Credit unions and banks say their business is a relationship business.
So how does artificial intelligence, which sounds impersonal and cryptic, enter into financial services?
Actually, AI using machine learning can help credit unions and others cultivate more of the right relationships — faster. That could simply mean signing up more ideal new members and then getting to know them well enough to understand how and when to provide them greater value and service.
But most financial institutions aren’t ready. Sure, AI is being discussed at the board or executive team level at half of all financial institutions, according to The Financial Brand. But only one in five bankers feels their institution has the necessary data analytics skills.
AI might sound like a mystery, but it can be simple. AI can help match financial institutions with their quintessential customers while reducing risk.
AI can help target potential new members and connect them with services. Credit unions already have a healthy understanding of their membership. AI can investigate that relationship further. Credit unions can use AI to learn the patterns in their current membership data and build profiles of their ideal members.
What are the qualities of loyal, longstanding members? Which members have also taken out loans? What do these members have in common?
From there, credit unions can use AI to predict which potential new members share the attributes of their ideal existing members and then actively market directly to this group.
Another possibility is using AI to target consumers who have recently experienced life events, such as buying a new home, getting married or having a baby. That’s because most of the time when people switch banks, it’s after a life event, according to 2015 research by Oliver Wyman and AOL.
Companies across industries have long deployed these so-called look-alike strategies of crunching data to identify their top customers. See, not so mysterious. The difference today is AI can go through a lot more data, identify a lot more patterns and ultimately build richer consumer profiles.
AI can also reduce the risk of fraud for digital account openings. Digital is a big push for banks and credit unions, and rightly so. Consumers appreciate the ease of opening new accounts anytime and from anywhere.
Therein also lies the problem. Financial institutions offering digital account set up must ensure consumers are who they say they are and protect against account-takeover fraud.
AI can help by “learning" the patterns from data in devices, transactions, digital channels and other applications that indicate fraudulent or risky activity. As consumers sign up or sign in, these algorithms can use this data to score the risk of fraud. Passive AI monitoring is a powerful complement to device-level security, such as
What classifies a borrower depends on the lender’s priorities and risk tolerance. AI algorithms can recommend approving or denying borrowers on either extreme and identify key drivers to help explain the recommendation. This allows a loan officer to spend more time understanding the cases in the middle.
Much like training a new employee on internal policies, AI is capable of learning from a lender’s past decisions. The AI algorithm can detect patterns from the credit union’s historical car and auto loan decisions. It can then score loan applications based on the application data, financial statement data, credit reports (especially the underlying tradeline data) and other data.
Moving from high-touch, rule-based decision to a predictive model can prove beneficial for lenders.
AI can help credit unions manage risk within their loan portfolio after loans are made. Today, AI software can analyze real-time economic data and borrowers’ past behavior to identify signs of risk up to 90 days in advance. These AI models can predict risk with greater accuracy than traditional financial ratios and benchmarks. Instead of waiting for a missed payment or broken covenant, lenders can use these early warning systems to act before default.
Credit union relationship managers, for example, can reach out to at-risk borrowers to assess the risk in greater detail or work out alternative payment schedules. Internally, lenders can manage this risk by adjusting loans’ risk ratings and controlling additional exposure.
Finally, AI can help credit unions match members with the right kind of credit card and the right kind of offer message. Credit unions can use AI to identify patterns that indicate which current members are more receptive to a card offer. This might be members with high monthly deposits, high monthly withdrawals, a wide range of withdrawal amounts per month and members without many standing bill pay or transfer orders.
A credit union can then cluster its target members into similar groups and determine the kinds of cards most relevant. For example, those who spend more on vacations might value travel rewards cards and business owners might value offers for corporate credit cards.
Bottom line: Artificial intelligence can be accessible and have beneficial implications for financial companies, including credit unions. It’s still about forming relationships. It’s just that AI helps you identify the relationships you should be building right now.
