Overview
A machine learning based risk underwriting solution that predicts the probability of a borrower to default on a BNPL type of loan.
Customer Pain Point
In 2022, the BNPL (Buy Now Pay Later) segment was growing rapidly, but there was no standardized reporting or underwriting framework for these loans. Lenders lacked visibility into borrower risk, leading to potential overexposure.
Product Goal
As a Product Manager, I was responsible for building a risk intelligence solution from scratch to help lenders underwrite BNPL customers despite the absence of formal data infrastructure.
Approach
Defined the data foundation: Partnered with banks and fintechs to structure BNPL data and built a usable dataset for modeling
Shaped the ML solution: Worked with data science to define the core risk metric (early default probability) aligned to short-term lending behavior
Drove usability & adoption: Collaborated with sales/customers, simplified the model into an interpretable risk index, and improved real-world usability
Tested & proved value: Launched via a closed user group and built an ROI calculator to demonstrate loss prevention and support go-to-market
Impact
Generated ₹X.X Cr in early revenue from initial adopters
Received strong validation from early users on model effectiveness
Established a new data foundation for BNPL underwriting within the ecosystem