Helped a lender become market leader through AI transformation
Challenge
The client was a major equipment financing company which specialized in construction and mining equipment. However, their business processes were extremely manual and process driven – which impacted their agility. The client wanted to leverage the huge data they have collected over time to create a business advantage.
Solution
We assessed the clients existing processes to identify three key areas for AI transfomation: better underwriting and pricing; better collections; and better reporting. Our entire solutions was focussed on delivering on these three key pillars
For underwriting, we used historical application data, credit bureau data, and loan repayment data we created a series of application scorecards for various segments. Along with these application scorecards, we also created an asset rating model. Asset rating model leveraged historical asset resale/ auction data to determine the quality of equipments and their ease of resale. We brought these two models together for a new risk based pricing model which was able to offer differential pricing based on the two risk models. FInally, we created an API based pricing engine which was able to provide pricing decisions within an online journey.
For improving collections, we built a series of behavior models and collection models. The behavior models focussed on current accounts and helped in identifying accounts with high probability of default. These suite of models were leveraged for early collection efforts. For already delinquent accounts, separate collection models were built focussed on prioritizing late/ hard collections. The entire suite of models were implemented on FlowXpert business rule engines and integrated with lender’s own and third party collection system. Feedback from both the collection systems were collected to make the algorithms self learning.
For better reporting, initially an assessment was done to understand the existing landscape of reports. Based on the initial assessment, prototypes were created for various reports using static data to ensure all the stakeholders are aligned on their requirements. Post requirement finalization, all the reports were created on Tableau and distributed throughout the organization in mobile as well as desktop through active directory management. To ensure adoption, multiple trainings were held across the organization and usage reports were tracked and analyzed.
Impact
With new underwriting models, the client was able to increase their market share of new loans by 3.5% in an expanding market while keeping the risk profile same. The client was able to attract newer customers by offering lower rates to less risky profile compared to their competitors. The client also approved some portion of previously rejected customers but at a higher price.
With collection models, client was able to reduce the collection leakages by 5%.