Some banks customers are already using virtual assistants powered by artificial intelligence to ask questions like “What’s my account balance?” without having to look it up themselves online or on a mobile device.
While strides have been made in natural language processing and machine learning, many chatbots still must hand off to a person when the conversation gets too complicated. But some banks are looking to virtual assistants to be all-in-one bankers: equal parts teller, contact center representative, investment adviser and mortgage broker.
“What we are doing, compared to many competitors globally, is using [a virtual assistant] in a much more proactive way,” said Peter Dahlgren, CEO of Nordnet,a digital-only bank with customers in Norway, Sweden, Denmark and Finland.
This fall the bank will roll out a virtual assistant called Amelia, powered by the artificial intelligence company IPsoft. Some telecom companies are using IPsoft’s technology already to analyze emotions in customer calls.
AI in overdrive, on overtime
“There are ... some things that humans do better than AI,” but "the beauty of Amelia is that she never sleeps," says Nordnet CEO Peter Dahlgren in describing the bank's ambitious new virtual assistant.Nordnet, based in Stockholm, was founded in 1996 partly as a reaction to the outdated legacy structures in banking, Dahlgren said. He said Amelia is a natural extension of that mission in an industry that is becoming more automated.
“We will use her to make the onboarding process easier — she can guide you through it in three minutes,” he said. “There will be less dropouts from the process because customers who may get stuck won’t have to stop” and call someone.
But Nordnet has even bigger plans for Amelia, Dahlgren said.
“We are going to use her for moving securities and trading stocks,” he said. “We are hoping to be the first bank in Europe to provide investment advice entirely through AI.” Amelia will be deployed in issuing mortgages, too, he said.
Though Dahlgren acknowledged humans will still play a role in many relationship-building aspects of banking — such as catering to high net worth clients — banks can operate much more efficiently with “digital employees” such as Amelia in many areas.
“There are of course some things that humans do better than AI,” he said. “But a lot of the standard queries and advice can be done through Amelia. And the beauty of Amelia is that she never sleeps.”
Still, some U.S. banks that are experimenting with virtual assistants say that while the technology holds much promise, its role will be to complement — not eliminate — human labor.
“Until AI looks and sounds exactly as humans do, I don’t think it will ever entirely replace the human factor,” said Darrius Jones, assistant vice president for enterprise innovation at USAA, which recently rolled out a service on Amazon’s Alexa where its customers can ask about things such as account balances and spending habits.
“What’s important for us is having that great digital experience along with human interactions,” he said. For example, Jones said, situations where empathy and understanding are involved will always necessitate person-to-person conversations, such as that with an investor worried about a major global economic shock such as Brexit.
“I think there are some financial conversation where you’d be more comfortable talking to a human and others where you could talk to an AI,” he said.
Different geographic expectations when it comes to banking could also come into play; Dahlgren says the Nordic region “is very advanced when it comes to the banking system and customers have a high level of digital expectation” and thus may be more comfortable conducting more complicated financial conversations with virtual assistants.
But any preference for human interaction still doesn’t mean AI can’t — or won’t — have more complicated conversations, said Hari Gopalkrishnan, client-facing platforms technology executive at Bank of America, which plans to deploy its virtual assistant, erica, later this year.
“So instead of asking [the AI] how much did I spend on groceries last week, perhaps it could proactively notice you spent more on groceries last month than previous months, [and] here are some tips to help you save,” he said.
Gopalkrishnan agrees with Dahlgren that AI can already shift much of the mundane interactions away from humans, thus freeing them up for more high-touch interactions.
“I was in a branch the other day and someone stood in line to reorder checks,” he said. “You can do that on the mobile app, but maybe they didn’t know that or didn’t know where to look. But with [AI] they can just say ‘erica, can you reorder me checks’?”
Ultimately, success will depend on banks' ability to connect data on the back end, said Emmett Higdon, director of digital banking at Javelin Strategy & Research.
“Banks haven’t knitted together data sources on the back end well enough,” he said. “There’s still too many silos that don’t talk to each other. That’s why you can use [AI] to give consumers fairly simple spending information, but when it comes to advice most fall short.”
Higdon said banks like Nordnet that grew up digitally, or perhaps a fintech granted an OCC charter, would find it easier to offer advanced virtual assistants since they do not have the decades of disparate data to aggregate.
“They will have fewer silos for sure, but what’s really important for any bank [creating a virtual assistant] is real-time data feeds,” he said. “When customers are dealing with a chatbot or an AI, they are looking for real-time answers. They don’t want to sit there watching those little dots waiting for an answer.”
But regardless of how advanced they become, having some kind of chatbot or virtual assistant capability will be essential for every bank in the near future.
“Within the next five to seven years this will be table stakes,” said Steve Ellis, executive vice president and head of the innovation group at Wells Fargo, which is among the banks that have created a Facebook Messenger chatbot.
Ellis said Wells, like Nordnet, is exploring ways AI can eventually aid customers in more complicated financial tasks, such as investing.
“As they get used millions of times, we’ll see the evolution of natural language processing tools,” he said. “It’s not unlike the evolution of the Internet; it started with simple things and went from there.”
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access