March 2025
March 2025
We are introducing a cutting-edge machine learning model designed to enhance ABSLI’s customer retention by predicting policies most likely to lapse due to non-payment of the first-year premium. By identifying these high-risk policies in advance, we can proactively engage with policyholders, offering tailored solutions to prevent lapses and boost persistency year after year.
Behavioural & External Factors: External influences like health changes or major life events affect lapsation, making them hard to model.
Dynamic Customer Base: Continuous model updates are required to address shifting behaviours and demographics.
Market Dynamics: ULIP policyholder behaviour is influenced by market trends and investor sentiment.
Improved Retention: The model's timely interventions have significantly improved customer retention by reducing lapsation rates among high-risk policyholders.
Optimized Marketing Spend: By targeting high-risk customers more effectively, the project has improved the efficiency of marketing spend, ensuring resources are allocated to those most likely to benefit.
Increased Revenue: The prevention of policy lapses has directly boosted premium collections and contributed to a noticeable increase in overall revenue.
Enhanced Risk Assessment: The model's accurate risk profiles have enhanced the underwriting process, allowing for better-informed decision-making and more effective risk management.
Better Customer Experience: Personalized engagement strategies have greatly improved customer satisfaction and loyalty, creating a more positive overall customer experience.
The model, deployed since FY25, has delivered a notable business impact, with a cumulative delta value of ₹249.9 crore as of December 2024, compared to ₹247.4 crore in FY24. The impact was measured using a test vs. control methodology, which demonstrated the model’s effectiveness. By deploying the model at critical policy milestones, we’ve maximized its impact, optimizing engagement, renewal rates, and business performance. Ongoing analysis and refinement will continue to drive success and scalability.
