🔍 Read the full analysis: AI's Pivot: 80% Focus On Services on ThorstenMeyerAI.com
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
TL;DR
A November 9 report argues that the service sector’s rise in the United States began decades before manufacturing employment peaked, and explains rising service costs through Baumol’s cost disease. It cites a study of 5,172 customer support agents in which an AI assistant increased issues resolved per hour by 15% on average, but says the effect on jobs remains unclear.
An AI assistant increased issues resolved per hour by 15% on average in a study of 5,172 customer support agents, according to a November 9 report examining AI’s potential effects on service work. The report also says services accounted for 86% of U.S. nonfarm payroll jobs in 2025, and leaves open whether productivity gains will reduce staffing or shift workers into other tasks.
The figures in the report show that the shift toward services predates the peak in factory employment. Based on annual averages from the U.S. Bureau of Labor Statistics’ Current Employment Statistics, it calculates that service-providing industries accounted for 62% of nonfarm payroll jobs in 1939, 72% in 1979 and 86% in 2025. The calculation excludes farm workers and self-employed people. Separately, International Labour Organization estimates put services at 79.6% of all U.S. workers in 2025.
The report uses the economist William Baumol’s theory of “cost disease” to explain why services can become more expensive even as other industries grow more productive. In some work, technology and larger-scale production can reduce the labor needed for each unit of output. In other work, including a live performance or personal care, the labor itself is a central part of the service, limiting how much production can be sped up without changing what customers receive.
As examples of price changes, the report says official, quality-adjusted U.S. television prices fell 98% from 2000, while hospital services rose 274%, college tuition 188% and car repair 143%. These figures apply to the stated period and categories; they do not by themselves establish why each price changed. The report also notes that haircuts and car repairs rose despite not relying on health insurance or student loans, while U.S. health spending reached 18.0% of the economy compared with 12.4% in Germany.
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 service sector is not a small remainder of an economy built around factories: it represents most U.S. employment by the measures cited in the report. Changes in how hospitals, schools, retailers and support teams use technology may therefore affect a broad range of workers and customers, even if manufacturing remains important.
The reported 15% average increase in resolved support issues per hour offers a specific measure of AI’s potential to alter service workflows. It is evidence of a productivity result in one study, not proof that the same gains will occur in every service job or that companies will reduce headcount. Productivity can lower the amount of labor needed for a given task, but demand, service quality and the creation of other tasks also shape employment.
The cost-disease argument helps explain why productivity gains in manufacturing do not automatically make personal services cheaper. If wages rise across the economy while a task still needs roughly the same human time, its price can rise relative to goods that become easier to produce. AI could change that pattern in some services, but the extent depends on whether it can take over tasks without reducing quality or changing the service people expect.
Products Worth Considering
Bose QuietComfort Headphones (2nd Gen), Wireless Over-Ear Noise-Cancelling Headphones, Eucalyptus Green
Sony WH-1000XM4 Wireless Noise-Canceling Headphones, Silver
DJI Mic 3 (2 TX + 1 RX + Charging Case) Wireless Microphone
As an affiliate, we earn on qualifying purchases.
From Orchestras to AI Tools
Baumol and William Bowen examined the economics of performing arts in their 1966 book, Performing Arts: The Economic Dilemma. The report recalls their example of a 45-minute Schubert string quartet: four musicians still require about three hours of combined labor to perform it. The example illustrates a limit on productivity in work where the human performance is the product, rather than an input that can be reduced without changing the output.
In a 1967 paper, Baumol extended the idea beyond the arts, distinguishing work in which technology can steadily raise output per labor hour from work where labor remains central. He described the distinction as a matter of degree, not an absolute divide. Retail, for example, can use self-service, while some education and care work still depends on direct human attention.
The report places that theory alongside the timing of U.S. employment change. Manufacturing jobs peaked in June 1979 at almost 20 million, according to the source, but services had already grown from 62% of nonfarm payroll jobs in 1939 to 72% in 1979. The service share continued increasing afterward, reaching 86% in the report’s 2025 payroll calculation.
““Human ingenuity has devised ways to reduce the labor necessary to produce an automobile, but no one has yet succeeded in decreasing the human effort expended at a live performance of a 45-minute Schubert quartet much below a total of three man-hours.””
— William Baumol, quoted in the report from his 1967 paper
Jobs After Productivity Gains
The report does not identify the customer-support study’s authors, publication details, date, or research design in the supplied material, so readers cannot assess its methods or how broadly its result applies. The 15% figure is an average for the studied agents; it does not show that every worker or company will see the same outcome.
It is also unclear whether AI-driven productivity will lead to fewer service jobs, more output from existing teams, different responsibilities, or some combination. The source frames this as an open question and provides no employment forecast, company-level staffing data or evidence that service prices will fall as a result of AI.
The employment percentages use different measures. The BLS calculation covers nonfarm payroll jobs and excludes self-employed people, while the ILO estimate concerns all workers. They should not be treated as interchangeable readings of one identical population. The price comparisons also describe changes over time, but do not isolate the contribution of wages, policy, market power, technology or other causes.
Evidence to Watch in Service Work
The immediate question is whether productivity findings like the reported customer-support result are replicated across employers, tasks and service industries. Useful evidence would include study methods, task definitions, the comparison group and measures of service quality, alongside the number of hours worked and any changes in staffing.
Future employment and price data can show whether AI adoption coincides with changes in service-sector jobs or costs, but those trends alone would not establish that AI caused them. The report gives no specific date for a next study or policy milestone. For now, its central point is that AI is being applied to a large and diverse part of the economy, while the consequences for workers and customers remain unsettled.
Key Questions
What is the report’s main development?
It connects the long-term growth of U.S. service employment and rising service costs with AI tools entering service work, citing a customer-support study that found a 15% average increase in issues resolved per hour.
What does Baumol’s cost disease mean?
It describes how costs can rise in work where productivity is difficult to increase, even as wages rise across the economy. A live performance is one example: cutting the musicians’ time would change the service itself.
Does the 15% result mean AI will eliminate customer-support jobs?
No. The reported figure measures issues resolved per hour in a study of 5,172 agents. It does not establish whether employers will cut jobs, expand output, change tasks or combine those responses.
How many U.S. workers are in services?
The report cites an ILO estimate that services represented 79.6% of all U.S. workers in 2025. Its separate BLS-based calculation puts service-providing industries at 86% of nonfarm payroll jobs that year; the two measures cover different populations.
What remains unknown about AI’s effects?
The available material does not establish how widely the customer-support result applies, whether AI changes service quality, or how adoption will affect staffing, wages and prices across other service industries.
Source: ThorstenMeyerAI.com
