Will AI replace customer service reps? The evidence
AI already answers a large share of routine chats and calls at some large companies, and at least one CEO says it let him cut thousands of support jobs. Complex, emotional or phone-based cases still go badly without people: Klarna brought humans back after quality fell, and most support leaders plan to keep human agents.
The job today
Task by task: what AI can actually do
In real use means deployed at scale; shown in tests or pilots means demonstrated but not yet routine; not yet means we found no credible evidence it does this reliably.
| Task | Status | Evidence |
|---|---|---|
| Answering routine chat questions (refunds, returns, payments) | In real use | Klarna said in Feb 2024 its AI assistant handled 75% of customer chats, about 2.3 million conversations, in its first month, the work of 700 agents. (Entrepreneur, May 2025) |
| Handling support cases end to end | In real use | Salesforce CEO Marc Benioff said AI agents handle about half of support calls and that he reduced support staff 'from 9,000 heads to about 5,000' (Sep 2025). (KTVU, Sep 2025) |
| Suggesting replies to human agents | In real use | A study of 5,179 support agents found an AI assistant raised issues resolved per hour by 14% on average and 34% for novices, with little effect for the most experienced. (Quarterly Journal of Economics, 2025) |
| Automated support workflows via API | In real use | Anthropic's Economic Index found customer service tasks prevalent in business API traffic, with Claude doing a high share of the tasks in automated workflows (Mar 2026). (Anthropic Economic Index, Mar 2026) |
| Complex, emotional or high-stakes cases | Not yet | Klarna's CEO said in May 2025 the AI-first approach gave 'lower quality' service and the company is rehiring humans so 'there will always be a human if you want'. (Entrepreneur, May 2025) |
Jobs evidence so far
Labelled by who says AI was the reason: the company itself, reporters, or a study.
- Salesforce: Benioff says he cut support headcount from about 9,000 to 5,000 'because I need less heads' as AI agents took on half of calls (company says AI) (KTVU, Sep 2025)
- Klarna, which said its bot did the work of 700 agents, reverses course and rehires human agents after quality complaints (company says AI, then reverses) (Entrepreneur, May 2025)
- Gartner predicts half of organisations that planned big customer-service workforce cuts will abandon them by 2027; 95% of 163 leaders polled plan to keep human agents (study) (Gartner, Jun 2025)
- Microsoft study of 200,000 Copilot conversations lists customer service representatives among occupations with the highest AI applicability (study) (Sinclair/Newschannel 9, Aug 2025)
What to do
- Specialise in the hard casesEscalations, complaints, retention and anything involving vulnerable customers are where companies have pulled back from full automation.
- Use the assistant, don't fight itThe best evidence shows AI suggestions make newer agents much faster; learning to use them well helps your numbers.
- Learn the systems behind the botRoles that tune knowledge bases, review bot transcripts and handle bot hand-offs are growing out of the same teams.
- Watch your employer's deflection rateIf the share of contacts handled by the bot keeps climbing, plan your move into a specialised or adjacent role early.
Questions people ask
Will AI replace customer service representatives?
It is already reducing headcount at some firms, and BLS projects a 5% decline to 2035. But no evidence shows companies running support without people: Klarna rehired humans and Gartner expects many firms to drop agent-less plans.
Which customer service tasks are most exposed?
High-volume, scripted chat and email questions such as order status, refunds and password resets.
Which tasks are safest?
Complex complaints, emotionally charged calls, retention and situations needing discretion or exceptions to policy.
Did Salesforce really cut 4,000 support jobs because of AI?
Its CEO said so: he described cutting support from about 9,000 to about 5,000 people as AI agents took half the calls. That is the company's own claim.
Does AI make agents better or worse?
In a large study, AI suggestions raised productivity 14% on average and 34% for novices, with customer sentiment improving.
How we grade
Each task is graded 'deployed' only when a named organisation reports AI doing it in real work at scale, 'demonstrated' when it works in tests, benchmarks or pilots, and 'not yet' when we found no credible evidence AI does it reliably. For job losses we separate what a company said in its own words ('company says AI') from what reporters or analysts attributed to AI ('reported as AI') and from research ('study'); employment and outlook figures come from the BLS 2025-35 projections.
Sources
- BLS Occupational Outlook Handbook: Customer service representatives
- NBER/QJE: Generative AI at Work
- KTVU: Salesforce cut 4,000 jobs due to AI, Benioff says
- Entrepreneur: Klarna hiring human agents again
- Gartner press release (Jun 2025)
- Anthropic Economic Index, Mar 2026
- Microsoft: Working with AI (arXiv)
Spot an error or something out of date? Tell us; corrections are made and logged. Written with AI assistance from public sources; every fact checked and approved by Colin Balfe. General information, not legal, medical or financial advice.