Breaking Stories

AI Adds Fuel To New Credit Scoring Data

There’s arguably no more important issue for financial service companies than customer acquisition.

Yes, it costs more to get a customer than to keep them. And yes, banks spend tens of millions of dollars on lead generation to find new business. But what if banks were missing out on opportunities that are already in front of them?

Part of the problem can be found in traditional credit scoring models — most obviously FICO — which rely on past payment history and other factors that don’t necessarily serve as accurate predictors of future performance. New technology, new software and most importantly new data are knocking on the door of redefining the criteria for valuable customer acquisition. It could be said that to fix the credit industry, the credit industry needs to fix the data.

That mission has been the focus of, said President and Founder Clint Lotz. The company offers a machine learning (ML) solution that allows lenders to leverage millions of dispute records and alternative data (and the company’s data set) through an application programming interface (API) and help predict future borrowing potential. That in turn helps offer better loans to customers. As Lotz told PYMNTS, the future of banking in the connected economy — particularly in extending credit — is more inclusive and high-tech driven.

That inclusive approach has been hobbled by data errors, which muddy visibility into a borrower’s creditworthiness. Expanding the data inputs and integrating the APIs into financial service firms’ existing infrastructure can help determine which negative credit items could be removed from a customer’s credit history, allowing for a better sense of creditworthiness — and extending offers to customers who might be otherwise be declined.

“There’s always been an ever-growing ecosystem of data,” said Lotz, adding the more data that can be considered in a lending decision, the better. “Maybe [certain bits of information] are not part of the decisioning process, but it needs to at least be acknowledged.”

Here in the states, he said, credit data can be problematic. That’s due in part to accuracy, but also to the fact that not everyone has a FICO number.

“Not every adult who’s working, holds down a job, pays her bills … has a credit score,” he said. “So not everybody can be decisioned or reduced down to that little three-digit number that makes lenders feel happy. We’re already moving towards using alternative data, where the increasing push into digital-first activities is creating reams of new, even real-time information that can help [artificial intelligence (AI)-driven] efforts to pinpoint errors in scores from the traditional agencies [which can prevent loan approval] as lenders interact with customers.”

Delving Into Dispute Data

He said that TrackStar’s data, spanning dispute data and other details, and access gleaned over 15 years enables firms to see patterns or trends across geographical locations and  income levels. He noted to PYMNTS that credit reporting efforts and regulations vary widely from country to country. But sourcing data from a global, online, reputable payment system or social media platform, “can give you probably a hundred times more data on that consumer than you could ever get from just their credit report.”

The gist of the ML component is that the advanced tech will identify erroneous items on someone’s credit report and determine whether certain, perhaps alternative, data points can be relied on just as heavily as other items on an existing report. In this way, a 680 score can be re-examined, and in fact, post-analysis, be a 720 — saving borrowers thousands of dollars over the life of a loan.

Advantages also accrue to the lenders using TrackStar, said Lotz.

“Since you were able to save them so much money compared to the other offers,” that consumer received elsewhere, they win new business. And then, he added, they can share the information with the consumer, illuminating how decisions were made (and even how individuals can improve their credit scoring).

“That’s totally different than just saying ‘you’re approved’ or ‘you’re denied,’” he explained. “Now, we’re having a conversation. And the only way to be able to do that at scale is through these AI tools that we have.”

That type of insight, said Lotz, can help lenders work out individual payment arrangements with their consumers, which can help in a time of economic stress (like the one we’re in now) — making sure that consumers can keep current on their loans and keeping lenders from losing money. Lenders thus are able to show empathy rather than simply seek to drive new business.

Looking beyond the pandemic, said Lotz, in the bid to widen credit availability, especially for first-time borrowers, point-of-sale (POS) financing remains an attractive option.

“It’s an easier, lower barrier of entry, to get into some sort of financing, and it’s usually for a much lower amount,” he said.

It’s a marked departure from the age-old way of building credit, where one might, for example, become an authorized user on a family member’s credit card.

“Through technology, we’re able to be more inclusive because no matter if you have someone who was born here, or just got here yesterday, it doesn’t really matter,” he said. “Both of them should have the same opportunity or the same tools available to them to build their credit, to build their identity, to build their history within our credit reporting system.”



About: Buy Now, Pay Later: Millennials And The Shifting Dynamics Of Online Credit, a PYMNTS and PayPal collaboration, examines the demand for new flexible credit options as well as how consumers, especially those in the millennial demographic, are paying online. The study is based on two surveys, totaling nearly 15,000 U.S. consumers.

What is your reaction?

In Love
Not Sure

You may also like

Leave a reply

Your email address will not be published. Required fields are marked *