Marketers everywhere are sick of hearing about the data-driven marketing prowess of Amazon and Netflix and Google, but deep inside everyone knows they want to — and need to — rise to the challenge.
Big tech and retail giants demonstrate daily the power of artificial intelligence software. Not the kind of incredibly human-like prototypes that get a lot of attention, but rather the AI that anticipates needs and makes relevant suggestions. It’s powerful stuff, and many financial marketers in the banking industry are working hard to figure it out.
The Financial Brand spoke with two very different institutions to get a sense of how long and difficult the road to AI in banking will be.
One thing is immediately clear: AI means very different things to different people depending on whom you speak with. A few would have you believe its most advanced capabilities are already being implemented, but the broader consensus is that the technology is still only in its infancy even though it’s been around in one form or another for decades. Nevertheless, the fundamental components in the umbrella concept of AI are well established, and financial institutions need to embrace them.
In many ways Citizens Financial Group, a $160-billion regional bank, seems quite far down the AI road. The institution has already scaled up a large team that includes data scientists to develop its data capabilities, according to Chris Musto, SVP and Head of Digital Strategy.
They are focused on weaving data from multiple sources together to yield a more robust view of consumers. This means drawing on internal and external data sets.
“Sometimes the best way to understand a customer’s needs is to go beyond their relationship with you and understand them more holistically,” Musto states. “That’s really challenging, but we’ve already put a lot of work into merging external data sources with internal data sources and structuring and staging the data.”
Is Citizens using artificial intelligence? Musto wouldn’t quite characterize what they’ve done thus far as true AI. “It’s really been good, sophisticated data analytics,” he explains. But he believes the institution is far along the curve of being at ready for AI when it finds the right opportunities for it.
( Read More: Success With AI in Banking Hinges on the Human Factor )
Inching Closer to the AI Goal Line
Union Bankshares, a fast-growing Virginia bank currently at $14 billion, is 11 times smaller than Citizens, and that affects where it falls on the data analytics spectrum. But the bank is not at square one, by any means.
Having spent 16 years at Capital One and three at Vonage, Union’s CMO Duane Smith is quite familiar with data-driven marketing insights. “Test-and-learn is in the DNA of those organizations,” he says. Smith says he has been using such techniques to enhance the analytical capabilities of his current employer for the past two years.
“With all things digital rising in importance, we hired a head of digital strategy who is helping us evaluate — among other things — the role AI can play in a bank,” says Smith. “But we’re early in that journey.”
“The next step is the addition of external data. Knowing a customer has a million-dollar IRA somewhere else would very much change the way we approach them.”
— Duane Smith, Union Bank
Everyone agrees the vast amount of consumer data now available and the ability to integrate multiple channels will be very helpful for both marketing and improving customer experience. But right now, too much information can bog you down. Smith and his team are concentrating on identifying the 20% of data that will give Union Bank 80% of the value.
“We already have good views of our customers, not just by products or as individuals, but by household,” notes Smith. But right now, the basis of that view is limited only to internal data. “The next step for us is the addition of external data. Maybe a customer only has a checking account with us, but knowing they have a million-dollar IRA somewhere else would very much change the way we approach them.”
BI, AI… and BS
Artificial intelligence implementation is hindered by ambiguous — even on occasion sloppy — terminology in the data science field. Musto calls it “nebulous.” There are many culprits, including vendors and consultants who sometimes use AI, data analytics, cognitive computing and machine learning interchangeably. There are countless sources of definitions of AI/ML.
Further, with the rise of AI, other data-related capabilities have been downplayed. “Some say business intelligence and data analytics are passé,” says Chandra Ambadipudi, Co-Founder and CEO of Clairvoyant, a data and decision engineering company. He says, however, that those functions are not mutually exclusive from artificial intelligence or machine learning. “They build on each other.”
“At the end of the day,” says Musto, “our focus has to be on providing highly personalized experience for our customers.”
The increase in storage capacity and processing power now available means that machine learning can be paired with very large data sets to generate advice and suggestions for consumers.
What Are They Actually Doing With Data Analytics and ML?
In addition to using machine learning for fraud prevention, Citizens is well along in using data analytics in various retail applications, including for its SpeciFI program, a digital “robo-adviser” for investments and wealth management.
“Our vision for AI ultimately is for customers to realize we’re being intelligent about looking out for them, calling out the important stuff and allowing them to spend less time engaged with us,” Musto explains.
“We don’t need customers to ‘engage’ with us more. We need customers to get more value out of us.”
— Chris Musto, Citizens Financial
“Vendors frequently say, ‘This solution would be a way to get customers engaged with you,'” Musto recalls. “But the best way to help people reach their potential is not to make them spend more and more time on their banking. We don’t need customers to engage with us more. We need customers to get more value out of us.”
That’s why Citizens does not measure digital product success by the number of times consumers log in. Rather success is achieved by making sure customers don’t leave Citizens, even if they move out of its market area.
“Why would you want to leave a bank that looks after you and looks out for you while you spend your time doing other things?” asks Musto.
Union Bank’s Duane Smith sees a great deal of potential for data analytics, and eventually for AI, to enhance customer experience. For now, however, Union Bank is using analytics for ad targeting and optimization — making sure the right products are presented to the right customer in the right channel.
“The channel piece is where things have grown more interesting,” says Smith. “We have the ability to target a specific customer or a type of customer across multiple channels. It’s no longer about just running a TV ad or dropping a piece of direct mail. It’s about following them in their digital shopping experiences. This is harder to do, but the data is there now to help make sure we’re connecting the dots.”
In one example, customer emails are provided to Google so that the bank can retarget ads to those customers as they search for, say, a home equity line. Or, if a person used the home equity calculator on the bank’s website, Smith says Union “can find them in other environments to be able to market that product right back to them, if we’ve cookied them appropriately.”
( Read More: Four AI Applications Banks and Credit Unions Can’t Ignore )
Next Steps on The Road to AI
With his big bank background, Smith is working to enable Union Bankshares to create a more personalized experience, appropriate to its level of resources.
“Personalization for us is going to be more of trying to serve up the right product to somebody — not necessarily a tailored product, but one that’s in the right category,” he observes.
He plans to begin to bring external data into the mix for this effort. In addition to data aggregators, which can provide some modeling based on geographic and other data, another approach is to use a data consortium in which member banks pool data anonymously to create a broader customer profile. Smith says Union Bank is investigating that option, but declined to mention with whom.
Musto envisions expanding a program Citizens uses for its branch and call center employees called Checkup. It involves having staffers work with customers to review their financial situation. By using artificial intelligence capabilities, the employees could initiate conversations with a customer based on spending patterns and other internal and external data that AI could track, much the way some fintech mobile apps do now. And, in fact, the service could also be provided on the bank’s mobile app. It’s part of the bank becoming a trusted financial advisor, Musto says.
“Ultimately,” he concludes, “you win as a bank by being able to fit your customer like a glove and being able to evolve with them, anticipate their needs.”