Case Study - AI-powered risk assessment for peer lending
FamilyFund is a peer-to-peer lending platform that needed smarter credit risk models to scale their operations while maintaining low default rates.
- Client
- FamilyFund
- Year
- Service
- AI Strategy, ML Engineering
Overview
FamilyFund came to us with a common scaling challenge: their manual underwriting process couldn't keep up with demand. Loan applications were taking 3-5 days to process, and they were losing potential borrowers to faster competitors.
We developed a custom machine learning model trained on their historical loan data to predict borrower risk. The model considers over 50 features including financial history, employment stability, and social verification signals unique to their platform.
The solution integrates directly with their existing systems, providing instant risk scores for new applications while flagging edge cases for human review. This hybrid approach ensures speed without sacrificing the judgment that complex cases require.
What we did
- Risk Modeling
- ML Pipeline
- API Integration
- Model Monitoring
Dogle helped us transform our underwriting process. What used to take days now happens in seconds, and our default rates have actually improved.
CEO of FamilyFund
- Faster approvals
- 95%
- Lower default rate
- 23%
- Loan volume increase
- 3x
- Annual savings
- $2.4M