Case Study

Optimized “Collections” for a Money Lending Company

Business Challenges

The collection department was running out of bandwidth and the client was considering to invest in additional collection agents but wanted to understand if technology could help

Our Solutions

SWStrategies built an ML (Machine Learning) model for the loans in collections with over 1000 features to analyze the loans
The model categorized the loans as “self-correcting”, “potentially-fraudulent”, and “collectable”
We built intelligence into our process which continuously file-tuned the model with a feedback loop

Results

Achieved 92% prediction accuracy
The collection department was able to prioritize focus to the “collectable” loans, reducing their workload by 50%
The collected amount increased by about 10% due to improved categorization and business focus

Technologies Used

Machine Learning
Python
AWS cloud

SWStrategies Difference

Thought Leadership
Unique technical solutions