Fifth Third Among Banks Using AI to Help Lure Customer Deposits
Fifth Third Bancorp, Huntington Bancshares Inc. and Valley National Bancorp are among regional US lenders that use artificial-intelligence tools to scrape customer data, helping them personalize deposit offerings as competition for customers’ money intensifies.
Increasingly digital-savvy consumers are hunting for better online alternatives and higher interest rates, and banks, which use deposits as their main source of funding for loans, have had to adapt to win and retain customers who can easily make a switch.
“We have seen positive trends, especially in bringing new customers to the bank,” Valley National Chief Data and Analytics Officer Sanjay Sidhwani said in an interview. “Use AI models and see what customers might be in the market for.”
Valley National has used machine learning for about the past nine months to predict whether a customer would be a good target for a product, according to Sidhwani. The Morristown, New Jersey-based firm uses AI software to tailor online messaging for customers, and its data-analytics system flags customers who are deemed well-suited for certain accounts, he said.
“This is growing the depth of the relationships with the existing customers,” he said.
Fifth Third’s offering personalizes product and service recommendations using more than 100 AI machine-learning models, and has increased customer engagement by 40%, according to Shawn Niehaus, executive vice president and head of consumer banking at the Cincinnati-based company.
Huntington Chief Financial Officer Zachary Wasserman said in a Bloomberg Radio interview that the bank uses AI to understand consumer behavior and figure out when to reach out to them with products that would be beneficial.
A McKinsey & Co. report published last month recommended that banks use machine learning to predict which of its customers are suitable for their interest rates, whether customers might be prone to moving their money elsewhere or if rates are out of sync with competitors.
“The deposit space is vibrant,” McKinsey partner Roque Echaniz said in an interview. “The banks are fighting for the customer relationship.”
Personetics Technologies Inc. helps banks win deposit business with its AI prediction models. One top 10 US bank snagged 20,000 new savings accounts from its existing customers, according to Personetics, which declined to name the bank.
The software can also help banks target customers who have accounts at other institutions, thanks to open-banking practices, which let third-party providers access data from customers who have given their consent.
“They basically offer a more effective product to get their customers to move money from the other banks to their bank,” Personetics Chief Executive Officer Udi Ziv said.
Royal Bank of Canada’s partnership with Personetics helped create NOMI Find & Save, a tool that identifies existing customers with surplus money in their checking account. It helps them set up a deposit account and then sets aside money automatically, with the customers’ permission. Users on average move about $495 a month into savings products, according to 2024 company data.
Hot Money
Curinos is one third-party firm offering AI products that help banks tailor deposits. It uses data such as spending habits and whether accounts were opened online or in person, and pools survey responses. One of its clients, a firm with a market capitalization of $100 billion that it declined to name, used the tool and produced more than $1 billion in additional deposits from its existing customers, according to the company. Curinos currently works with six financial institutions with assets of at least $25 billion.
“Not all deposits are equal,” said Sarah Welch, a managing director at Curinos. “Are you going to bring me hot money, or money that is less valuable to me as an institution?”
Despite the initial success stories, some see pitfalls. Christopher Peterson, a professor of consumer finance law at the University of Utah and a former official at the Consumer Financial Protection Bureau, said AI algorithms could start generating discriminatory outputs if the machines learn off a biased data set. Or, he said, they may target vulnerable customers, such as those in banking deserts, with deposit accounts that have higher fees.
“I worry machine learning might also sharpen predatory marketing practices,” he said. “They could be disproportionately targeting people of color, people from disadvantaged socioeconomic backgrounds or rural communities that are struggling.”
Wasserman at Huntington said the issue is one the bank focuses heavily on.
“It’s probably the most important place we look before we ever institute any kind of machine learning or AI model,” he said. “What is the impact on the customer in terms of fairness, clarity, and adherence to compliance and regulatory rules.”
