resolve more issues could support higher volumes with existing staff, change the skills employers seek, or reduce the number of workers needed for some duties. The source provides no workforce or wage figures from the study, so it cannot show which outcome occurred. The broader question is whether AI makes service work more productive while preserving the human attention that customers value.
Products Worth Considering
Samsung Galaxy Tab A11+ 11-Inch Wi-Fi Tablet, 6GB RAM, 128GB Storage, Gray
Sony WH-1000XM6 Wireless Noise-Canceling Headphones, Black
Apple MacBook Air 13.6-inch Laptop with M5 Chip, 24GB Memory, 1TB SSD, Midnight
As an affiliate, we earn on qualifying purchases.
From Quartets to Customer Support
Economists William Baumol and William Bowen examined performing arts in their 1966 book, Performing Arts: The Economic Dilemma. A year later, Baumol developed a wider account of unbalanced productivity growth in an American Economic Review paper. The familiar example is a string quartet: a performance of a 45-minute Schubert piece still requires four musicians playing for that length of time. The work cannot be made faster without changing the performance itself.
By contrast, manufacturing has repeatedly found ways to produce more with less labor per item. The source compares this difference with official US price changes since 2000: quality-adjusted television prices fell 98%, while hospital services rose 274%, college tuition 188% and car repair 143%. These figures describe different categories and do not, on their own, explain every cause of price changes. Baumol’s framework is one way to understand why costs can rise in labor-intensive fields even when productivity advances elsewhere.
The source also cautions against explaining service costs only through inefficiency or pricing power. It points to haircuts and car repairs, which do not depend on health insurance or student loans, as services whose prices more than doubled over the period it cites. At the same time, US health spending was 18.0% of the economy, compared with 12.4% in Germany, according to the article. That gap indicates that national institutions and policies can shape costs alongside broad productivity pressures.
“An AI assistant raised the number of issues resolved per hour by 15% on average.”
— Thorsten Meyer AI article, summarizing the customer-support study
Limits of the Reported AI Gain
The source does not identify the customer-support study or provide its date, methodology, comparison group, task types or duration. It is not clear whether the 15% average reflects a particular company or setting, how performance was measured, or whether customers experienced improved service. The result cannot establish how AI affects support workers’ pay, employment or workload.
It is also unknown how widely the reported productivity gain applies beyond customer support. Some services involve judgment, physical presence or sustained personal relationships, and the article does not show that AI can increase output in those settings without affecting quality. The historical employment and price figures provide context, but they do not predict the pace or scale of AI adoption.
Evidence Needed on Wider Adoption
The next step is to examine the underlying study and determine how its result was measured, which support tasks were included, and whether the productivity effect persisted over time. Follow-up evidence on service quality, customer satisfaction, worker outcomes and operating costs would help clarify what the 15% figure means in practice.
For the wider economy, the central test will be whether AI can raise productivity across a range of services without weakening the human contribution those services depend on. The source describes that as an open question; it gives no forecast or announced timeline for when broader effects might become clear.
Key Questions
What is the reported AI productivity result?
The source says a study of 5,172 customer-support agents found that an AI assistant increased issues resolved per hour by an average of 15%. It does not provide enough study details to judge how broadly that result applies.
What does Baumol’s cost disease mean?
It describes the difficulty of increasing productivity in some labor-intensive services, such as live performance or personal care, compared with activities where technology can reduce labor per unit of output. When wages rise across the economy, costs in the harder-to-speed-up services can rise too.
How much of the US workforce is in services?
The source cites an International Labour Organization estimate that services represented 79.6% of all US workers in 2025. Its separate calculation from BLS payroll data puts service industries at 86% of nonfarm payroll jobs that year; the two figures use different measures.
Does the study show that AI will cut support jobs?
No. The reported outcome is issues resolved per hour, not job losses, staffing levels, wages or total costs. The source does not say what happened to employment among the agents studied.
Will AI lower the cost of services?
The reported productivity result does not establish that prices will fall. Whether businesses pass productivity gains on to customers depends on costs, competition, service quality and other factors not covered in the source.
Source: Thorsten Meyer AI
