A recent Gartner survey found that 64 per cent of customers prefer companies avoid using artificial intelligence (AI) in their customer-service interactions. For businesses, the stakes are high.
More than half of those surveyed said they would consider switching to a competitor if they discovered a company was using AI to power its customer service, exposing a trust gap between businesses eager to adopt the technology and the clients they value.
At the heart of these concerns are fears around losing the benefits of human interaction. Customers worry about difficulties reaching a human agent, the risk of AI displacing jobs, and the potential for receiving inaccurate information.
The skepticism could originate from prior interactions with older AI systems that often failed to meet expectations, says Rob Dunlap, partner and generative AI lead at IBM in Toronto. “The concern for customers [is] generally, ‘I’m going to get a bot that doesn’t know how to help me … I might get an incorrect answer … and I’m still stuck in the queue waiting for [human] chat.’”
Some firms, on the other hand, are finding success with chatbot technology. Lori Bieda, BMO’s Toronto-based chief data and analytics officer, says the company’s chatbots have handled more than two-million interactions in the past two months alone, with a self-reported satisfaction rate of 89 per cent to 92 per cent.
“What we know from our chatbots is that when a customer receives knowledgeable service, their net promoter score (NPS) increases by 33 per cent,” she says. An NPS shows how loyal customers are by measuring the likelihood they’ll recommend a business.
Trust, says Ms. Bieda, is foundational to rolling out any AI initiatives. BMO’s adoption, especially in its chatbot program BMO Assist, uses the “human in the loop” approach, ensuring that even as AI takes over routine inquiries, human intervention remains available. “It’s [this] collaboration that creates relevancy and personalization for the customer,” Ms. Bieda says.
Concerns about job loss are not without merit. Ms. Bieda says she acknowledges AI has streamlined many processes but insists it’s about shifting resources rather than reducing them. AI can take over busy work, allowing companies to hire more knowledge workers.
“In cases where customers are looking to retrieve information, we want to do that in the most efficient way possible,” she says. “But we are also seeing a big demand for advice – portfolios, trade-offs, new life stages. So, what we’re doing is optimizing and shifting our work force to accommodate those emerging needs.”
Elsewhere, job losses are becoming a reality. This year, brands such as L.L. Bean, American Airlines, Best Buy and Verizon have all made cuts to their customer-service teams as they introduce AI. Fintech company Klarna also recently released internal statistics that claim its customer-service AI assistant is “doing the equivalent work of 700 full-time [customer-service] agents.”
Accountability is also a significant issue with AI adoption, says Dr. Gene Moo Lee, associate professor of information systems and analytics at the University of British Columbia. “AI is very good, but it’s not really explainable or interpretable. … It’s a black box. We don’t know what’s going on inside,” he says.
Companies can be held accountable for any inaccurate information provided by their AI systems. Earlier this year, Air Canada was held liable after its chatbot promised a passenger a discount that wasn’t available to them.
For many customers, it’s not just about the use of AI, but how it’s deployed. “People want it in a way that’s acceptable to them,” says Bern Elliot, research vice-president at Gartner in Philadelphia, who focuses on natural language technologies, pointing out that poorly designed AI systems often miss the mark by over-automating or failing to provide clear pathways to human agents.
There’s also what Mr. Elliot refers to as the “creepiness factor.” When AI mimics human characteristics too well, it can feel invasive and uncomfortable for consumers. “Even if you know what someone wants to order for dinner, you should still ask them what they want.”
Despite these challenges, businesses such as IBM are optimistic about the future of AI in customer service. Mr. Dunlap says many organizations are taking a cautious approach by testing AI internally before deploying it to customers. “A lot of organizations are going to prove it internally and then ask, ‘How can we make this powerful for our customers as well?’”
He suggests a focused approach to adoption, starting small and ensuring human agents are still accessible when needed. “Don’t try to solve for everything at once.”