“Can I afford a Tesla? If not, can you help me save for one?”
These questions were once reserved for financial advisers. Technically they’re still questions for a financial adviser, but nowadays that adviser may be a bot. Artificial intelligence has come a long way from simply answering balance inquiries. It’s even going beyond the money transfer function, doing more than just moving money: It’s also managing it.
According to Kasisto Chief Product Officer and Co-Founder Dror Oren, financial illiteracy is at an all-time high, making it more important than ever for banks to reach customers wherever they happen to be, using language that’s easy to understand. Oren believes it’s not enough anymore for banking bots to answer basic questions or perform basic tasks. Modern virtual assistants, he said, need to be multi-skilled, offering contextual and actionable insights with educational value.
“Two-thirds of Americans cannot pass a basic financial literacy test,” said Oren. “There should be a better way to make financial information more accessible, digestible and actionable. That’s where conversational AI can help.”
That’s the gap Kasisto wanted to help banks fill when it built its conversational AI platform, KAI, Oren said. The virtual assistants and bots powered by KAI can decode jargon, make recommendations, enable people to take action on spending and account data it provides and even help save up for that Tesla.
Banks adapt the KAI bots and assistants to fit their brand, but Oren said the one thing that stays the same across the board is KAI’s trustworthy, authoritative persona — and the carefully crafted sense of the bot being not “just a bot,” but also a customer advocate. Banking technology that instills in customers the feeling their financial institutions are for them rather than against them goes a long way toward inspiring customer loyalty, Oren said.
Kasisto took a two-pronged approach with KAI. First, the platform focuses on giving actionable insights to help people manage their money and increase their financial well-being. Sending a payment to Comcast or checking account balances is good, but even better? Letting a customer know his Comcast bill has spiked, or notifying him that his checking account has gotten low and asking if he would like to transfer funds from a savings account.
KAI is also trained to make contextual offers. Did the customer just make a major purchase at Best Buy? If her bank’s credit card offers a discounted extension on the warranty, KAI will let her know about it and ask if she would like to take advantage.
Or maybe the customer bought plane tickets from JetBlue. KAI might then ask if he or she wants to also purchase travel insurance. Depending on the cost of the tickets, the AI banking tech might even surmise the customer is traveling internationally and suggest applying for a credit card that will not hit him with any foreign fee charges.
The second prong is education around financial literacy. The average person doesn’t understand finances, Oren explained, and trying to make sense of it on his or her own can be discouraging or even downright intimidating with alienating jargon. Oren believes a good bot doesn’t just do the work for customers, but also helps them navigate the financial weeds and imparts best practices and knowledge along the way.
That’s why Oren said Kasisto felt strongly that the KAI platform be able to make suggestions to encourage people to act in ways that will improve their financial well-being. For example, it may inform a customer that other people like him with the same demographics are saving, on average, 20 percent more than he is. The bot will then ask if the customer wants to increase his savings based on that information; and, if the voice command is yes, KAI will suggest ways to do that. The bot can even help customers save by setting goals like not spending more than $200 a month dining out and informing them when they are getting close to that limit.
Information is not the most valuable asset here, Oren said. The value lies in how that information is being used — how to make it actionable. Bank customers already receive static information via bank statements sent either by mail or email, but both are going largely unread, as “historical” records then are not terribly actionable.
On the other hand, he said, when those same people receive a contextual nudge from an AI banking bot — like after a major purchase or the tenth $4 latte that week — they can take action in the moment: like purchasing the extended warranty or making coffee at home.
According to Oren, banks have a ton of data on their customers, yet they’re leveraging very little of it to the customer’s benefit. There are certainly FinTech companies out there trying to make the most of their customers’ financial data as they help them save and invest, but those FinTech apps are starting from scratch without years of checking, credit card, mortgage and loan information about their customers. On the other hand, Oren argues no one is better positioned to take the data-driven, customer-centric approach than the banks themselves.
On the bank’s side, when they leverage conversational AI and deploy a banking bot, it increases efficiency, reduces costs and streamlines the customer experience. Oren reported that Kasisto’s customer, DBS Bank, has a KAI-powered assistant in their mobile-only bank, digibank, handling 82 percent of their customers’ requests that would have otherwise been dealt with by live agents. Of course, there are cases, like reporting fraud, when a customer should speak to a person, so in the case of digibank, the remaining 18 percent of requests are handed off to live agents.
Despite advances in technology, Oren noted that some banks seem stuck on very specific, one-trick voice command bots focused on customer support. Going forward, he believes banks will need to take a more holistic approach, incorporating literacy as well as brand loyalty, and even upselling.
“Banks are quickly understanding that it’s up to them to educate [consumers], because educated customers save more, and that’s what banks want you to do,” Oren said. “It’s easy to make a bot that tells you about the weather, and an entirely different league to make a bot that can give solid financial recommendations — but that’s where banks need to go, and the AI-powered bots need to support it.”