Main points
- Customer service roles are most at risk. Anthropic’s new AI workforce analysis finds that customer service representatives are the second most impacted occupation by AI task automation.
- Automation pressures are already shaping hiring. The impact on unemployment has so far been limited, but early data suggests recruitment into high-risk jobs may slow, particularly among younger workers.
- CX leaders are facing a workforce redesign. As AI increasingly handles routine interactions, contact center leaders must rethink training, workforce planning, and escalation workflows.
Another day, another report predicting that customer service representatives and contact center agents will be displaced by artificial intelligence. The researchers this time come directly from one of the companies responsible for that AI technology, through observations on their own platform.
Customer service representatives continue to emerge as one of the most at-risk roles in the AI economy, according to a new labor market analysis from Anthropic.
The study, “AI’s Labor Market Impact: New Measures and Early Evidence,” introduces a new measure called . Observed exposure — A framework that combines theoretical AI capabilities with real-world usage data extracted from millions of interactions with Anthropic’s Claude AI system.
The findings place customer service representatives as one of the most exposed jobs to AI in the economy, highlighting the growing potential for automation of routine support interactions.
For customer experience and contact center leaders, this report confirms what many organizations are already experiencing first-hand. AI is not only enhancing customer service workflows, but is also beginning to reshape the workforce behind them.
“This report introduces a new metric for understanding the labor market impact of AI, studying its impact on unemployment and employment,” Antropic researchers said. “Jobs are exposed to AI as far as it is theoretically doable using LLM and observed on our platform in automated work-related use cases. We found that computer programmers, customer service representatives, and financial analysts were most exposed.”
table of contents
Customer service representatives near the top of AI exposure
This report ranks occupations based on how much of their task portfolio can be automated or accelerated by AI systems. Customer service representatives rank second overall and are responsible for over 70% of tasks. 1st place? computer programmer.
The main reason is simple. Many of the tasks that support agents perform, such as answering questions, providing information, troubleshooting common issues, and processing requests, are increasingly being handled by AI systems.
Anthropic points out that these tasks are already frequently appearing in the use of AI in production environments through enterprise API integration and workflow automation.
For contact centers, this is consistent with industry changes that have accelerated over the past three years. Generative AI agents now routinely do the following:
- Solve Tier-1 customer questions
- Summarize conversations with customers
- Generate suggested responses for agents
- Automate knowledge searches
- Process routine service transactions
As these capabilities expand, customer service organizations are increasingly moving toward hybrid service models where AI solves the majority of interactions and humans handle the rest of the complexity.
Related article: Is this the year of artificial intelligence call centers?
The most exposed profession in the human world
The following table reproduces the key exposure rankings from Anthropic’s report. Source: Anthropic, “AI’s Labor Market Impact: New Measures and Early Evidence.”
| Occupation | Observed exposure | Key automation tasks |
|---|---|---|
| computer programmer | 74.5% | Creating, updating, and maintaining software programs |
| customer service representative | 70.1% | Meetings with customers to provide information, receive orders, and respond to complaints |
| data entry keyer | 67.1% | Read source documents and enter data into the system |
| medical records specialist | 66.7% | Compile, abstract, and code patient data |
| Market Research Analyst and Marketing Specialist | 64.8% | Create reports and insights documenting your findings. |
| Sales person (wholesale/manufacturing) | 62.8% | Contact customers to demo products and solicit orders |
| financial and investment analyst | 57.2% | Analyzing financial information and creating forecasts |
| Software Quality Assurance Analyst and Tester | 51.9% | Modify the software to fix errors or improve performance |
| information security analyst | 48.6% | Perform risk assessments and test the security of data processing |
| computer user support specialist | 46.8% | Answer user questions regarding the operation of software or hardware |
AI has not yet caused mass unemployment
The report reveals that despite the exposure levels, there has been no clear increase in unemployment rates in high-exposure occupations so far.
Rather, early signs of labor seem more subtle. And in fact, it could be a key factor in the 14% increase in output in 2025, according to the U.S. Bureau of Labor Statistics. An analysis of BLS numbers released on March 5 shows that customer service productivity has increased. “This is not conclusive evidence that AI is the cause. The macro data is noisy, but the micro-level evidence points in the same direction. Tasks that AI can perform increase productivity by 14% for customer service, 26% for developers, and about 25% for consultants.”
Researchers found suggestive evidence that recruitment into high-exposure occupations, including customer service roles, may be slowing, particularly among younger people entering the workforce.
Current Population Survey data suggests job search rates for young workers in high-risk occupations will decline by about 14% compared to 2022 levels.
This may indicate that AI adoption is impacting workforce growth before impacting layoffs. This is a pattern that economists have seen in previous changes in technology.
Will AI lead to a revolution in the role of customer service?
Generative AI has the potential to automate routine tasks such as issuing tickets, taking reservations, and handling basic inquiries.
However, this transformation appears to be more subtle than widespread exclusion. Industry research shows that more than 80% of organizations expect to reduce their agent workforce within the next 18 months, but a similar number plan to move agents into new roles (such as automation supervisors, escalation specialists, and AI trainers) rather than eliminating positions entirely.
Gartner and Forrester research shows that companies that focus solely on reducing headcount often find themselves rehiring in similar roles within a few years. According to the researchers, automation strategies often overestimate the capabilities of AI and underestimate the complexity of customer needs.
Key reports examining the impact of AI on customer service jobs
The following research from leading consulting firms, analysts, and academic institutions explores how generative AI and automation will impact customer service representatives and contact center agents.
Why customer service operations can be automated so much
Anthropic’s methodology analyzes job exposures at the task level rather than the occupation level.
Customer service jobs score higher because their core tasks closely align with AI strengths.
- language understanding
- Information search
- Process description
- Workflow automation
According to the task exposure metric used in the report, AI systems can already complete many support interactions twice as fast as humans in structured scenarios.
However, anthropology researchers stress that the adoption of AI remains far below its theoretical potential. Even in the most at-risk sectors, current AI coverage only scratches the surface of what could eventually become possible.
Related article: Microsoft AI CEO says marketing will be automated within 18 months
What CX leaders should do now
For customer experience leaders, this report highlights the strategic changes already underway in contact centers.
Rather than asking whether AI will impact support teams, a more relevant question is how quickly organizations can redesign their service models.
Three priorities stand out.
1. Redesign agent roles around escalation
As AI handles routine requests, human agents are increasingly focused on complex interactions such as escalations, sensitive customer issues, and multi-system troubleshooting.
This requires different recruitment profiles and training programs.
2. Invest in AI supervision skills
Agents are increasingly responsible for monitoring AI responses, correcting output, and ensuring compliance with policies.
In other words, their role evolves from answering questions to monitoring automated service flows.
3. Prepare for changes in workforce composition
As hiring for high-risk roles slows, organizations may gradually move to smaller field teams supported by automation.
This does not necessarily mean that the total number of CX employees will decrease, but it does mean that their roles will be different.
The real story: AI will reshape entry points
The report’s most important signal may not be unemployment, but labor force entry.
Customer service roles have traditionally served as an entry point into the workforce and a stepping stone to other business functions.
As AI increasingly handles the mundane interactions that once trained new employees, organizations will need to rethink how they develop their talent pipelines.
For CX leaders, the challenge may prove to be more important than automation itself. As contact center AI continues to evolve, the focus must shift from simply implementing the technology to strategically redesigning how humans and AI work together to deliver superior customer satisfaction.
Agents themselves have expressed misgivings about the shift. Contact center research shows that while many people fear losing their jobs, they also worry that the human element in customer interactions will be lost, leaving only the most emotionally taxing cases.
extended approach
There is a growing consensus that AI is best used to augment, rather than replace, human agents, allowing them to manage mundane tasks while freeing them to handle complex or sensitive issues. What did our report find? It’s important for organizations to invest in upskilling and redesigning workflows to maintain service quality and employee engagement during the transition.
