Ping An Health achieves 98% AI diagnostic accuracy and first full-year profit in 2024

China's healthcare system faces chronic resource constraints — a shortage of qualified physicians relative to a rapidly aging population creates significant pressure on digital health platforms to scale without proportional headcount growth. Ping An Health, operating across family doctor, specialist, and health manager services, needed to serve a growing national user base while maintaining diagnostic quality. Manual review of health checkup reports, consultation triage, and chronic disease follow-up were bottlenecks that limited throughput. Without AI augmentation, the cost of delivering consistent, accurate care across millions of patient interactions was unsustainable and threatened both service quality and the company's path to profitability.

Ping An Health built a proprietary multi-modal medical AI system — the Ping An Medical Master® model — trained on five industry-leading medical databases and 1.44 billion historical online consultations, giving it a domain-specific foundation that general-purpose models lack. The accompanying Ping An Doctor's Home™ workbench integrated this AI directly into physician workflows, enabling assisted diagnosis, intelligent triage recommendations, and automated health checkup interpretation at scale. The company also validated and deployed DeepSeek's large language model for structured medical scenarios including chronic disease management. Predictive ML models were applied across the care continuum — from intake to ongoing condition monitoring — embedded within existing service workflows rather than deployed as standalone tools, enabling incremental rollout across doctor and health manager teams.

AI-powered health checkup interpretation reached 98% accuracy, reducing manual physician review time significantly. AI-assisted diagnosis accuracy exceeded 95%, while intelligent recommendation accuracy hit 99% — metrics validated across production-scale patient volumes. Chronic disease management improvement rates reached 90%. Service efficiency gains were substantial across all care roles:

• Family doctor services: ~62% efficiency improvement
• Specialist doctor services: ~42% efficiency improvement
• Health manager services: ~55% efficiency improvement

Financially, 2024 marked the company's first full-year adjusted net profit of RMB158 million. Customers enrolled in integrated insurance-plus-healthcare products held 1.6x more contracts and showed 3.9x higher AUM per capita, demonstrating that AI-driven service quality directly influenced commercial outcomes.

• Proprietary training data (1.44B consultations) creates a durable moat — generic models cannot replicate domain-specific accuracy without equivalent data assets.
• Embedding AI into existing clinician workflows (rather than building parallel systems) drives faster adoption and measurable efficiency gains across diverse care roles.
• Integrating AI across the full care continuum — checkups, diagnosis, chronic disease management — compounds returns more than point-solution deployments.
• Linking AI-enhanced services to insurance products creates a measurable commercial feedback loop: better health outcomes increase product engagement and AUM per customer.