Why Manulife got into banking with AI

Register now

When Manulife, one of the largest insurance companies in Canada, launched a digital bank this month, it leaned on artificial intelligence to help it attract the millennial consumers it sought.

It used a customized version of Kasisto’s AI-based virtual assistant, Kai, which answers questions and offers advice on spending and savings. Manulife named its chatbot Mai — a mashup of the name Kasisto uses for the technology and Manulife.

The rollout is an example of how digital-first banks can experiment with newer technologies like AI-based chatbots, learn from the experience, then roll the technology out throughout the organization. It also adds to the list of banks accepting the idea of letting a virtual assistant interact with customers, joining Bank of America (with Erica), Ally Bank (Ally Assist), Regions Bank (Reggie) and others.

Additionally, it’s another case of a digital bank trying to be a customer advocate. One of the first “insights” — or automatically generated pieces of advice programmed into the chatbot tells customers how to avoid the bank’s monthly fee. Manulife Bank’s new digital account comes with a monthly $10 charge, unless the customer saves $100 within the month. Every time a customer logs into their banking app, Mai reminds them how much more they need to save to avoid paying the $10 fee.

Not many U.S. banks would push out an alert to consumers on how to help them avoid bank fees.

“This is a new approach to consumers that we can build trust with them and hopefully they'll trust us with more and more of their business,” said Rick Lunny, CEO of Manulife Bank.

Manulife plans to add insights that will tell customers that at the rate they're spending money on their credit card, they’re likely to exceed their limit before the payment date, and ask if they’d like to make a payment or apply for an increased limit.

“All our products and services are designed to be in the customer's best interest, which is not necessarily the way traditional banking is done,” said Lunny.

Manulife is taking on the six large Canadian banks with what it believes are more consumer-friendly products.

“Because of the changing way of banking in use in Canada and the fact that technology is starting to become a lot more mature in terms of digital experiences, they could provide a much more compelling banking offering to Canadian customers than the traditional banks,” said Stephen Epstein, chief marketing officer at Kasisto. TD Bank is also a Kasisto customer.

“They’re using a digital-first strategy with AI to create this humanizing virtual chatbot to this demographic that wants to use technology to bank in the first place,” Epstein said.

Lunny said Manulife has an advantage over fintechs in that the company’s brand is widely recognized. About a third of Canadians use Manulife’s group benefit and retirement plans. Lunny wants to interest those clients in Manulife’s digital bank, and then ultimately to its wealth management offerings.

“The other advantage is that we don't have the burden of the cost of a branch network, and we don't have the burden of legacy systems,” Lunny said. “We can continually roll out enhancements so we're not standing still.”

Seeking millennials

Lunny hopes the simplicity of the digital bank will help attract younger customers to the 150-year-old Manulife brand. Focus groups helped the company make its user interface consumer friendly.

For instance, participants suggested that Manulife change the way it asks compliance questions in its banking app. Instead of suddenly hitting the user with a bunch of legalistic questions, the app now says, “Help us prevent fraud by answering these questions,” gently preparing the user for the onslaught of queries.

Focus group members also gave Manulife designers tips on how to give advice.

“One of the pieces of advice was, 'Don't tell us what's discretionary; let us choose what's discretionary,' ” Lunny said. “In the morning, my Starbucks coffee gets mandatory. We can provide relevant advice where people want the advice. That's the key to the human-centric design of this.”

Lunny said he expects Mai to draw a younger audience with its conversational interface and natural language understanding.

“That's why we chose Kasisto, we really liked the way that interface works with consumers,” he said. “You can say, how much did I spend on fast food last month in Toronto? Those are three questions.” Mai parses out an answer.

Customers can see where their money is going in the app, which Manulife plans to update every two weeks. It plans to keep adding insights that can be pushed out to the customer, such as warnings when it looks like someone is going to run out of money before their next paycheck.

The bot should also appeal to millennials because it lets them do everything through texting.

“The people we are going to attract, the last thing they want to do is talk to somebody — that’s the absolute last thing,” Lunny said. “If they have to talk to someone, they view that as a failure. And so we want to build the natural language understanding working with Kasisto and a conversational interface so that they don't have to do that. We feel that will be a strategic advantage.”

Customizing the chatbot to Manulife

Like any company that implements an AI-powered virtual assistant, Manulife had to tailor the AI engine to its business and locale. For one thing, it wanted to make Mai sound Canadian. If you chat with Mai, you will notice the engine ends many sentences with, “eh?”

The company also added, in addition to references to Starbucks, mentions of Tim Hortons, a popular coffee chain north of the border. The company had to add content about Manulife’s products into Mai, such as its mortgage and travel insurance.

In the future, Manulife plans to program different personalities for Mai. One might be for private clients, another might be for millennials.

Lunny said he doesn’t worry about Mai misfiring and giving a customer wrong information, because it’s set up to answer only simple questions.

“We're not trying to replace financial advisers,” he said. “It's simple advice: learning from your habits about how and where you spend your money.”

If Mai doesn't understand a question, it says, "I don't understand your question." Lunny plans to roll out Mai to all Manulife customers.

“We’re working with Kasisto on how do we achieve this in the insurance business or the group-benefits space where there's traditionally a call center interface, and we feel there could be some tremendous advantages in terms of improving the consumer experience,” Lunny said.

For reprint and licensing requests for this article, click here.
Artificial intelligence Machine learning Digital banking Consumer banking Digital Assistant Mobile banking Manulife