Are chatbots mature enough to handle payments?
Today’s chatbots are best for straightforward banking tasks like changing an address, finding bank hours or updating payment card details. But there are limits as to what bots can do.
Although bots match up queries and provide suggested results as best they can, they can’t interpret information, understand pleasantries or small talk or solve complicated issues for customers.
“There’s a great place for them. You just have to be careful and not buy into the hype that they’ll replace all the people and provide a magic solution,” says Rurik Bradbury, global head of research and communications at LivePerson, a Manhattan-based provider of cloud mobile and online business messaging solutions.
Certainly, chatbots have a lot of promise. For starters, there’s substantial cost savings potential. The cost of a call center transaction can vary; industry experts peg it in the vicinity of $5 to $8 overseas and $12 to $15 for a U.S.-based call center.
And many financial services companies including PayPal, Lydia and RBC are experimenting with their own chatbots on commonly used platforms such as Slack and Facebook Messenger.
Done right, chatbots have the potential to reduce call center activity by at least 15 percent, says Mahi de Silva, co-founder and chief executive of Botworx, a Palo Alto, California-based maker of AI driven technology.
“It’s going beyond what you can do with a ‘like’ or a ‘follow’ experience and making it something that leads to engagement of the consumer and a transaction,” he says.
Chatbots also help simplify the user experience. In banking, for instance, there can be significant cost savings by using bots to help customers perform routine tasks like checking their account balances. There are also aggregator bots such as MyKAI that allow customers to access information from multiple financial institutions without visiting each website or using multiple apps.
“It’s a whole new world,” says Rena Nigam, president of global solutions and services at Incedo Inc., a technology services and outsourcing company based in Santa Clara, Calif. “It’s got promise, but there’s still a long way to go.”
Customers can get frustrated when interactions with bots don’t go as expected, so companies have to find a way to build in “handshake moments” where customers are efficiently transferred from bot to human, says David Hewitt, global mobility strategy lead at consultancy SapientRazorfish. “Part of the responsibility in crafting [chatbot] experiences is to know when and how to hand off to a person,” he says.
Bradbury of LivePerson suggests the ideal process should be similar to a group chat, where the brand’s bot takes on menial work and queries, and the brand's human agents take on more complex parts of the conversation—all within one thread. That’s what LivePerson is trying to accomplish with a bot it plans to launch soon for Royal Bank of Scotland.
LivePerson research also suggests that transparency is critical; 80 percent of consumers polled by the company want to be told clearly whether a human or bot is on the other side of a conversation.
Ultimately, every customer service strategy has its limitations, and the challenge is to find ways for bots and humans to work together effectively, says Eran Livneh, vice president of marketing at Personetics, a fintech company that provides an AI-powered bot to financial institutions.
“There are things that bots can do better than people. There are things that people can do better than bots,” he says.
“It’s not an all-or-nothing proposition,” says Beerud Sheth, the chief executive of Gupshup, a chatbot maker in San Francisco. “You would never ask employees to take on more than they are capable of, and companies shouldn’t expect bots to do that either. You want to position your employees and your bots for success.”