Dive Brief:
- Nearly 90% of customers say companies that use generative AI in customer support must also ensure a human agent is available, according to a Gartner survey of more than 3,500 B2B and B2C customers released earlier this month.
- While customers are in agreement that they want a human service option, there is little consensus when it comes to their views of AI. Half of customers say their interactions are easier when companies use generative AI.
- This split view on generative AI is nothing new, Eric Keller, senior director analyst in Gartner’s customer service and support practice, told CX Dive. “When we look at data about customer preferences, you get about half of customers, for a while, have been enthusiastic towards self-service options and efficiency options, and they don't want to talk to somebody. And about half of customers for a while have been leaning towards the ‘I want to talk to a person’ side of a fence.”
Dive Insight:
While generative AI has a lot of potential, it all too often fails to meet expectations. Consumers aren’t against the technology writ large — they just want the technology to work.
About half of consumers say they would be “delighted” if AI could handle their problems quickly even if it “sounded like a toaster,” according to a recent PolyAI survey of more than 1,000 consumers. Nearly 9 in 10 of consumers say fast and accurate resolutions determine whether they buy from a brand.
Making sure generative AI can provide fast and accurate resolutions also requires knowing when to — and when not to — use it.
“AI-powered virtual agents, chatbots and other self-service tools can handle routine, predictable interactions quickly and at scale — whether that’s checking an order status, confirming an appointment, resetting a password or answering a straightforward question,” CallMiner CMO Eric Williamson said in an email. It can also aid human agents, providing them information and context during calls so they can better answer the customer’s inquiry.
More broadly, there are three key principles to successful, helpful generative AI in customer service, according to Keller. It needs to: transparently convey it is AI, explain its answer and process, and provide an off ramp to a human.
If the AI tries to disguise itself as a human, “that puts a sour taste in a customer's mouth” and erodes trust, Keller said.
If a customer begins with AI but wants to talk to a person, it's important that the customer has the ability to do that.
“A lot of companies hesitate because they have these strong goals around containing that interaction and self-service and reducing that cost, but if you make it easier to switch to a person, it helps with longer-term adoption,” Keller said. “Customers will go back and use that over time because they know it's a good place to start, and if they need more help, it'll be easy to connect there.”
The handoff between AI and humans needs to be seamless as well, according to Williamson. “Virtual agents and bots should be able to resolve the interactions they’re well suited for, but when a customer needs a person, the transition should be seamless. The human agent should have the context of what the customer has already tried so they aren’t forced to start over.”
