Significant improvement in accuracy over the existing solution
Substantial annual revenue retention
Early identification of churn signals for proactive action
Scalable across markets
A leading global energy company’s B2B Mobility team faced increasing customer attrition across seven key markets. Internal models lacked predictive accuracy and lead time, resulting in missed retention opportunities and inefficient account management.
Key challenges
Short action window
Low prediction accuracy
Lack of explainability – no actionable insights
No structured value tracking framework
The solution
Customer360 Platform
Customer360 platform with 400+ attributes in Databricks Feature Store
Unified internal and external data for enriched customer insights
Designed for scale and future evolution into a recommendation engine
Predictive Modeling
Created three XGBoost models for Large, Small, and Micro customer segments
Used dynamic segmentation to boost accuracy and reduce model maintenance
Predicted volume decline over the upcoming period and grouped customers by churn risk
1
Prioritization framework
Focus on high value
Segment by risk
Use sales trends
2
Explainability with SHAP
Portfolio-level drivers
Customer-level insights
Transparent predictions
3
Power BI dashboard
Holistic customer view
Early risk signals
Self-serve analytics
Business outcomes
Significant improvement in accuracy
Substantial cost savings
Positive market feedback
Strategic benefits
Proactive retention model
Scalable across geographies
Explainable AI insights
Adoption drivers
Win-rate tracking
Test-control validation
AM performance metrics
Expand models to additional markets with automated retraining
Integrate Next Best Product, CLTV modeling, and share-of-wallet estimation
Introduce Agentic AI-driven hyper-personalized insights for 1:1 customer engagement
