Why Affirm is incorporating LLM tech into its underwriting

Affirm is changing how it decides which customers get approved for financing. The company is now using Large Language Models (LLMs)—the same technology behind ChatGPT—to refine its underwriting process. While these models are famous for generating text, Affirm uses them to predict consumer repayment behavior with higher accuracy. For retailers, this means Affirm is getting better at spotting 'good' borrowers who might have been declined by traditional credit scores. The shift to LLM-based underwriting allows Affirm to process vast amounts of unstructured data quickly. They are moving away from rigid, older formulas toward a system that understands patterns in spending and debt. This technology helps the lender adjust to economic changes in real-time. If inflation rises or employment shifts, the model catches those trends faster than a manual update. For your business, this translates to higher conversion rates without necessarily increasing the risk of defaults. Better underwriting means more 'yes' decisions for your customers at the point of sale. Affirm claims this tech helps them maintain low delinquency rates while keeping credit lines open even when other lenders are tightening their belts. Expect other BNPL providers to follow suit as they race to automate and improve approval odds.

Source: American Banker — Top News

← Return to the library· Submit a correction