Will AI Replace Customer Service Agents? The Evidence
The evidence points to a gradual fall in customer service jobs, with AI taking simple, repetitive contacts and people keeping complex, emotional and regulated ones. US projections show customer service representative jobs falling 5 percent from 2025 to 2035, a large field study found that an AI assistant made agents 14% more productive, and at least one company that cut human support has hired people back. Regulators are moving toward disclosure of AI and a route to a human.
What does the evidence say at a glance?
Studies and labor data agree that AI reduces the human work customer service needs, and none shows the function running without people.
| Source | What it measured | Finding |
|---|---|---|
| US Bureau of Labor Statistics, August 2026 | Employment projections, 2025 to 2035 | Customer service representative jobs projected to decline 5 percent |
| Brynjolfsson, Li and Raymond, NBER, 2023 | AI assistant given to 5,179 support agents | Issues resolved per hour up 14% on average, 34% for novice and low-skilled agents |
| Brynjolfsson, Chandar and Chen, Stanford, August 2026 | US payroll data through June 2026 | Fewer young workers in AI-exposed jobs, including customer support |
| Humlum and Vestergaard, NBER, revised March 2026 | Danish workers in 11 exposed occupations, including customer support | Null effects on earnings and hours, ruling out effects above 2% two years after ChatGPT's launch |
| McKinsey, August 2026 | Survey of 1,719 participants | 14 percent said AI contributed to a decline in workforce size in the past year |
What do labor statistics project for customer service jobs?
US labor projections show a gradual decline. The Bureau of Labor Statistics, on a page last modified August 27, 2026, states: "Employment of customer service representatives is projected to decline 5 percent from 2025 to 2035." Its table shows 2,666,000 jobs in 2025 and 2,524,100 in 2035, a drop of 141,800.
The page gives the cause: "There is expected to be less demand for customer service representatives, especially in retail trade, as their tasks continue to be automated." It also expects that "Some companies will continue to use in-house service centers to differentiate themselves from competitors, particularly for complex inquiries such as refunding accounts or confirming insurance coverage." Despite the decline, BLS projects "about 289,500 openings for customer service representatives" each year, all from replacing workers who leave.
An earlier BLS analysis named AI directly. A July 16, 2026 note on the 2024–34 projections lists customer service representatives among occupations where "demand is expected to be limited" as AI brings "additional efficiency gains".
BLS also warns against reading its new AI exposure ratings as forecasts. Its AI exposure categories page (last modified August 27, 2026) says: "An exposure category is not a forecast of employment growth or decline."
What happened when a call center gave agents an AI assistant?
A large field study found that AI assistance made agents faster, with the largest gains for new staff. In "Generative AI at Work" (NBER Working Paper 31161, April 2023, revised November 2023, later published in the Quarterly Journal of Economics), Erik Brynjolfsson, Danielle Li and Lindsey Raymond studied an AI conversational assistant rolled out to "5,179 customer support agents".
Their abstract reports: "Access to the tool increases productivity, as measured by issues resolved per hour, by 14% on average, including a 34% improvement for novice and low-skilled workers but with minimal impact on experienced and highly skilled workers." They also found that "AI assistance improves customer sentiment, increases employee retention, and may lead to worker learning."
The tool assisted agents, who still handled every conversation. Even so, a 14% gain in issues resolved per hour means the same contact volume needs fewer agent hours, unless volume grows.
Is AI already reducing customer service employment?
Early data shows fewer entry-level hires in AI-exposed jobs, while other studies find little effect on pay or hours so far. The two can both be true, because they measure different things.
The hiring signal comes from US payroll data. In the August 2026 revision of "Canaries in the Coal Mine?", Erik Brynjolfsson, Bharat Chandar and Ruyu Chen report that "employment of young workers (ages 22–25) in AI-exposed occupations now stands 19% below where it would be had it kept pace with that of their less-exposed peers". They write that employment "has declined substantially for early-career workers in the occupations most exposed to AI, including but not limited to software development and customer support." The change operates "primarily through reduced hiring of young workers rather than increased separations". The authors call their results "early, descriptive indicators" rather than causal estimates.
Danish data points the other way on pay and hours. Anders Humlum and Emilie Vestergaard's NBER Working Paper 33777 (May 2025, revised March 2026) covers 11 exposed occupations, including customer support specialists. It reports "precise null effects on earnings and recorded hours at both the worker and workplace levels, ruling out effects larger than 2% two years after the launch of ChatGPT."
Employers' expectations have also run ahead of results. McKinsey's State of AI 2026 survey (August 25, 2026) found that "Just 14 percent of respondents from organizations using AI report that AI contributed to an overall decline in workforce size in the past year", less than half the 32 percent who had expected reductions a year earlier.
What happened at companies that replaced agents with AI?
Company results are mixed. Some firms cut support staff after deploying AI, and one widely reported case has since rehired people.
Klarna announced on February 27, 2024 that its AI assistant "has had 2.3 million conversations, two-thirds of Klarna's customer service chats" in its first month and was "doing the equivalent work of 700 full-time agents". By May 2025 the company had changed course. Fortune reported on May 9, 2025 that chief executive Sebastian Siemiatkowski told Bloomberg the company was recruiting humans again after the AI approach led to "lower quality". He was quoted as saying: "[I]t's so critical that you are clear to your customer that there will be always a human if you want."
Salesforce went the other way. CNBC reported on September 2, 2025 that chief executive Marc Benioff said of the support organization, "I've reduced it from 9,000 heads to about 5,000, because I need less heads." In a statement, the company said that "we've seen the number of support cases we handle decline and we no longer need to actively backfill support engineer roles."
Analysts expect more reversals. In a press release dated June 10, 2025, Gartner predicted that "By 2027, 50% of organizations that expected to significantly reduce their customer service workforce will abandon these plans". A Gartner poll of 163 service leaders in March 2025 found that "95% of customer service leaders plan to retain human agents".
What do customers want from AI customer service?
Customers mainly want to reach a person when self-service fails. A Gartner survey of 5,728 customers, published July 9, 2024, found that "Sixty-four percent of customers would prefer that companies didn't use artificial intelligence (AI) in their customer service." Their top concern was "that it will get more difficult reaching a person".
Self-service leaves most issues unresolved, by Gartner's measure. A Gartner press release of August 19, 2024, drawing on the same survey, reported that "Only 14% of customer service and support issues are fully resolved in self-service", and only 36% of issues customers describe as "very simple". Gartner sells advisory services to service leaders.
What do regulators require for AI calls and chats?
Regulators are moving toward three requirements: consent for AI-voice calls, disclosure that the customer is dealing with AI, and a way to reach a human. The rules differ by country and by channel.
| Jurisdiction | Rule | Status | What it means |
|---|---|---|---|
| United States (FCC) | Declaratory Ruling FCC 24-17: the TCPA's "artificial or prerecorded voice" covers "current AI technologies that generate human voices" | Adopted February 2, 2024 | Outbound AI-voice calls need prior express consent, absent an emergency purpose or exemption |
| United States (FCC) | Proposed rule requiring callers using AI-generated voice to "clearly disclose to the called party that the call is using AI-generated technology" at the start of each call | Proposed, Federal Register, September 10, 2024 | Covers outbound AI-generated calls; check the docket for final rules |
| United States (Congress) | S. 2495, Keep Call Centers in America Act, as introduced July 29, 2025 | Bill | Would require disclosure "that a nonhuman, artificial intelligence or machine is being used for customer service" and a transfer to a US-based human on request |
| California | Business and Professions Code §17941 | In force | Bars bots that mislead people about their artificial identity to drive a sale or influence a vote; a business that "discloses that it is a bot" is not liable under the section |
| European Union | AI Act transparency rules | Apply from August 2026 | People using chatbots "should be made aware that they are interacting with a machine" |
| United States (consumer finance) | CFPB report "Chatbots in consumer finance", June 6, 2023 | Research report | Warns that institutions "risk violating legal obligations" when chatbots fail customers |
The CFPB report also gave the common failure a name. Chatbots can leave customers in "continuous loops of repetitive, unhelpful jargon or legalese without an offramp to a human customer service representative", which it calls "doom loops". These rules are general summaries; take legal advice for your channels and markets.
Where is a human handoff still necessary?
A human handoff is still necessary when the AI cannot resolve the issue, when the customer asks for a person, and when the stakes are financial, legal, medical or emotional. Gartner's Keith McIntosh described the design in the July 2024 release: chatbots "must communicate to the customer that they will connect them to an agent in the event that the AI cannot provide a solution", then hand over to "an agent chat that picks up where the chatbot left off."
Hand off to a person when:
- The customer asks for one, in any wording, including saying "agent" on a call.
- The AI has failed to resolve the issue after a set number of turns.
- The request involves a refund, a dispute, a complaint or a change to a contract.
- The customer appears distressed or vulnerable, or a decision affects money, health, housing or legal rights.
- The AI's confidence or retrieval results fall below a threshold.
Pass the conversation history and the reason for transfer, so the person does not start from zero.
What should a business do now?
Automate simple, high-volume contacts such as order status, bookings and password resets. Give human agents an AI assistant for drafting, search and summaries, the use the NBER study measured. Track resolution and repeat contacts alongside the share of contacts the AI handled, tell customers when they are dealing with AI, and staff for the complex cases that remain.
Key takeaways
- BLS projects US customer service representative jobs to fall 5 percent from 2025 to 2035, with about 289,500 openings a year.
- An AI assistant raised issues resolved per hour by 14% on average and 34% for novice agents in an NBER field study.
- Payroll data shows fewer young workers in AI-exposed jobs, including customer support. Danish data rules out effects on pay or hours larger than 2% so far.
- Klarna said its AI did the work of 700 agents in 2024, then began recruiting humans again in 2025.
- Rules are converging on consent for AI-voice calls, disclosure of AI, and a route to a human.
Frequently asked questions
Will AI replace customer service representatives?
Not entirely, on current evidence. BLS projects a 5 percent decline in US customer service representative jobs from 2025 to 2035 and expects human service to continue for complex inquiries.
Will AI replace call center agents?
AI is taking over simple, repetitive calls and helping agents with the rest. In an NBER study of 5,179 support agents, an AI assistant raised issues resolved per hour by 14% on average.
What is an AI customer service agent?
It is software that answers customers by chat or voice, looks up information in company systems and completes tasks such as bookings. It should hand the conversation, with its history, to a person when it cannot help.
Do companies have to disclose AI in customer service?
In some places, yes. The EU AI Act's transparency rules apply from August 2026. In the US, outbound AI-voice calls need prior consent under the FCC's 2024 ruling, and California bars bots that mislead people about being bots in order to drive a sale.
Does AI make customer service agents more productive?
Yes, in the NBER field study. It found a 14% average gain in issues resolved per hour and a 34% gain for novice and low-skilled agents, with minimal change for the most experienced.
Easital Technologies Ltd. builds AI customer service systems, so we have an interest in this topic. Easital built Calldone for a client in the United States: AI voice agents that answer business phone calls, book appointments, answer questions and qualify leads. See voice AI agent development, AI chatbot development, the Calldone case study and our work. For how voice agents differ from phone menus, read AI voice agents vs IVR.
Sources
All sources were opened and checked on October 5, 2026.
- US Bureau of Labor Statistics, Occupational Outlook Handbook, "Customer Service Representatives", last modified August 27, 2026. https://www.bls.gov/ooh/office-and-administrative-support/customer-service-representatives.htm
- US Bureau of Labor Statistics, The Economics Daily, "Artificial intelligence, information technology, and employment, 2024–34", July 16, 2026. https://www.bls.gov/opub/ted/2026/artificial-intelligence-information-technology-and-employment-2024-34.htm
- US Bureau of Labor Statistics, "Artificial Intelligence (AI) exposure categories", last modified August 27, 2026. https://www.bls.gov/emp/publications/ai-exposure-categories.htm
- Erik Brynjolfsson, Danielle Li, Lindsey R. Raymond, "Generative AI at Work", NBER Working Paper 31161, April 2023, revised November 2023; published in The Quarterly Journal of Economics 140(2), 2025. https://www.nber.org/papers/w31161
- Erik Brynjolfsson, Bharat Chandar, Ruyu Chen, "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence", Stanford Digital Economy Lab, revised August 2026. https://digitaleconomy.stanford.edu/publication/canaries-in-the-coal-mine-six-facts-about-the-recent-employment-effects-of-artificial-intelligence/
- Anders Humlum, Emilie Vestergaard, "Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI" (previously "Large Language Models, Small Labor Market Effects"), NBER Working Paper 33777, May 2025, revised March 2026. https://www.nber.org/papers/w33777
- McKinsey & Company (Dan Tinkoff, Lieven Van der Veken, Michael Chui, with Tara Balakrishnan), "The state of AI in 2026: On the road to ROI", August 25, 2026. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
- Klarna, "Klarna AI assistant handles two-thirds of customer service chats in its first month", press release, February 27, 2024. https://www.klarna.com/international/press/klarna-ai-assistant-handles-two-thirds-of-customer-service-chats-in-its-first-month/
- Fortune (Irina Ivanova), "As Klarna flips from AI-first to hiring people again, a new landmark survey reveals most AI projects fail to deliver", May 9, 2025. https://fortune.com/2025/05/09/klarna-ai-humans-return-on-investment/
- CNBC, "Salesforce CEO confirms 4,000 layoffs 'because I need less heads' with AI", September 2, 2025. https://www.cnbc.com/2025/09/02/salesforce-ceo-confirms-4000-layoffs-because-i-need-less-heads-with-ai.html
- Gartner, "Gartner Predicts 50% of Organizations Will Abandon Plans to Reduce Customer Service Workforce Due to AI", press release, June 10, 2025. https://www.gartner.com/en/newsroom/press-releases/2025-06-10-gartner-predicts-50-percent-of-organizations-will-abandon-plans-to-reduce-customer-service-workforce-due-to-ai
- Gartner, "Gartner Survey Finds 64% of Customers Would Prefer That Companies Didn't Use AI For Customer Service", press release, July 9, 2024. https://www.gartner.com/en/newsroom/press-releases/2024-07-09-gartner-survey-finds-64-percent-of-customers-would-prefer-that-companies-didnt-use-ai-for-customer-service
- Gartner, "Gartner Survey Finds Only 14% of Customer Service Issues Are Fully Resolved in Self-Service", press release, August 19, 2024. https://www.gartner.com/en/newsroom/press-releases/2024-08-19-gartner-survey-finds-only-14-percent-of-customer-service-issues-are-fully-resolved-in-self-service
- Federal Communications Commission, Declaratory Ruling FCC 24-17, CG Docket No. 23-362, adopted February 2, 2024, released February 8, 2024. https://docs.fcc.gov/public/attachments/FCC-24-17A1.pdf
- Federal Communications Commission, "Implications of Artificial Intelligence Technologies on Protecting Consumers From Unwanted Robocalls and Robotexts", proposed rule, Federal Register vol. 89, no. 175, September 10, 2024. https://www.govinfo.gov/content/pkg/FR-2024-09-10/html/2024-19028.htm
- US Senate, S. 2495, "Keep Call Centers in America Act of 2025", introduced July 29, 2025, text as introduced. https://www.govinfo.gov/content/pkg/BILLS-119s2495is/html/BILLS-119s2495is.htm
- California Legislature, Business and Professions Code §17941. https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=BPC§ionNum=17941
- European Commission, "AI Act", Shaping Europe's digital future, last updated August 3, 2026. https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
- Consumer Financial Protection Bureau, "Chatbots in consumer finance", June 6, 2023. https://www.consumerfinance.gov/data-research/research-reports/chatbots-in-consumer-finance/chatbots-in-consumer-finance/