We built a supervised machine learning model that ingests wearable, financial, and social data streams and consolidates them into a single Health Risk Score per applicant, replacing static averages with a dynamic, individualized assessment. On top of the scoring engine, we implemented ML-based application routing, automatically directing lower-complexity applications through streamlined paths while flagging higher-risk cases for underwriter review. Underwriters were given real-time dashboards surfacing the risk score, its contributing factors, and routing decisions, all built within an IRDAI-compliant data architecture to meet regulatory requirements for the Indian insurance market.