Leveraging AI/ML Models at State Bank of India for Loan Approval

Organizer: INFORMS Practice Section

## Analytics-Based End To End Digital Pre-approved Personal Loan For Non-Salaried Customers at State Bank Of India

The State Bank of India is the country’s largest Bank by customer base and network. It was established by the Government of India in 1955 In March 2023, the bank launched an end-to-end digital pre-approved personal loan product for non-salaried customers, using AI/ML-based credit risk modelling. Prior to this, non-salaried customers did not have a consistently maintained personal loan product. To enable this, an AI/ML model was built using customer transaction behavior data, cash flow characteristics and demographics to predict credit default probability. The methodology used was Two-Step Clustering, to handle the large volume of data and the lack of substantial credit history. Further, clustering was conducted twice, with riskier clusters being eliminated in each phase based on delinquency concentration and expert insight. The findings suggest that cluster analysis can effectively flag credit risk even in cases with complex and inconsistent cash flows.

This approach reduced the processing time, manpower cost, and product risk. This was achieved while extending formal credit to a previously underserved customer base. During the period March 2023 to October 2024, the product had (created) additional loan disbursements worth ₹ 8,400 Cr. (USD 1 Billion) spread across 5,36000 account holders, which were all sanctioned digitally, with sanctioned digitally, with no branch visit or manual intervention.

Speakers:
Divya Gopalkrishnan Nair, SBI
Komaragiri Srinivas Jagannath, SBI

Wednesday, March 12, 2025
India Time: 6:30 pm India
US Time: 9 am Eastern

Link: https://us06web.zoom.us/webinar/register/WN_HrJsMr5HQwikTkekSqZPRg#/registration