Artificial intelligence may be transforming customer service faster than many expected, but opinions remain sharply divided on whether humans will stay in the loop. While AI firms insist that automated systems and humans will collaborate, tech leaders and analysts suggest that an era of predominantly machine–run customer support may be only a few years away.
Last year, Tata Consultancy Services chief executive K Krithivasan predicted that AI could soon leave “minimal need” for call centres in Asia. Research firm Gartner expects AI to independently resolve 80% of routine service queries by 2029 – a figure that could reshape entire industries built on human customer support.
This shift is being driven by the rise of “AI agents” – advanced systems capable of acting autonomously, making decisions, and moving beyond the limitations of traditional rule-bound chatbots. Current chatbots are limited to pre-set responses, a problem many customers continue to experience firsthand. When a delayed parcel was queried with courier firm Evri’s chatbot, the automated system incorrectly stated the item was delivered, showed a photo of a package at the wrong address, and offered no option to continue the conversation. In response, Evri says it is investing £57 million to upgrade services and insists that most customers get instant solutions from its system, though many remain unconvinced.
Rival courier DPD had a different problem – upgrading to a more advanced AI that ended up insulting users and criticising the company. It was forced to shut the chatbot down, highlighting how easily AI can go off-script.
Despite such incidents, AI adoption continues at breakneck speed. According to Gartner, 85% of customer service leaders are already experimenting with or deploying AI chatbots. Yet only 20% believe their systems are meeting expectations. Analysts warn that while AI allows more natural conversations, it can also produce false information or “hallucinate”, creating fresh risks for businesses.
Cost savings are driving much of the shift, though experts warn AI is not automatically cheaper. Companies need huge amounts of high-quality training data, strong knowledge-management systems and constant oversight. Salesforce, which has deployed its AgentForce AI platform for clients including Formula 1, Prudential and Reddit, says customer satisfaction scores are now higher with AI than humans. It claims 94% of users opt for the AI agent when given the choice, and that the system saved the company $100 million in service costs. While reports suggest thousands of human roles were cut, Salesforce states many were redeployed.
Not everyone is convinced. Some firms are using AI not to replace people, but to improve working conditions. QStory, which works with eBay and NatWest, uses AI to offer call centre workers more flexible schedules. Its founder, Fiona Coleman, argues that emotional intelligence is still a human strength. Handling debt, mortgages, and personal crises, she says, requires empathy current AI cannot provide.
Governments are also stepping in. Proposed US legislation would require businesses to disclose when callers are interacting with AI and allow them to request a human. Gartner forecasts the EU could adopt a legal “right to talk to a human” by 2028 as part of consumer-protection reforms.
The future of customer service may depend on how fast AI develops – and how much consumers are willing to accept it. Businesses see huge cost savings and faster interactions. But for millions of customers, the difference between convenience and frustration may still come down to speaking to a real person on the other end of the line.
