Mei Ri Jing Ji Xin Wen·2025-11-13 14:01Core Insights - The true value of large models lies in deep specialization within vertical fields, which enhances the overall intelligence density of the ecosystem [1] - Ping An's AI strategy, termed "AI in All," aims to redefine service margins and transition from a labor-intensive model to one led by a few skilled professionals managing a group of robots [1][2] AI as a Value Creation Center - AI technology is the core engine driving Ping An's "Finance + Technology" and "Finance + Ecosystem" strategies, evolving through three stages: small models, strong thinking models, and now scaling capabilities [2] - The AI-driven insurance system has achieved a 4000-fold increase in damage assessment speed, processing over 40,000 claims daily, and is the only automated damage assessment platform in large-scale production [2][3] Applications in Various Sectors - In smart city initiatives, Ping An has developed a multi-layered financial risk warning and economic decision-making system, achieving over 90% recall rate for defaulting companies and 85% accuracy in macroeconomic predictions [3] - The "Yibotong" model in healthcare has improved consultation and diagnosis accuracy to 99% and 94%, respectively, utilizing 200 billion medical tokens [3] Balancing Privacy and Data Value - Ping An focuses on vertical specialization in AI, establishing a robust research and organizational structure to support rapid model development and deployment [4][5] - The company employs a dual-track framework of "technology + system" to balance data value extraction with user privacy protection [5] Competitive Advantages and Future Directions - Ping An's "Five Wisdom" strategy emphasizes intelligent marketing, services, operations, management, and business, leveraging open-source models for vertical innovation [6][7] - The transition from "perceptual" to "operational" AI applications has led to the creation of over 57,000 intelligent agents, significantly enhancing operational efficiency [7][8] Systemic Impact of AI - The deep application of AI models is creating a chain reaction, leading to the emergence of "AI-native applications" and systemic intelligent agents that share capabilities through a unified model base [8][9] - The focus on model safety and interpretability is critical for compliance and ethical standards, especially in finance and healthcare sectors [9]