TL;DR
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
The source article reports that services accounted for 79.6% of U.S. employment in 2025, while service industries made up 86% of nonfarm payroll jobs. It links the long-term shift to differences in productivity and wages, and cites a study in which an AI assistant increased customer-support issues resolved per hour by 15% on average. The material does not establish that AI will reduce service employment.
Services accounted for an estimated 79.6% of U.S. employment in 2025, according to International Labour Organization estimates cited in a November 9 article by Thorsten Meyer AI. The report links the sector’s long-term growth to productivity differences across the economy and asks whether artificial intelligence can raise productivity in service work—without claiming that it will eliminate service jobs.
The article gives a second measure from the U.S. Bureau of Labor Statistics: service-providing industries made up 86% of nonfarm payroll jobs in 2025. That figure is based on the report’s calculation from annual averages in the Current Employment Statistics survey, which excludes farm workers and self-employed people. The share was 62% in 1939, 67% in 1967 and 72% in 1979, the year U.S. manufacturing employment reached its reported peak of almost 20 million.
The report argues that the shift toward services began well before manufacturing employment peaked and continued afterward. It uses Baumol’s cost disease to explain why some service costs can rise even as other industries become much more productive. In activities such as manufacturing, technology can increase output per hour; in work centered on personal attention or live performance, adding speed or reducing labor may change the service itself.
As evidence of a possible change, the article cites a study of 5,172 customer-support agents in which an AI assistant raised the number of issues resolved per hour by 15% on average. The source does not provide the study’s full citation or enough detail to assess the result’s scope. It presents AI as a technology directed at work that has historically been hard to make more productive, while leaving open whether higher productivity will change employment or staffing.
From AI to Services How 80% Ended Up There
Services account for about four in five U.S. workers. AI may raise productivity in some service work, but the evidence does not show that it reduces jobs.
Two measures, different coverage
The figures describe different populations, so they are not interchangeable.
- ILO estimate: 79.6% worked in services in 2025.
- This is the measure behind the article’s “about 80%” framing.
- BLS-based calculation: 86% were in service-providing industries in 2025.
- Excludes farm workers and self-employed people.
A long shift toward services
Service industries’ share of nonfarm payroll jobs grew over decades, before and after manufacturing employment peaked.
Why some services resist speed
Baumol’s cost disease describes work where labor is central to the service, so reducing time or staff can change what customers receive.
A 45-minute Schubert quartet takes four musicians a combined three hours of performance labor.
Technology can raise output per hour in some industries; faster performance may alter a service built around personal attention or live work.
Retail and travel have also changed through innovations such as self-service supermarkets and faster transport.
What the AI result shows—and doesn’t
The cited customer-support study reports a productivity gain, not an employment outcome.
| Reported | Not established |
|---|---|
| AI increased issues resolved per hour by 15% on average among 5,172 agents. | Whether staffing fell, prices dropped, waits shortened, quality changed, or wages rose. |
| The article presents AI as a possible tool for service productivity. | Whether the result applies across services or persists over time. |
AI Targets Service Productivity
The size of the service sector makes its productivity consequential for workers, businesses and household costs. If a tool allows staff to resolve more customer issues per hour, employers may be able to handle more work with the same staff, reduce waiting times, or change how many workers they need. The cited study reports a productivity result, but the source does not show which of those outcomes followed.
The distinction matters because a measured increase in output per hour is not the same as a fall in prices or a reduction in jobs. Whether AI changes costs depends on how organizations use it, whether the work can be automated without reducing quality, and how any gains are distributed. The article frames that as an open question, not a settled forecast.
The report also cautions against explaining rising service costs through a single cause. It compares U.S. spending on health care—18.0% of the economy—with Germany’s 12.4%, and notes that haircuts and car repairs also rose in price. Those comparisons suggest that productivity pressures alone do not account for every difference in prices or spending; the figures do not, by themselves, establish why the countries differ.
Products Worth Considering
Apple AirPods Pro (2nd Generation) Wireless Earbuds with USB‑C Charging, Up to 2X More Active Noise Cancelling, Transparency Mode, Adaptive
Bose QuietComfort Headphones (2nd Gen), Wireless Over-Ear Noise-Cancelling Headphones, Black
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Twilight Blue
As an affiliate, we earn on qualifying purchases.
From Quartets to Service Jobs
The report’s central example comes from economists William Baumol and William Bowen’s work on performing arts. In their 1966 book, Performing Arts: The Economic Dilemma, they described the labor required for a live performance of a 45-minute Schubert quartet: four musicians still spend a combined three hours performing it. That example illustrates work where cutting the time or number of performers may alter the product audiences value.
In a 1967 paper, Baumol applied the idea to the broader economy, distinguishing activities where labor is mainly an input from those where labor is central to the service. The theory is not a claim that every service is immune to technology. The source itself notes that retail and travel have changed through innovations such as self-service supermarkets and faster transport.
The employment measures in the report are not interchangeable. The BLS figure covers nonfarm payroll jobs and reports the share in service-providing industries; the ILO estimate covers workers more broadly and puts the services share at 79.6% in 2025. The article’s title refers to the roughly four-in-five estimate, not the 86% payroll figure.
What AI Has Not Settled
The source does not identify the full study behind the 15% average increase in customer-support issues resolved per hour, including its publication details, timeframe, comparison method or whether the result persisted over time. The figure should be treated as a reported study result, not as a general measure of AI’s effects across services.
It is also unclear whether the reported productivity gain led to lower prices, shorter customer waits, changes in service quality, higher wages or fewer jobs. The article offers no evidence that AI has already caused a broad decline in service employment. Its employment data describe the sector’s size, while the customer-support study concerns output per hour; they do not establish a causal link between the two.
The article’s price examples and international health-spending comparison also do not isolate the causes of rising costs. The material does not provide the underlying calculation details or a full account of how national health systems differ, so those comparisons cannot establish that the same economic forces explain each price change.
Evidence to Watch in Services
The next evidence needed is whether AI productivity gains extend beyond the customer-support setting described in the source and whether they persist in real workplaces. Useful follow-up measures would include output per worker, service quality, customer wait times, prices, wages and staffing, reported over a clear period and against a defined comparison group.
For the broader employment question, future labor data can show whether services continue to account for a similar share of U.S. work. Those figures should be read with their coverage stated: payroll jobs, all employed workers, and specific service industries measure different populations. The source article does not announce a new official forecast or a scheduled policy decision, so no specific next milestone is confirmed.
Key Questions
What does the claim that about 80% of U.S. workers are in services mean?
The source cites an ILO estimate of 79.6% of U.S. workers in services in 2025. A separate BLS-based calculation says service-providing industries accounted for 86% of nonfarm payroll jobs that year; the figures use different coverage and should not be treated as the same statistic.
Why does the article use a string quartet to explain service costs?
A 45-minute quartet takes four musicians a combined three hours of performance labor. The example illustrates Baumol’s argument that some work cannot be made faster without changing the service, unlike production where technology can reduce labor per item.
Did the cited AI study show that customer-support jobs are disappearing?
No. The source reports that an AI assistant increased issues resolved per hour by 15% on average in a study of 5,172 agents. It does not say whether staffing or employment fell, or provide enough study detail to assess longer-term effects.
Does Baumol’s theory explain every increase in service prices?
No. The theory describes how costs can rise in some labor-intensive services when wages increase while productivity is harder to raise. The source also points to differences in U.S. and German health spending, but does not establish a single cause for those differences or for every price increase.
Source: Thorsten Meyer AI
