Few customers enjoy speaking with a bank’s interactive voice response system—and when you think about it, why should they? For starters, the systems have limited communication abilities and it can take multiple rounds of queries and responses for customers to get the information they need. What’s more—and perhaps more poignantly—the consumer themselves might have limited communication abilities: if English isn’t their first language, for example.
But new technologies hitting the market could offer banks expanded options to automate some customer service tasks in a more efficient manner. Through machine learning and conversational artificial intelligence, automated systems can now hold “human-like” conversations.
Of course, human-like isn’t human just yet. Nor can chatbots mind their manners like their flesh-and-blood counterparts, as evidenced by the infamous 2016 incident where an otherwise friendly Microsoft chatbot named Tay was taught to swear and spew racist remarks on Twitter.
A lot can change in two years. And today, big institutions beyond Microsoft are leading the way to a chatbot-driven future. Still experts say it’s not too early for small banks to dip their toes in the water.
“The hype surrounding chatbots remains sky-high,” Forrester Research concluded in a recent report. Bank of America, Ally Bank, and USAA have developed promising “conversational banking solutions” with chatbots, and from the boardroom to the watercooler there’s talk that the technology has the potential to revolutionize the banking industry.
And no wonder: Research from the 2017 BAI Banking Outlook revealed that among all sizes of financial institutions—community, super community, regional, super-regional and large—“improving digital capabilities” scored as their top priority for raising the customer experience bar.
And so enter the chatbot. Eventually.
The chatbot conversation
A chatbot is simply a technology that enables humans to better communicate with machines, says Ron Shalit, director of product innovation at Personetics. Chatbots can understand user requests and messages via text or voice and respond to them in a meaningful way; think of Apple’s Siri and Amazon’s Alexa. They work through a complex process called natural language processing (NLP) to identify intent from hundreds or thousands of options.
If you’re thinking, “Your call is very important to us, please continue to hold,” then hold on. NLP is a far cry from the antiquated interactive voice response (IVR) systems many small banks use to drive customers to yes/no answers, numerical responses—and tear-your-hair-out frustration. The Conversation surveyed more than 13,000 consumers and found only 10 percent were satisfied with their IVR experiences.
“[Past technologies] had a pre-defined flow and didn’t really let you conduct a conversation,” says Dror Oren, chief product officer at Kasisto. “These systems are now fielding open-ended questions where you can ask what you want, and they will understand you.”
While simple chatbots can offer account balances and due dates on payments, smarter bots will step up the game, Oren predicts. For example, data may indicate that a customer’s payment will be drafted in 24 hours but they don’t have enough money in the account to cover it. The bot could then contact the customer via SMS text, email or voice and ask if they’d like to move money from their savings account or defer the payment.
“That’s another layer of application that could be more proactive,” Oren says.
Chatbots, competitive advantage and community banks
So far, the dominant adopters of chatbots have been big banks. But through third-party providers, banks could possibly adopt the technology more easily than they think—and at moderate prices.
“There’s no reason a small bank can’t,” Oren contends. “You could argue that some of these smaller banks can be much more agile and could work with vendors more quickly and differentiate themselves by using this type of technology.”
In fact, chatbots arguably flip the conventional question—“Can we afford this new tech?”—on its head. Conversational artificial intelligence (AI) can help reduce costs and pressure on call centers while it enhances customer service, Oren says. And happy customers equal returning customers, which builds the bottom line.
Chatbots can also enable small banks to more easily scale up or offer new services, while helping call center agents predict customer inquiries, improve collections and detect fraud.
Varo Money, which launched last June and recently raised $45 million in private equity, offers FDIC-backed accounts and loans. The bank uses a mobile chatbot called “Val” that performs basic banking functions and serves as a digital money coach.
And outside of the U.S., digibank launched in India in April with an AI-driven virtual assistant that handles 80 percent of customer requests—without human intervention. The bank has already grown to 2 million customers.
Chatbot best practices
Many challenges stand in the way of full-scale adoption and banks should remain wary of bringing chatbots to market faster than their systems can support. Many bots aren’t yet ready to handle all banking tasks; thus they fail because of unclear purpose and poor planning, Forrester says. As data represents the single most important component of conversational banking services, a lack of data infrastructure and governance models could also create major issues at smaller banks. They’ll also need APIs (application programming interfaces) to access and control data, and talent to manage it all.
Yet that complex mix of requirements does not excuse a gradual sharpening of focus. Forrester recommends skipping generalist applications and moving toward niches that meet specific customer needs: “Even chatbots services from leading financial providers are often positioned for too broad a set of customer activities—which is one of the reasons why even the best bots still fall on their face occasionally.”
Several best practices can ensure that chatbots go over well with customers. Banks should first identify:
- who they want to serve
- define their business objectives
- build a strategy
- decide which services to offer, and
- choose whether to deploy a chatbot, human or combination.
Banks shouldn’t force consumers to use chatbots but rather deploy them as one channel in a omnichannel strategy, Shalit says. And the best bots proactively approach consumers with tips, advice and information.
“A chatbot that just sits there and waits for customers to ask questions,” Shalit says, “will only gain very small traction with the customer base.”
Want more Banking Strategies? Sign up for our free newsletter!
Craig Guillot is a business writer who specializes in retail and finance. His work has appeared in such publications as the Wall Street Journal, CNBC.com, Bankrate.com and Better Investing.
If you enjoyed this article, check out the webinar: Banking in 2018 - From the Customers' Perspective.