A credit decision engine is the software that decides whether a loan is approved or declined, how much is offered and at what interest rate. It is also called a decisioning engine or a loan decisioning platform. It is the place where bureau data, the lender's own rules and its scoring models meet.
The engine works in layers. First come the rules, the lender's fixed policies. For example: the borrower must be 21 to 58 years old, earn at least ₹25,000 a month and have no loan more than 90 days overdue. A borrower who fails a rule is declined at once. Next comes the scorecard. A scorecard turns each piece of data into points and adds them up. For example, five years with the same employer might add 30 points, and a recent late EMI might subtract 50. Last come AI models, which find patterns that a simple scorecard misses. The engine combines all three into one decision.
Take, for example, an application from a lab technician in Hubballi for a ₹1.5 lakh loan. The engine receives her credit report from the credit bureau, with a credit score of 742. It receives a profile from bank statement analysis, showing a monthly salary of ₹48,000 and existing EMIs of ₹6,000. She passes all the rules. Her scorecard total puts her in the second-safest of five risk groups. So the engine approves ₹1.5 lakh at 15% interest, for 18 months. A borrower in a riskier group would be offered a higher rate or a smaller amount. Setting the price according to the risk is called risk-based pricing.
The engine also checks for fraud, such as a synthetic identity, a fake person built from a mix of real and invented details. And it calculates how much more EMI the borrower can afford, given the EMIs she already pays.
The decision goes back to the loan origination system, which shows the offer to the borrower. The same engine can set the limit for a credit line on UPI or a buy now, pay later account.
Because the engine decides about people's money, its models must be checked. A lender must be able to explain why a loan was declined. It must also check that each model stays accurate as borrowers change. Model risk management (MRM) software tracks how each model performs over time. The general practice of judging a borrower's risk is called underwriting.