• Outdated Rate Cards: Reliance on legacy assumptions that don't reflect current mortality/morbidity patterns.

• Information Silos: High-quality census and occupation class data remaining underutilized for dynamic pricing.

• Pricing Uncertainty: Slower response times and higher uncertainty when quoting for unfamiliar industry segments.

• Manual Audits: Lack of automated deviation alerts between tool-generated rates and actual experience.

We delivered a unified AI platform designed to augment actuarial rigor with machine learning precision:

• Data Lakehouse Integration: Centralized storage of inquiry history, census datasets, and claims archives on Databricks with real-time API communication.

• Dual-Rate Modeling: Simultaneous calculation of technical rates and experience-based rates with credibility weighting.

• Benchmarking Engine: Automated cohort analysis to identify similar historical groups based on industry, size, and OC distribution.

• AI Pricing Assistant: An LLM-integrated layer that provides real-time insights and captures user feedback for continuous model learning.

Technologies Used  

• AI/ML: Experience-based Modeling, Similar-group Matching, LLM Pricing Assistant  

• Data Engineering: Databricks, ETL Pipelines  

• Integration: Real-time APIs, Side-by-side Comparison UI  

‍

‍

• Increased Win Ratios: Real-time competitive benchmarking allows for more accurate and aggressive bidding.

• Portfolio Health: Immediate identification of underperforming segments and emerging loss patterns.

• Underwriter Confidence: Direct access to industry norms and benchmarks for all team members.

• Data-Forward Strategy: Successful transition from reactive pricing to a predictive, evidence-based model.