About our client

Our client specializes in providing low-ticket size home loans to the priority sector in India. This is a relatively unbanked segment of the Indian population — with varied income streams and fluctuating income levels. At the same time, the volume of transactions runs into millions, making this a massive market with enormous potential. 

Why they come to us

Since Priority Sector Lending in India is extremely complex and diverse, assessing and managing credit risk becomes very difficult. The client knew that they would have to use AI in credit risk management and bring down their default rates.

The problem

The default rate for our client’s portfolio stood at a whopping 6%. This translated into NPA (Non-Performing Assets) worth USD 3 million. In order to grow and reach the desired scale, our client needed to control this percentage in a big way. 

our strategy

Our business-tech experts, at Valiance, decided to use Machine Learning to build a platform that would detect the high-risk applicants during loan origination itself. This would ensure that the client would be able to have a comprehensive credit risk assessment before approving the loan. 

the implementation

The Risk Detection platform generates a Default Score between 0-100 — the higher the score, the greater the likelihood of default. We knew that the platform would have to be able to generate a Default Score that was not only comprehensive but also dynamic. 

The platform:

  • Ingests 200+ structured and 30+ unstructured consumer application data points.
  • Includes Machine Learning algorithms. These are built on more than 3 years of historical loan portfolio data and have been tested across different population segments.
  • Has robust Machine Learning algorithms that monitor the data periodically and identify any new risk patterns.

the transformation

  • We reduced the default rate by an incredible 66% — from 6% to 2% — in just 18 months
  • With credit default coming under control, the client was able to increase loan disbursements by 10% — to USD 55 million.