Churn Prediction and Fraud Risk Analytics
About the Customer
As a large insurer, Allianz Life Indonesia manages millions of policies and claims. Retaining customers and minimizing fraudulent claims are critical to protecting margins and maintaining customer trust.
Key Business Challenges
- Rising competition and price sensitivity driving higher churn risk
- Increasing volume and complexity of claims, making manual fraud review unsustainable
- Need for an integrated, data-driven approach rather than ad-hoc analysis
What We Automated
- Predicting which customers are at high risk of churn and flagging them early
- Surfacing key churn drivers so retention teams know what to address
- Scoring claims for potential fraud and routing suspicious cases to investigation teams
- Providing dashboards that combine churn and fraud risk for management
How We Solved It
- 1Designed and led multiple advanced analytics initiatives
- 2Built and supervised production Churn Prediction models integrated into customer management workflows
- 3Acted as Technical Project Leader for the Fraud Detection project, from model design to operational integration
- 4Embedded both solutions in an MLOps-ready framework using AWS cloud services and BI dashboards for business teams
"With predictive churn and fraud scores available in our workflows, our teams can intervene sooner and spend less time on low-value reviews."
Key Results & Impact
Churn Prediction models helped reduce customer churn rates by around 3%
Fraud Detection models improved the claim fraud pipeline and reduced the fraud rate by about 1%
Retention and risk teams can focus their time on the highest-risk customers and claims
Teams can intervene sooner and spend less time on low-value reviews
Ready to Protect Your Revenue?
Let's explore how churn and fraud analytics can protect your revenue and improve customer experience.
Start Your ProjectReady to Reduce Churn and Fraud?
Let's discuss how predictive analytics can help you retain customers and detect fraud before it impacts your business.