A new report from PYMNTS Intelligence and Thredd highlights a shift in credit underwriting from borrower-level evaluation to transaction-level analysis. Traditional models rely on historical data, such as past payment delays or revolving debt, which often fail to capture the economic context of specific purchases.
Artificial intelligence allows lenders to integrate real-time transaction and cash flow data into their decision-making. Instead of assigning a fixed risk tier to a customer, systems can now evaluate the borrower, the specific purchase, and current financial behavior simultaneously to determine creditworthiness.
This approach enables lenders to adjust repayment schedules, durations, and pricing dynamically based on the nature of each transaction. For instance, a customer might be offered revolving credit for one purchase and an installment plan for another, rather than being locked into a single product structure.
The report suggests that this architecture helps identify creditworthy borrowers who were previously excluded by conventional risk assessments. By focusing on the ability to support a specific obligation rather than general historical patterns, lenders can expand access to credit while pricing risk more accurately.
