In a short time, artificial intelligence has gone from fodder for Hollywood blockbusters to an everyday tool inside businesses of all kinds, including many community banks. These institutions and their vendors are putting machine learning technology to work in nearly all areas of the bank, from compliance to fraud prevention. Just as quickly, this once-futuristic tech is becoming a must-have for many banks.
By Judith Sears
This year’s widespread quarantining in response to COVID-19 has accelerated the online migration of both businesses and consumers. Professional teams now routinely meet via teleconferencing technologies, and consumers have also moved online. “Over 50% of transactions today are online,” says Aaron Lazor, CEO and cofounder of Finscend, an online bank dispute resolution platform based in Tel Aviv, Israel.
Changing practices are changing expectations, too. Community banks can’t afford to be left behind as businesses and customers move to new practices and develop new expectations. Leveraging more advanced technology is a must—not a nice-to-have.
Artificial intelligence, or AI, is the umbrella term for the technology wave that’s arriving. AI can be summed up as “making intelligent machines,” whereas machine learning, a subset of AI, focuses on machines’ ability to receive data and learn for themselves without being programmed.
What AI does, fundamentally, is provide faster, better data. And because community banks are voracious gatherers and users of data, there’s a use for AI in nearly every area of the bank.
Consider loan application and processing. Community banks faced the ultimate stress test with the Paycheck Protection Program (PPP), and they passed with flying colors. Of particular note is $2 billion-asset Cross River Bank in Fort Lee, N.J., which ranked third in the country in PPP unit volume loan.
To do that, Cross River Bank processed more than 198,000 loans, with the average loan size coming in at less than $33,000. “The fact [is] that we did it at a scale that surprised us—especially relative to much larger banks,” says Jesse Honigberg, the bank’s senior vice president and technology chief of staff. “Whether it’s with AI or machine learning, with the right combination of technology and business direction, this showed that you can really punch above your weight.”
For the PPP, Cross River Bank worked with more than 30 partners, including Divvy, Veem and Intuit, to build a platform that could interact with Small Business Administration (SBA) through the Cross River APIs, or application program interfaces. Integrating tech from New York City-based fintech Ocrolus into Cross River’s technology stack provided the capability to quickly review bank statements, extract payroll expenses, compare cash flow and spot variabilities. By collaborating with multiple partners, Cross River Bank had the capability for origination-to-funding visibility in its loan process.
Honigberg believes community bankers should drive technology discussions in collaboration with partners. To facilitate online account opening, for example, Cross River Bank worked with its core and another fintech, Mantl, to create a new solution. “We started with the problem that we want to be able to do high-volume online opening deposit products,” Honigberg says. “We brought business partners together to solve the problem.”
Community banks can also improve the pricing and profitability of their loan portfolios with the use of AI-powered tools. As with loan processing, there is an increasing number of vendors eager to work with community banks. “You don’t have to figure it out yourself. Start simple: Collaborate with Excel files,” Honigberg says. “There are companies that will help you do that. Small changes can make a big difference in how you process things.”
5 tips for getting started with AI
The pace of technological development is making artificial intelligence, or AI, more affordable and accessible. Community banks can take advantage of AI capabilities. Some first steps include:
- Partner with fintechs and other vendors. There are many small AI-focused companies with easily implemented solutions. For example, the dispute resolution platform of Tel Aviv, Israel-based Finscend creates application programming interfaces (APIs) to pull required information from wherever it’s stored. “There’s not a bit of integration on the bank’s side, and you don’t need a development team,” says CEO Aaron Lazor. “Our system is modular and easily accessible.”
- Structure flexible partnerships. Quontic Bank in New York City, for example, pays vendors a bit more per month in return for the option to quickly exit the contract. “That’s the risk mitigation for us to have the innovation,” says chief innovation officer Patrick Sells.
- Create a specific position with access to the CEO and board to identify and champion technology opportunities. “You have to have a platform for creativity to exist and thrive within a bank,” says John Owen, CEO of SimplyBank in Dayton, Tenn.
- Start with a business problem, not technology. “For example, we want to better understand small business bank statements. How can we figure out a way to do that at scale?’” says Jesse Honigberg of Cross River Bank in Fort Lee, N.J.
- Start small. “You don’t have to boil the ocean,” cautions Keith Henkel, CEO of FI Works in Little Rock, Ark. “You might just need to know what percent of your checking accounts don’t have debit cards.”
The value of preventing fraud
Before the pandemic hit, fraud and security were already two of the biggest concerns for banks. Today’s increased activity online has, unfortunately, come with increased risk. “With the pandemic and the shift to digital by both bank customers and employees, it’s really increased the attack surface for cybercriminals,” says Mike Venaccio, vice president of bank technology solutions for UFS, a community bank technology outfitter located in Grafton, Wis., that’s owned by and exclusively services community banks.
One of the most effective new cybersecurity technologies fall into the category of Endpoint Detection and Response (EDR). Venaccio explains that EDR looks at the behavior of programs rather than specific patterns or guarding against specific pieces of code.
“One of the great things about this technology is it’s low-cost and fairly easy to implement,” he says. “It’s a technology that can enhance the security posture of any community bank.”
In Dayton, Tenn., $361 million-asset SimplyBank started its ventures into AI-powered technology with an array of new security technology, including Alert Logic, a ransomware preventative. “Ransomware continues to be a real threat for most organizations,” says Pat Maloy, the community bank’s chief information officer.
Among other things, the Alert Logic agent protects these endpoints by intelligently blocking attacks with a unique combination of machine-learning attribute analysis and real-time behavior analysis. It will immediately shut down a laptop or a desktop if there is a malicious attempt on it.
SimplyBank pays a nominal amount of money per year for ransomware protection for close to 200 users. Indeed, it’s peanuts when you consider that Coveware, a ransomware incident response platform, reports that the average cost to recover from a ransomware attack was $84,116 in Q4 2019. “If you don’t have something like this implemented, you probably should,” Maloy adds.
“We’re using AI to analyze everything known about a data breach—the risks, the best courses of action to keep people safe—and we’re doing that in milliseconds.”
—Al Pascual, Breach Clarity
Breach Clarity, a fraud prevention and detection technology firm in Walnut Creek, Calif., has developed an algorithm that rates the severity of data breaches on a scale of 1 to 10. The tool provides a detailed assessment of how extensive the breach was and what types of fraud or identity theft are likely.
“Machine learning has its roots in fraud prevention and offers an incredible amount of value,” says Al Pascual, chief operating officer of Breach Clarity. “We’re using AI to analyze everything known about a data breach—the risks, the best courses of action to keep people safe—and we’re doing that in milliseconds.”
Compliance on the fast track
Compliance issues are often linked with fraud and security and are another area ripe for AI-related solutions. “They can improve your [know-your-customer (KYC)] process by an order of magnitude,” Honigberg says.
Earlier this year, Cross River Bank used Alloy, a compliance aggregation platform, to help fulfill its KYC requirements in rolling out an online account opening capability. “Over 80% of the accounts were opened with only minimal human intervention,” Honigberg says.
In New York City, $1.2 billion-asset Quontic Bank uses AI to monitor its community development financial institution (CDFI) status, which requires 60% of its lending to be to low-income borrowers, according to chief innovation officer Patrick Sells. “That is an enormous struggle to manage and report all this data,” he adds.
“[Through AI-powered efficiencies, we] provided 30% more mortgages to low-income individuals or those living in low-income census tracts in 2019 than in 2018.”
—Patrick Sells, Quontic Bank
Working with vendors and the community bank’s own IT development staff, Quontic Bank developed the capability to automatically check a number of criteria, such as property, zip code and income to determine if a loan qualifies as part of the CDFI target.
Further, Quontic Bank now has visibility into its loan portfolio along a variety of metrics. “I can look at my pipeline and know if I’ve qualified in general, as well as by area of the country or by loan officer,” Sells says. “That used to be a full-time job. Now, not only do we not have the cost, but we can manage our business so much more easily.”
The resulting efficiencies have enabled Quontic Bank to provide better service to low-income customers. “We provided 30% more mortgages to low-income individuals or those living in low-income census tracts in 2019 than in 2018,” Sells says. “This made a huge difference to the families who were able to get homes.”
How Encore Bank grew with AI
Encore Bank, a $620 million-asset community bank in Jonesboro, Ark., has successfully deployed AI-powered technology, developed by FI Works of Little Rock, Ark., to drive growth and exceed its strategic business goals.
Tonya Gossage, CEO of Gossage Performance Consulting, LLC, also of Little Rock, helped Encore Bank implement the FI Works software. She says FI Works’ Intelligent Targeting analyzes transaction patterns and identifies those prospects most likely to use certain products. More precise targeting saves time and resources.
In one initiative, FI Works identified top growth opportunities from a list of shareholders. Encore Bank assigned a concierge banker to each shareholder. Bankers contacted shareholders and discussed opening accounts with the bank. The campaign had an 89% acceptance rate.
That level of success motivates bankers, Gossage says. “When you are calling customers and nine out of 10 accept, the bankers are so excited because they are getting yes, yes, yes. That builds their confidence,” she adds.
Gossage also reports that the additional information better positions employees to be more personal and available to customers, and that personalized approach bolsters your bank’s reputation. “When you offer needs-based products, you can better be a trusted advisor,”
Encore Bank also found success when it achieved a 20% acceptance rate in a campaign for electronic statements. “We rely heavily on the FI Works platform to initiate campaigns, generate business and drive sales,” says Burt Hicks, senior executive officer and chief operating officer of Encore Bank.
Fueling customer relationships
The potential to create an enhanced customer experience and, therefore, increased loyalty may be one of the best benefits of AI for community banks. If you’ve ever used Siri or Alexa, you’ve interacted with a chatbot, the most visible consumer-facing use of AI. Third-party providers are making this technology available and more affordable for community banks.
But banks have many other, less-obvious customer touchpoints. SimplyBank deployed a risk-based analytics system for an overdraft protection (ODP) program. ODP software quickly analyzes customer behavior using direct deposit and other records to identify changes in the customer’s cash flow.
Quick and accurate decisions about increasing or decreasing overdraft protection provide better customer service while minimizing risk. “This is a very practical, simple example that’s available to every bank out there,” says Blake Swafford, senior innovation officer for SimplyBank.
Transaction disputes may be another area where community banks can quickly and easily deploy AI-powered tech to improve customer satisfaction. Finscend’s platform, for example, can review years of a cardholder’s data to provide context for dispute resolution. “Within a matter of minutes, a banker can see the percentage of likelihood that a dispute is valid and winnable,” Lazor says.
Finscend also creates a portal for the customer, empowering them to track the status of their complaint. That makes for a more satisfied customer while easing the bank’s burden.
The possibilities for AI applications in banking are only going to multiply. John Owen, CEO of SimplyBank, looks forward deploying AI technology that uses real-time data to enhance the customer experience and internal efficiencies.
“I’d like to put a system in place that can review a customer’s behavior and habits right when the customer steps up to a teller’s window or goes online,” he says, “and then make immediate recommendations to the banker to enhance the customer’s banking experience.”
The importance of community banks adopting AI is not that customers consciously select a bank based on whether it uses AI-powered technology. It’s the superior service that this technology empowers community banks to deliver.
“Where RPA can really impact the bank is in speed to market,” Venaccio says. “An AI/RPA-enhanced process can reduce the time it takes for a commercial loan or even a mortgage. If I’m a realtor and I know the community bank takes 45 days to close the loan and the bank across the street can close in five days, that’s going to influence my choice.”
Community banks have always emphasized strong customer relationships, and that won’t change with the adoption of AI and machine learning. Many community banks will—and should—embrace this new technology to meet customer needs and strengthen relationships.u
Judith Sears is a writer in Colorado.