Comprehensive AI-enabled credit policy outreach solution for one of India’s largest retailers, launching a check-out (POS) financing product based on transaction and alternate data.
Challenge
Our client was one of the top three retailers in India with sales across multiple formats, such as hypermarkets, small stores and fashion, with its own closed-loop payments wallet. The client wanted to launch a small credit-line product for its customers to provide a monthly credit line within its network shops ,payable through the payment wallet.
To further expand the business they wanted a credit policy that would use their customer’s transaction, wallet and alternate data for underwriting. The client was keen to maximize approval
Solution
To help the client with expansion, we integrated data from multiple sources, such as SKU-level transactions, invoice-level payments, wallet usage, loyalty programs and alternate data based on the mobile device.
Further, based on the transaction pattern and identifiability of customers, we shortlisted the master list of eight million customers.
We then hypothesized what retail behavior aspects can mimic credit risk signals and then created around 1,600 features around customer behavior. For are presentative sample, we collected past credit bureau information to build a series of scorecards to rank-order customers based on expected credit loss.
We pre-approved around three million customers for product launch, refreshed it on a monthly basis with a strong focus on credit monitoring to build the next generation models.
Impact
A significant high approval/pre-approval rate was achieved using a hybrid approach, whichcould not have been possible with the bureau-based approach given the low bureau penetration rate.
We were also able to target new credit customers for small-ticket loans, thereby increasing customer loyalty and overall customer sales.
