Join The Discussion | Ai4 Finance 2019
September 1-2, 2020 | MGM Grand, Las Vegas
Traditionally credit scoring has been a fairly straightforward process where five key factors – including bill payment history, level of debt, age of credit history, the types of credit on the report, and the number of credit inquiries – are accounted for in a linear model.
While these rather simple and static models have done a sufficient job in the past, they are becoming outdated by models leveraging machine learning technology. These AI-based models, in many cases, have been able to reveal previously unseen connections between variables beyond the traditional five.
One of the major benefits of ML models is that they’re able to recalibrate in real time when access to new data, allowing updates to the credit scoring process to be made instantaneously and without the back and forth between credit experts and data analysts.
In some cases, credit issuers have been able to leverage new types of credit scoring data using machine learning techniques to open lines of credit to entirely new groups of individuals, which has major implications for both the credit issuer and the customer.
Here’s an example of how AI is being used for credit scoring:
Providing credit to the previously uncredited: There’s an estimated 1 billion people worldwide who don’t have access to credit cards due to their lack of financial histories, with around 30 million people in the U.S. These people don’t have the traditionally scored data points of bill payment history, level of debt, and age of credit history, so lenders are unable to evaluate them for credit. With AI, lenders can take into account data points such as smartphone habits and use of financial apps to determine a person’s financial stability. Beyond analyzing financial-transaction data, these AI lenders can consider whether you avoid one-word subject lines (implying that you pay attention to details) and ratio of photos in your smartphone library are selfies, indicating youth, so the lender can divide people into customer segments.