TL;DR
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
US services employment expanded for decades before manufacturing employment peaked in 1979, according to the source’s Bureau of Labor Statistics calculations. The article links that shift to Baumol’s cost disease and asks whether AI can raise productivity in work where human labor is a central part of the service.
A November 9 analysis by Thorsten Meyer argues that the long rise of service work in the United States began well before manufacturing employment reached its peak in 1979, and asks whether AI can change the productivity limits that help explain the shift. The article says services accounted for 79.6% of all US workers in 2025, using estimates from the International Labour Organization, while its calculations from Bureau of Labor Statistics payroll data put service industries at 86% of nonfarm payroll jobs.
The two percentages describe different measures, not a contradiction: the ILO estimate covers all workers, while the BLS payroll survey cited by Meyer excludes farm workers and the self-employed. On that payroll measure, services accounted for 62% in 1939, 67% in 1967, 72% in 1979 and 86% in 2025. The figures support the article’s central point that services were gaining share well before the manufacturing job peak.
The analysis uses economist William Baumol’s cost disease to explain why labor-intensive services can become more expensive relative to goods produced with rising productivity. Some work can be reorganized or mechanized so that fewer labor hours produce more output. In other work, the time or personal attention provided by a person is part of what customers value, limiting how much labor can be removed without changing the service.
Meyer points to US price changes since 2000 as an illustration: official quality-adjusted television prices fell 98%, while hospital services rose 274%, college tuition 188% and car repair 143%. These are figures reported in the source article; its excerpt does not provide the underlying price series or additional methodological details. It also reports that an AI assistant increased issues resolved per hour by an average of 15% in a study of 5,172 customer-support agents. That result concerns the study’s particular setting and does not by itself show how widely the gain applies.
How Did 80% of AI End Up in Services?
Services made up about four in five US workers in 2025. AI may help raise productivity in service work, but one study does not show that jobs will disappear or prices will fall.
Two measures, different populations
The 2025 estimates are not contradictory: they count workers differently.
| Measure | Service share | Population counted |
|---|---|---|
| ILO estimate | 79.6% | All workers |
| BLS payroll calculation | 86% | Nonfarm payroll jobs |
The service shift began earlier
Services gained share well before manufacturing employment reached its reported peak in 1979.
Why some services resist productivity gains
Baumol’s cost disease describes how labor-intensive work can become relatively more expensive when productivity rises faster elsewhere.
- Technology can help workers produce more in less time.
- AI may help staff handle more routine tasks.
- Personal time can be part of what customers value.
- Removing labor may change the service or its quality.
What the AI evidence does—and does not—show
The reported productivity result comes from a particular customer-support study, not a forecast for all services.
- The study covered 5,172 customer-support agents.
- The source does not provide enough detail to assess how widely the gain applies.
- It does not establish that AI reduced prices, cut jobs or improved service quality.
- Evidence across other service settings would help show what changes in practice.
AI Meets Labor-Intensive Services
The analysis matters because services account for most US employment, and their costs affect everyday expenses such as health care, education, repairs and personal care. If AI helps workers handle more tasks per hour, it could moderate some cost pressures or allow organizations to serve more people with existing staff. The cited customer-support study offers a specific example of a measured productivity gain, but it is not evidence that AI has already reduced service prices or displaced workers across the economy.
The larger question is how gains are distributed. AI might take over routine tasks while people continue to handle complex cases, or employers might use higher output to serve more customers. It could also change the quality or nature of the service. The source frames these as possibilities, not established outcomes. Its evidence does not establish that AI will eliminate service jobs or reverse the long-term expansion of services.
Products Worth Considering
Canon IVY 2 Mini Photo Printer Bundle, Blush Pink
Dell 34-inch Curved USB-C Monitor S3425DW, WQHD, 120Hz
Logitech MX Mechanical Wireless Illuminated Full-Size Keyboard, Clicky
As an affiliate, we earn on qualifying purchases.
The Service Shift Predates 1979
The article challenges a familiar explanation: that services grew mainly because factory jobs disappeared. Manufacturing employment in the United States peaked at nearly 20 million jobs in June 1979, according to the source, but its payroll figures show services had already risen from 62% of nonfarm jobs in 1939 to 72% in 1979. The service share then continued to increase, reaching 86% of nonfarm payroll employment in 2025 under the article’s calculation.
Baumol and economist William Bowen examined the economics of performing arts in their 1966 book, Performing Arts: The Economic Dilemma. Their example was a live performance of a 45-minute Schubert string quartet: four musicians still require three hours of combined labor to perform it. In a 1967 paper, Baumol developed the broader argument that productivity can rise quickly in some industries while remaining comparatively constrained in others. The source stresses that this is a matter of degree: technology can change some service work, but not every productivity improvement preserves the same experience.
How Far AI Gains Will Reach
The source does not provide enough detail to independently evaluate the customer-support study, including when it was conducted, how the AI assistant was deployed or whether the reported average applied equally across workers and tasks. A 15% increase in issues resolved per hour in that study should not be treated as a forecast for all service industries.
It is also unclear whether productivity improvements will lower prices, increase output, change service quality or reduce staffing needs. The source’s price figures describe changes since 2000 but the supplied material does not identify all series or explain how each measure was calculated. The employment shares also use distinct populations and data sources, so they should not be compared as though they were identical measures.
Evidence Needed Beyond Customer Support
The next test is whether results like the reported customer-support gain appear in other service settings and persist outside a study. Evidence on prices, output, service quality, worker responsibilities and employment would help show what AI changes in practice. The source does not identify a scheduled follow-up study or a specific policy decision tied to its analysis.
For now, the article presents AI as a potential new tool for work historically less amenable to productivity increases, not as proof that the service economy is about to shrink. The central question is whether AI can raise output while maintaining the human attention and quality that define many services—and who benefits if it does.
Key Questions
What does the article mean by saying about 80% of US workers are in services?
The source cites an ILO estimate of 79.6% of all US workers in 2025. Separately, its calculation from BLS payroll data says services made up 86% of nonfarm payroll jobs that year. Those figures use different worker populations.
Did services grow only after manufacturing jobs declined?
No. The source’s BLS payroll figures show services rising from 62% of nonfarm jobs in 1939 to 72% in 1979, when manufacturing employment reached its reported peak. The shift began before that peak and continued afterward.
What is Baumol’s cost disease?
It is an economic explanation for why labor-intensive activities can face rising costs relative to sectors where technology lets workers produce more in less time. In some services, personal attention or time is part of the product, limiting how much labor can be cut without changing what customers receive.
Does the 15% AI result mean service jobs will disappear?
No. The source reports an average 15% increase in issues resolved per hour in a study of 5,172 customer-support agents. That result does not establish that jobs will disappear, that all service workers will see similar gains, or that prices will fall.
Source: Thorsten Meyer AI
