On every long drive for a business trip or a vacation, chances are you’ll pass by hundreds of billboards along the way. Gas stations, restaurants, local attractions, and even the billboard companies themselves are vying for eyeballs hoping to catch the attention of drivers.
Often, bank marketing can seem like a billboard on a lonely highway. Broadcasting the bank’s need-of-the-month whether it’s mortgages, CDs or auto refinancing. Advertising to the masses has been the standard through billboards, newspaper ads or television commercials, but with the progression of artificial intelligence, there’s a new way to reach audiences who are in the market for your bank products.
For decades, propensity models have been one of the most popular tools in banking analytics. Those models try to help banks and credit unions answer an important question:
"Which customers are most likely to buy a product?"
Sure, that's useful, and, in many cases, it's far better than broad demographic segmentation or generic marketing campaigns. However, as artificial intelligence evolves, a new question is emerging:
What if likelihood isn't the most important thing to know?
Because while propensity models can estimate what a customer might do, they often fail to explain why. And in banking, understanding why the real competitive advantage is.
Propensity models are designed to predict probability.
How likely is a customer to open a credit card? Will they apply for a home equity loan? Will the purchase a CD? Will they even respond to a campaign?
By analyzing historical behavior and comparing customers to similar profiles, these models generate scores that help prioritize outreach. It does make marketing efficient, and they can improve targeting, and they do represent a significant improvement over traditional segmentation. The problem isn’t that propensity models are wrong; the problem is that they are incomplete.
Imagine two customers with the same high propensity score for a home equity loan.
On the surface, they appear equally attractive, but their situations may be entirely different.
Customer A recently purchased materials from home improvement stores and hired contractors. They're planning a renovation.
Customer B has experienced rising expenses and declining balances. They're looking for liquidity.
They have the same propensity score, but different motivations, financial needs and goals. Yet, many models would treat them as the same opportunity. What the models are missing is the context of their needs.
Banking isn't simply predicting transactions; it's about understanding people.
Every transaction leaves clues. A new payroll deposit may indicate a job change.
A series of childcare payments may signal a growing family. Increasing travel expenses may suggest changing priorities.
Shifts in spending patterns may reveal financial stress long before a customer asks for help. These signals tell a story, and stories create context. Context helps institutions understand not just what customers might do, but why they're doing it.
The next generation of banking AI is moving beyond prediction. It combines probability with understanding, behavior with intent, and signals with context.
Rather than asking, "Who is most likely to buy?", institutions can begin asking, "What is happening in this customer's life right now?".
That's a fundamentally different approach, and it's the foundation of customer-level intelligence.
When institutions understand customer motivations, they can deliver more relevant experiences, more relevant offers, more timely engagement, and more meaningful relationships.
Instead of pushing products, banks can solve problems and instead of reacting to outcomes, they can anticipate needs.
That's how deposits grow, retention improves, and how modern institutions create a competitive advantage that compounds over time.
Propensity models aren't going away. They remain useful tools, but the future of banking won't belong to institutions that simply know what customers are likely to do. It will belong to institutions that understand why customers behave the way they do.
Because probability identifies opportunities, but understanding creates relationships. And relationships are where growth happens.