TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A November 9, 2026, article by Thorsten Meyer says services account for about 79.6% of US employment in 2025, based on International Labour Organization estimates. It links the sector’s growth and rising costs to Baumol’s cost disease, then points to a study in which an AI assistant helped 5,172 customer-support agents resolve 15% more issues per hour on average. How that productivity gain affects workers, prices and service quality remains uncertain.
A November 9 report by Thorsten Meyer examines how AI might change service-sector productivity, a question made more consequential by estimates that services accounted for 79.6% of US employment in 2025. The article draws on economist William Baumol’s account of why some work, such as live performance and personal care, is difficult to make faster, and cites a customer-support study reporting a 15% average increase in issues resolved per hour with an AI assistant.
The employment figures show that the shift toward services began well before US manufacturing employment peaked in 1979. Meyer calculates from the Bureau of Labor Statistics’ Current Employment Statistics that services made up 62% of nonfarm payroll jobs in 1939, 67% in 1967, 72% in 1979 and 86% in 2025. Those figures cover payroll jobs outside farming, not every worker. Separately, the International Labour Organization’s estimates put services at 79.6% of all US workers in 2025.
The article connects the sector’s growth and costs to Baumol’s cost disease. In a 1967 paper, Baumol described a difference between work where technology can raise output per labour hour and work where human effort is part of the product. A factory may make more goods with fewer hours; a string quartet still needs four musicians to perform together. When wages rise across the economy, work with slower productivity growth can become more expensive by comparison.
For possible change, Meyer cites a study of 5,172 customer-support agents in which an AI assistant increased issues resolved per hour by 15% on average. The article also compares official US price changes from 2000: quality-adjusted television prices fell 98%, while hospital services rose 274%, college tuition 188% and car repair 143%. The figures illustrate different trends; they do not by themselves establish why each price changed.
80% In Services How AI Got There
Services accounted for about 79.6% of US employment in 2025. The report examines whether AI can lift productivity in work where human attention is part of the service.
Services grew over decades
BLS-based payroll estimates show services became the majority of nonfarm jobs well before manufacturing employment peaked.
Why some services cost more
Baumol’s cost disease describes why wages can rise even where productivity is harder to increase.
- Technology can help produce more goods with fewer labour hours.
- Factory output can rise without matching increases in work time.
- A live performance still requires musicians to perform together.
- AI can change tasks without removing the human work in every service.
Price changes point in different directions
Official US price changes from 2000 to 2025 illustrate contrasting trends, but do not establish their causes.
| Category | Price change |
|---|---|
| Quality-adjusted televisions | −98% |
| Hospital services | +274% |
| College tuition | +188% |
| Car repair | +143% |
What the AI result does—and does not—show
Higher hourly output
In the cited study, an AI assistant helped customer-support agents resolve 15% more issues per hour on average.
Broader effects
The result does not establish effects on prices, service quality, staffing, or other industries.
AI Meets the Costliest Service Work
The report’s central question is whether AI can improve productivity in jobs where human attention is a large part of what customers pay for. If tools help workers resolve more cases in the same time, businesses could handle demand with fewer labour hours, serve more customers, or change prices and staffing. Which outcome follows depends on employers’ decisions and the nature of the work.
The stakes extend beyond call centres. Services include healthcare, education, transport, retail and government, but an efficiency gain in one task does not automatically translate into lower costs or better service across an entire industry. The customer-support result is a specific measured outcome, not proof that AI can replace the human work involved in every service.
The historical price comparisons also caution against reducing rising service costs to one cause. Meyer notes that haircuts and car repairs became more expensive despite not relying on health insurers or student loans. At the same time, he contrasts US healthcare spending, reported as 18.0% of the economy, with Germany’s 12.4%. Those figures suggest that national institutions and spending choices also matter alongside productivity pressures.
Products Worth Considering
Bose QuietComfort Headphones (2nd Gen), Wireless Over-Ear Noise-Cancelling Headphones, Black
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Twilight Blue
Apple AirPods Pro (2nd Generation) Wireless Earbuds with USB‑C Charging, Up to 2X More Active Noise Cancelling, Transparency Mode, Adaptive
As an affiliate, we earn on qualifying purchases.
Services Grew Before Factory Decline
The report pushes back on the idea that services expanded only because manufacturing jobs disappeared. BLS payroll data cited by Meyer show service-providing industries already made up most nonfarm payroll employment in 1939, decades before US manufacturing employment reached nearly 20 million in June 1979. The service share continued rising after that peak. The measures are not interchangeable: the BLS series excludes farm workers and self-employed people, while the ILO estimate refers to all workers.
Baumol and William Bowen developed the underlying argument while studying the economics of performing arts. Their 1966 book, Performing Arts: The Economic Dilemma, used orchestras and theatres to show why some work resists the productivity gains found in manufacturing. Baumol formalized the idea in his 1967 paper, Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis. The report emphasizes that the boundary is a matter of degree: technology can change tasks in retail or teaching without eliminating the human work involved.
““When someone purchases an air conditioner he neither knows nor cares how much labor went into it.””
— William Baumol, as quoted in Meyer’s article from his 1967 paper
Productivity Gains, Unsettled Outcomes
The 15% average increase reported in the customer-support study does not settle whether AI will reduce service costs, improve customer experience or cut jobs. The material provided does not specify the study’s date, workplace setting, comparison period, or how the results varied among agents. It also does not establish whether the measured gain persists over time or applies to other industries.
The employment and price figures have limits as well. The BLS service share is a calculation from payroll employment data and excludes self-employed workers and farm workers; the ILO figure uses a different population measure. The price changes cited cover 2000 to 2025, but the article does not give a full account of the methods or factors behind each category’s increase or decline. It remains unclear how much of the growth in service costs is attributable to productivity, wages, policy, market structure or other causes.
Evidence Needed Beyond Call Centres
The next test is whether the customer-support productivity result can be reproduced in other settings and sustained as workers and employers gain experience with AI tools. Readers will need details about the study’s design, the tasks involved and its effects on service quality and staffing to judge how far the result can be generalized.
Meyer frames the broader question as whether AI will reduce the number of workers needed in services or shift people into the tasks and roles that remain. That answer is not established by the figures in the report. Tracking employment, prices, output and service quality over time will show whether AI changes the economics of service work or mainly changes how it is performed.
Key Questions
What is the main development in the report?
The November 9, 2026, report brings together figures showing services’ large share of US employment, Baumol’s explanation of slow productivity growth in some services, and a study reporting higher hourly output among customer-support agents using an AI assistant.
How many US workers were in services in 2025?
The report cites an International Labour Organization estimate of 79.6% of US workers in 2025. A separate BLS-based calculation puts services at 86% of nonfarm payroll jobs that year; that measure excludes self-employed people and farm workers.
What does Baumol’s cost disease mean?
It describes how costs can rise in activities where it is difficult to produce the same service with fewer labour hours, even as productivity and wages increase elsewhere. Live performance is one example; the report says the distinction varies by activity rather than being absolute.
What did the customer-support study find?
In a study involving 5,172 customer-support agents, an AI assistant was associated with an average 15% increase in issues resolved per hour, according to the report. The supplied material does not provide enough study details to judge how broadly that finding applies.
Does the report show that AI will eliminate service jobs?
No. The reported productivity result does not establish whether employers will reduce staffing, serve more customers, change prices or improve service. The employment effects remain uncertain.
Source: Thorsten Meyer AI
