AI Can Be Detrimental to CX

As companies continue to invest heavily in artificial intelligence to help improve the customer experience, those investments might not help as much as those companies hope, according to a new Gartner survey.

“While new technology opens up exciting possibilities, customer behavior is also changing,” Keith McIntosh, senior principal quantitative researcher on Gartner’s Customer Service & Support Practice, notes. “To make the right decisions, service leaders need to stay informed of customer behavior and avoid investing based on technology fads or optimistic vendor promises.”

Customers have major concerns about the use of AI in customer service, especially service organizations using AI to make it more difficult to reach an agent, the firm found in its latest customer service and support priorities survey.

The research found that 64 percent of consumers would prefer that companies not use AI for customer service. Additionally, more than half (53 percent) would consider switching to a competitor over the issue.

People perceive AI as a self-service wall between themselves and a human agent, says Joshua Feast, CEO and cofounder of Cogito. “They’re objecting to that.”

“Any time there’s new technology, there’s a certain amount of natural skepticism,” McIntosh says. “Service leaders need to be aware there’s a certain amount of apprehension out there. If they want to successfully integrate AI into their service channel experience, they need to pay attention to the concerns that customers have.”

Consumers don’t necessarily have a problem with self-service and are fine with companies using AI in that context, according to Feast. “But at some point, the self-service options may not work, and then they want to speak to a human.”

Once customers exhaust self-service options, they are ready to reach a person, McIntosh agrees, and don’t want to be blocked from doing so. “The onus is on service and support leaders to show customers that AI can streamline the service experience,” he says.

Customer service organizations go wrong when they fail to provide customers with a simple way to reach a human agent. Consumers don’t mind AI being used initially to collect and confirm some initial information, like their name and account info, the issue to be discussed (to help drive intelligent routing), and a few other essentials that then can be presented to the agent to speed the interaction.

“But when it goes beyond that, they get frustrated because it’s like the company is being patronizing,” Feast says.

But the biggest consumer issue, according to Feast, is when a company provides a phone number that gives the impression that this will enable a connection to a real-life agent only to find that the number connects to a chatbot or interactive voice response (IVR) system instead, with any connection to a live agent too complex to accomplish.

“They feel cheated,” Feast says.

Intentional barriers to customer-agent contact don’t work, Gartner says, noting that they harm the customer experience and don’t lower customer service costs.

When companies do roll out AI, they can build critical customer trust by ensuring that their general AI capabilities follow best practices of service journey design, according to McIntosh. “Customers must know that the AI-infused journey will deliver better solutions and seamless guidance, including connecting them to a person when necessary. AI-infused chatbots must communicate to the customer that they will connect them to an agent in the event that the AI cannot provide a solution. It must then seamlessly transform into an agent chat that picks up where the chatbot left off.”

But while user acceptance of customer service AI is low right now, customers will become more accepting of AI in customer service as the technology becomes more commonplace, according to Gartner.

McIntosh maintains that companies also put too much emphasis on customer channel preference rather than what the customer truly wants. Some might just want the fastest service possible, regardless of channel. Opportunities to improve self-service can be found with less traditional self-service channels like mobile apps and IVRs, he says.

Along those same lines, many companies offer too many interaction choices, leaving customers confused about the optimal path to take, according to McIntosh. “So many potential options lead to a negative customer experience. The best practice is making a clear pathway and reducing ambiguity. This often comes in starting the journey in a single channel and then providing the path that leads customers to where they need to be. It’s all about clarity and customer confidence.”

McIntosh recommends that customer service organizations take the following actions:

• Position customer-facing AI as a facilitator of customers’ best-fit solutions rather than as an agent replacement.
• Ensure that customer service journeys that start in self-service are low effort by letting customers access humans agents when necessary.
• Use successful content on third-party sites to elevate the solutions created for customers.
• Drive more low-cost resolutions by promoting overlooked self-service channels like mobile and IVR.

And finally, it helps to know that customer sentiment and customer behavior can sometimes be two very different things, McIntosh says. “Be aware that customers have apprehension about [AI] technology and they’re willing to make a switch. It all adds up to a potentially high-risk, high-reward situation. Everybody knows the potential rewards that can come from AI, but if you get it wrong, there’s a potential negative consequence.”