Humanizing FinTech #5: How can AI help lenders deliver customized products based on personalized underwriting?

Shamolie Oberoi   /    Content Marketing Specialist    /    2022-11-03

LinkedInLinkedIn

Here’s a quick question. How much has your life changed in the last six-twelve months?

Maybe you got that raise you’d been eyeing. Or a better paying job. Maybe your expenses have shot up because you bought a new car, or moved into an expensive neighborhood.

You get the drift - financial circumstances are dynamic. Everyone reading this is most likely in a different financial position than they were a year ago, be it good or bad.

And while this seems rather obvious, underwriting technology hasn’t quite caught up to the notion.

As lenders - banks, NBFCs, and FinTechs - focus on pushing out their next cutting-edge product, underwriting innovation falls by the wayside. And that’s why, they’re losing out - on 5-15% more revenue,  20-40% higher efficiency, and have to suffer 20-40% lower credit losses.

As financial products evolve alongside consumer expectations and lifestyles, so must the models that any sort of lending decision-making is based on.

Dynamic underwriting that keeps up with changing borrower circumstances doesn’t just improve the quality of lending decisions. It creates opportunities for lenders to serve their customers in different ways, be it offering personalized products or optimizing lending terms on-the-go.

This would involve leveraging Artificial Intelligence (AI) and Machine Learning (ML) models that can:

  • Track data points in real-time - cash flow, location, etc

  • Minimize human bias

  • Self-learn based on continuous data inputs

What can lenders do with dynamic, real-time underwriting?

Customize product experience 

Gamify the customer’s engagement journey to upsell 

Access to behavioral data opens up a plethora of opportunities for lenders to upsell to customers. Simple challenges and milestones based on your products can be used to unlock ‘rewards’ of sorts, such as a higher credit limit or added credit card benefits.

Take for example, Emirates NBD. In an effort to increase deposits, the bank decided to reward customers' active lifestyles with higher savings rates. Customers were thus required to open a special fitness account with the bank's mobile app, sync the app to compatible fitness devices and achieve daily goals expressed in a number of steps. With 12,000 steps per day, customers could get a 2% interest rate.

Personalize collections messaging

Risk assessment is a continuous process that brings immense value even at the collections stage. Risk and prioritization buckets based on real-time cash-flow data give lenders insight into the best time to communicate with a borrower (E.g. right after the borrower’s salary is credited). Behavioral data can also provide insights into the kind of messaging that will work best, and on which platform.

Of course, the process involves a fair bit of A/B testing to refine and pin-point what kind of messaging works for each customer cohort. For example, a message like “Salary credited? Why not pay off your EMI for the month!” may put off some users who could feel like their privacy is being violated. On the other hand, some borrowers may appreciate the casual and friendly tone of the message.

Offer dynamic credit products

When you have the right data at the right time, your products become living, evolving solutions that adapt to changing borrower circumstances. Take for example a credit line product that consumers use with their favorite food and grocery delivery apps. Once you’ve observed their repayment behavior over a certain period, you can tweak the credit line accordingly. Regular repayers get a higher limit, while those who have been irregular can have their limit reduced.

Cross-sell relevant products

AI and ML models that pick up on borrowers’ behavioral patterns can help lenders recommend products relevant to them at any given time. For example, if an existing customer just bought tickets to travel out of the country, that becomes the perfect opportunity for a notification on a travel insurance offer. If a borrower has recently spent money at a clinic or hospital, that could mean they’re in need of an emergency loan.

Insights like this allow lenders to move from being simply service providers to allies in their customers’ financial journeys.

Conclusion

According to McKinsey and Company, “banks follow a product-centric approach to credit models, only analyzing the data relevant to that product. A customer-centric approach, which combines the data signals from all product areas where the customer interacts, nearly.”

That succinctly sums it up. Dynamic, self-learning credit models allow lenders to move to a customer centric approach that puts borrower wellbeing front and center without compromising on business objectives.