Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
Customer service jobs have long been considered among the most vulnerable to the roll out and adoption of increasingly powerful artificial intelligence tools — a technology hailed by a senior executive at one of the world’s largest insurers as a “superpower”.
Germany’s Allianz has embraced the use of AI-powered applications, including the launch of a voice assistant in September, which provides roadside support in the event of breakdowns. It works in more than 20 languages, and is trained to prioritise the most urgent calls — such as those from a lone parent with young children in a car that has broken down in the middle of the night — and immediately transfer them to human agents.
In July, the insurer’s Australian business launched an app to speed up the processing of claims by homeowners for food spoilt in their refrigerators following power cuts caused by severe weather. “Generative and agentic AI is like a superpower which we use in order to improve the lives of our customers,” says Josef Teglas, Allianz’s group head of data and AI. “We are able to reduce the turnaround time and be there whenever needed, and resolve problems and claims much faster.”
But he also emphasises the importance of rolling out AI in a “responsible and safe manner”. Analysts say this caution reflects widespread feedback that AI still struggles to deal with more complex, real-world scenarios for customers, making it unlikely that call centres will fall quiet any time soon.
A recent report by research company Gartner concluded that a fully automated customer service function was “both unlikely and undesirable”. The findings contradict the views of Sam Altman, chief executive of AI developer OpenAI, who earlier this year echoed warnings that the widespread adoption of AI meant some job categories, including customer service, would be “totally, totally gone”.
The study, Agentless Customer Service Should Not Be Your Goal, predicts that “by 2027, 50 per cent of organisations that expected to significantly reduce their service workforce due to AI will drop these plans”.
“AI and automation are transforming how customer service organisations serve customers, but human agents are irreplaceable when it comes to handling nuanced situations and building lasting relationships,” says Kathy Ross, a co-author of the Gartner study.
She believes a hybrid approach is the future. “When customers need that human connection, when they need to have an issue which is complex or sensitive resolved, AI will likely work as a teammate so [the company] can provide higher-quality customer service,” she says.
Jo Causon, chief executive of the London-based Institute of Customer Service, agrees that what she calls a “blended approach” is the way forward for AI-powered customer service. “I don’t think that we would want to see AI as a pure replacement for human skills. It’s more about the tool to make humans more effective,” she says.
I don’t think that we would want to see AI seen as a pure replacement for human skills
Jo Causon, CEO, Institute of Customer Service
At travel portal Expedia, for example, chief product officer Shilpa Ranganathan says AI resolves more than 50 per cent of customer queries. But clients are transferred to a member of staff where required, such as when dealing with more complicated travel itineraries. “We use AI as our first touch but, at the same time, whatever information the AI collects is seamlessly transitioned to the human agent,” she explains.
Jonathan Corbin, chief executive of MavenAGI, which builds AI agents, agrees that the technology is unlikely to fully replicate the roles of customer service staff. “I don’t think it’s actually a case of us saying, ‘Hey, we’re replacing humans.’ We’re actually letting humans be able to engage with the customers who need it most: folks like my mom who really need that human touch.”
John Leighton, customer service director at easyJet, says the European low-cost airline uses AI in tandem with staff. The technology offers recommended responses to enquiries via telephone, live chat and email. Staff then review these and tweak them before sending a reply. The implementation of AI “is an investment in customer experience from our perspective, rather than looking at a full kind of cost-cutting exercise”, Leighton says.
Recommended
The airline is also using AI-powered social media scanning to identify customers most urgently in need of support. The tool was rolled out in March 2023 and was used during the serious wildfires in Greece later that year, which led to airlines scrambling to put on extra flights to evacuate tourists from the worst-hit parts of the country. The tool flagged terms such as “stranded” and “elderly”, allowing easyJet staff to identify and provide assistance to its most vulnerable customers, says Leighton.
Nick McBrien, financial services chief executive at The Very Group, the UK-based online retailer, says staff were initially “quite apprehensive” about the introduction of AI, but that has faded since they realised it was helping them perform better.
He attributes the company’s highest-ever customer satisfaction result, based on its own metrics, in the financial year to the end of June 2025 at least partly to the guidance and information the technology provides to employees. “They quickly saw that we weren’t trying to replace them and we were trying to support them,” McBrien adds.
