🔍 Read the full analysis: Why 80% Of AI Efforts Go To Services on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A November 9, 2026, article from ThorstenMeyerAI.com argues that the long rise of U.S. service employment is not simply a result of factory decline. It links the shift to Baumol’s cost disease, in which work that is hard to make more productive can become more expensive as wages rise elsewhere, and points to AI as a new tool aimed at some service tasks.
A November 9 analysis from ThorstenMeyerAI.com says about four in five U.S. workers were in services in 2025, and argues that this long-running shift cannot be explained only by shrinking factory employment. The piece connects the trend to Baumol’s cost disease—the tendency for services that are difficult to automate or speed up to become more expensive as wages rise elsewhere—and asks whether AI can change that pattern.
The analysis draws on two measures of service employment. Its calculation from Bureau of Labor Statistics payroll data says service-providing industries made up 86% of nonfarm payroll jobs in 2025. Separately, it cites International Labour Organization estimates that services accounted for 79.6% of all U.S. workers that year. The figures use different worker populations and should not be treated as interchangeable.
The historical series in the article shows services already represented 62% of nonfarm payroll jobs in 1939, rising to 67% in 1967 and 72% in 1979, the year U.S. manufacturing employment reached its cited peak. That timeline supports the article’s point that service employment was expanding well before the factory-job decline became a defining feature of the economy.
To explain why some services get more costly, the article revisits economist William Baumol’s 1967 theory. Manufacturing can often produce more per worker through machinery and process changes. Some services—such as live performance, personal care, teaching and medical attention—depend on human time, making similar productivity gains harder. The article says U.S. prices since 2000 illustrate the contrast: its cited official, quality-adjusted television price fell 98%, while hospital services rose 274%, college tuition 188% and car repair 143%. These are price changes over that period, not direct measures of the causes of inflation.
Infographic 02 · Baumol's cost disease
Why the workers moved to services
Baumol's mechanism in four steps, and what it looks like in US jobs.
- 1966Baumol and Bowen: the quartetA 45-minute Schubert quartet still takes about three hours of human work
- 1967Two kinds of workWhere labour is a tool, machines cut it again and again; where labour is the product, they can’t
- WagesPay rises everywhereTeachers, nurses and musicians must earn more too, or they leave
- ResultCosts rise, people moveIf we keep buying the same mix, more and more workers end up in the slow sector
Service industries, share of US nonfarm payroll jobs
US jobs, annual average 2025 (millions)
Payroll shares exclude farm workers and the self-employed (BLS establishment survey). Health spending: national health expenditure as a share of gross domestic product.
Sources: W. J. Baumol, “Macroeconomics of Unbalanced Growth”, American Economic Review 57(3), 1967, BLS Current Employment Statistics (own calculation of shares), CMS, National Health Expenditure, Table 1.
Powered by Thorsten Meyer AI
AI Meets Labor-Intensive Services
The argument matters because services now account for most U.S. employment, while many of the costs households encounter most directly—health care, education, repairs and personal services—are in sectors where labor remains central. If productivity gains are harder to achieve in those areas, their costs can rise even while manufactured goods become cheaper. Baumol’s account offers one explanation, but it does not establish that productivity constraints are the sole cause of rising prices.
The article highlights AI as a possible change in that pattern. It cites a study involving 5,172 customer-support agents in which an AI assistant increased issues resolved per hour by an average of 15%. That result suggests a productivity gain in a particular support setting; it does not show that AI will have the same effect across health care, education, arts or other service work. Nor does a higher number of resolved issues by itself show whether service quality, costs or employment will improve.
The broader question is how efficiency gains are distributed. AI could help workers complete more tasks, change the mix of jobs, or reduce demand for some kinds of labor. The source presents those possibilities as an open issue, not a settled forecast. For readers, the distinction matters: a tool that improves output per hour in one workplace does not automatically make the wider service economy cheaper.
Products Worth Considering
Apple MacBook Air 13.6-inch Laptop with M5 Chip, 24GB Memory, 1TB SSD, Midnight
EcoFlow DELTA 3 Portable Power Station, 1024Wh, 1800W
Samsung 49-Inch Odyssey OLED G9 (G91SD) Dual QHD Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
From Quartet to Cost Theory
The analysis uses a 45-minute Schubert string quartet as an example of work whose labor requirement is difficult to reduce without changing the performance. It traces the example to William Baumol and William Bowen’s 1966 book, Performing Arts: The Economic Dilemma, which examined the financial challenges facing performing arts organizations. A year later, Baumol developed a broader account of uneven productivity growth in an article in the American Economic Review.
In that account, some activities are technologically progressive: machinery, scale and innovation can raise output per labor hour. In others, the labor itself is a major part of what customers receive. The article stresses that this is a difference of degree, not an absolute division; retail, for example, has changed through supermarkets and self-service. Its central point is that productivity does not rise at the same pace in every sector.
The source places that idea alongside the longer employment shift. It notes that U.S. manufacturing employment peaked at nearly 20 million jobs in June 1979, while the service share had already been increasing for decades. Its payroll calculation excludes farm workers and self-employed people, a limitation that helps explain why its 86% figure differs from the ILO’s 79.6% estimate for all workers.
What AI Productivity Can Change
The source does not establish whether AI will substantially reduce costs across the service economy. The cited customer-support study reports a 15% average increase in issues resolved per hour among 5,172 agents, but the supplied material does not give the study’s publication details, duration, comparison method or effects on service quality, wages and staffing. The result should therefore be read as a reported finding in one setting, not a general forecast.
It is also unclear how much of service-sector price growth the cost-disease explanation accounts for. The source notes that haircuts and car repairs rose in price too, while comparing U.S. health spending at 18.0% of the economy with Germany’s 12.4%. Those figures point to differences in national spending, but the material does not provide the underlying year, definitions or a full explanation of the gap. The relative roles of wages, financing, regulation, market structure and other costs remain unsettled here.
Finally, the employment percentages reflect different data sources and definitions. The source identifies the payroll series as excluding farm workers and self-employed people, while the ILO figure covers all workers. The precise share depends on which population is counted.
Evidence to Watch in Service Work
The next step is to see whether AI gains reported in customer support carry into other service settings and persist over time. Relevant evidence would include output per worker, service quality, prices, wages and staffing, rather than task counts alone. The source frames the employment consequences as a question for its ongoing series, not as an announced outcome.
Readers should also watch for comparable, clearly defined labor and price data. Updated BLS and ILO figures can show whether the service share continues to rise, while sector-level research can test whether AI changes productivity in health care, education, repairs and other labor-intensive services. Until those results are available, the article’s case is best understood as an explanation of a long-term trend and a prompt for evaluating AI claims, not a prediction that service jobs or costs will fall.
Key Questions
What does the article mean by saying about 80% of U.S. workers are in services?
It cites an International Labour Organization estimate of 79.6% for 2025, covering all workers. The article separately calculates that services made up 86% of nonfarm payroll jobs, a different measure that excludes farm workers and self-employed people.
What is Baumol’s cost disease?
It describes how costs can rise in work that is difficult to make more productive when wages increase across the economy. In a live performance, for example, cutting the musicians’ time may change the product rather than simply make it more efficient.
Does the cited AI study prove that AI will cut service costs?
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 by itself establish lower prices, better quality or fewer jobs, and it may not apply to other industries.
Did services grow only after U.S. factory employment fell?
No. The source’s BLS-based calculation says services were already 62% of nonfarm payroll jobs in 1939, reaching 72% by 1979. The figures indicate that the shift toward services began well before the cited manufacturing employment peak.
Source: ThorstenMeyerAI.com
