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 essay by Thorsten Meyer argues that services made up about four in five U.S. workers in 2025 and connects that shift to Baumol’s cost disease. It cites a study in which an AI assistant increased customer-support issues resolved per hour by 15% on average, but says the effect on service jobs and costs remains uncertain.
Services accounted for 79.6% of U.S. workers in 2025, according to International Labour Organization estimates cited in a November 9 essay by Thorsten Meyer, which links the sector’s growth to long-running productivity differences and asks whether AI can change the economics of service work. The essay also cites a study of 5,172 customer-support agents in which an AI assistant raised issues resolved per hour by 15% on average.
Meyer’s central point is that the rise of services did not begin only after U.S. factory employment peaked. Bureau of Labor Statistics payroll data, as calculated in the essay, show service-providing industries made up 62% of nonfarm payroll jobs in 1939, 67% in 1967, 72% in 1979 and 86% in 2025. The payroll survey excludes farm workers and self-employed people, so those shares are not directly interchangeable with the ILO estimate covering all workers.
The essay uses Baumol’s cost disease to explain why some services can become more expensive even as productivity rises elsewhere. In work such as manufacturing, technology can reduce the labor required per unit. In a live performance, teaching or personal medical care, reducing labor may also change the service itself. If wages rise across the economy while productivity in these activities grows more slowly, their relative costs can increase.
For a price comparison, Meyer reports that the official, quality-adjusted U.S. price of televisions fell 98% between 2000 and 2025, while prices for hospital services rose 274%, college tuition 188% and car repair 143%. These figures describe different categories and should not be read as proof that one explanation accounts for every increase. The essay argues that labor costs are part of the picture, while acknowledging that health care and education have additional factors.
The Story Behind AI’s 80% in Services
Services accounted for about four in five U.S. workers in 2025. A customer-support study reports an AI productivity gain, but its wider effects on jobs and costs remain uncertain.
A long shift toward services
BLS payroll calculations show service-providing industries gaining share well before the 1979 manufacturing employment peak.
Why service costs can rise
Baumol’s cost disease links rising relative costs to work where productivity is harder to increase without changing the service.
- Manufacturing technology can reduce labor needed to produce goods.
- Productivity gains can lower the labor required for each unit.
- A live performance, teaching, or personal medical care still depends on human effort.
- Wages can rise while productivity grows more slowly, increasing relative costs.
A stark price contrast
Reported U.S. price changes from 2000 to 2025 differ widely across categories; they do not show that one cause explains every increase.
What the AI result does—and doesn’t—show
The reported customer-support finding is a measured example, not proof of economy-wide change.
More issues per hour
The essay reports a 15% average increase with an AI assistant among 5,172 agents.
Jobs and costs
The result alone does not establish effects on staffing, prices, service quality, or other industries.
Quality over time
Further research can test whether gains persist across tasks and account for workers’ and customers’ outcomes.
AI Meets Labor-Intensive Services
The question matters because services account for a large share of U.S. employment and include work people rely on, from customer support and retail to health care and education. If AI tools help workers complete more tasks in the same time, businesses may be able to increase output or handle demand with fewer labor hours. That could affect operating costs, service availability and the mix of jobs, although the cited productivity result alone does not establish any of those outcomes.
The customer-support study offers a measured example rather than evidence of economy-wide change: 5,172 agents resolved 15% more issues per hour on average with an AI assistant, according to the essay. It does not follow that all service jobs can achieve similar gains. Some work depends on physical presence, personal judgment or direct human interaction, and productivity measures may not capture quality or customer experience.
Products Worth Considering
Bose QuietComfort Wireless Noise-Cancelling Headphones, Twilight Blue
Bose QuietComfort Headphones (2nd Gen), Wireless Over-Ear Noise-Cancelling Headphones, Eucalyptus Green
DJI Mic 3 (2 TX + 1 RX + Charging Case) Wireless Microphone
As an affiliate, we earn on qualifying purchases.
From Quartet Economics to AI
Economists William Baumol and William Bowen examined the finances of the performing arts in their 1966 book, Performing Arts: The Economic Dilemma. Their example was a 45-minute Schubert string quartet: four musicians still require about three hours of human labor to perform it, while manufacturing has found ways to produce goods with fewer labor hours. Baumol developed the broader argument in a 1967 paper on uneven economic growth.
The theory describes a difference in degree, not a fixed division between jobs that can and cannot improve. Meyer notes that technology has changed some services, including retail and travel, but argues that productivity gains can be harder when labor is central to what customers are buying. The essay’s employment figures place the long-term shift toward services decades before the U.S. manufacturing employment peak of nearly 20 million jobs in June 1979.
““When someone purchases an air conditioner he neither knows nor cares how much labor went into it.””
— William Baumol, quoted from his 1967 paper
How Far AI Gains Will Reach
The material provided does not identify the customer-support study’s authors, publication venue, date, experimental design or the range of tasks covered. It reports an average productivity result but does not provide detail on variation among agents, changes in service quality, or whether the gains persisted over time. Those details are needed to judge how broadly the finding applies.
It is also unclear whether AI-driven productivity improvements will reduce service prices, increase output, change staffing levels or lead to different kinds of work. The employment shares and price comparisons describe past conditions; they do not establish what AI will do next. Meyer frames that labor-market outcome as an open question, not a settled forecast.
Evidence to Watch in Service Work
The next test is whether measured gains in specific settings persist outside a study and extend to other service industries. Readers should look for further research that identifies the tasks involved, measures output and quality together, and reports effects on workers and customers over time.
For the wider argument, future employment and price data will show whether the service sector’s share continues to grow and whether productivity gains change costs in labor-intensive activities. The essay does not announce a policy change or a new AI deployment; its immediate contribution is to connect a historical economic theory with a reported customer-support result and set out what remains unresolved.
Key Questions
What does “AI’s 80% in services” refer to?
It refers to the essay’s cited estimate that 79.6% of U.S. workers were in services in 2025, according to the International Labour Organization. The essay also gives a separate 86% figure from BLS payroll data for service-providing jobs among nonfarm payroll employment.
What is Baumol’s cost disease?
It is an economic explanation for why costs can rise in activities where labor productivity improves slowly. Wages may increase alongside those in more productive industries, even when the work still takes similar amounts of human time.
What did the customer-support study report?
The essay says a study of 5,172 customer-support agents found that an AI assistant increased issues resolved per hour by 15% on average. The source material does not provide enough study details to assess how broadly the result applies.
Does the essay show that AI will eliminate service jobs?
No. It presents job displacement and changes in the kinds of service work people do as unresolved possibilities. The reported productivity result does not, by itself, establish that employers will cut jobs or that AI can perform all service tasks.
Source: Thorsten Meyer AI
