"Super AI Search": A New Paradigm for User Experience Qorus-NTT DATA Innovation in Insurance Awards 2026

Super AI Search stands as an AI-driven comprehensive search system independently developed by Ping An Property & Casualty Insurance (Ping An P&C), facilitating the company's paradigm shift from conventional keyword matching to proactive intent prediction. Its core objective aims to surmount the limitations of traditional search engines contingent upon keyword matching. By embedding large language model (LLM) technology deeply within insurance business logic, the system accomplishes nuanced understanding of user intent coupled with proactive service recommendations, covering insurance product selection, claims counseling, service guidance, video and article recommendations. It establishes a novel user interaction model of "search-as-service", propelling insurance services from passive response to active, intelligent provision.

This upgrade stems from the inadequacy of traditional search in addressing complex, ambiguous, or cross-scenario user demands. Concomitantly, breakthroughs in LLMs regarding task planning, multi-agent collaboration, and dynamic decision-making have provided the technical feasibility for constructing cognitively capable search systems. Capitalizing on the trend, Ping An P&C is committed to resolving critical challenges inherent to insurance contexts, namely intent ambiguity, service misalignment, and conversion friction — consequently enhancing user service efficiency and experience.

Across the general search domain, tech giants such as Google, Baidu, and ByteDance are intensifying their deployment of AI search and agent technologies. In the face of these competitors with formidable general AI capabilities, Ping An P&C has opted to focus on the vertical insurance field to build differentiated competitive advantages. By deeply embedding LLM capabilities into the insurance value chain, the company has created a closed loop of scenario understanding, intelligent decision-making, and service delivery, forming insurmountable industry barriers and achieving a positioning upgrade from general search to exclusive intelligent advisor.

The system design draws inspiration from the task orchestration mechanism pioneered by Nano AI Super Search Agent, drawing on its task decomposition and dynamic scheduling mechanisms. Developing from this groundwork, Super AI Search innovatively constructs a cognitively elastic agent collaboration network: global decision-making through Meta-Agent, dynamic negotiation via the A2A (Agent-to-Agent) protocol, and standardized access to various tools and services through the MCP (Modular Context Protocol). This architecture accomplishes a systemic leap from fixed workflows to dynamic planning, and from tool connection to intelligent decision-making.

This project is led by the tech departments of Ping An P&C, adopting a twin-engine innovation framework of progressive business optimization and architectural paradigm reconstruction to guarantee synchronous implementation of technological upgrades and business value. Furthermore, the system design reserves open interfaces to support future expansion into ecosystem scenarios such as health insurance and pension finance.

To date, Super AI Search has achieved multimodal input understanding (text, voice, image) and personalized content delivery, comprehensively serving nearly 100 million users of the Ping An Auto Owner APP. The system has significantly improved search click-through rates and service conversion rates cumulatively contributing 220 million (RMB) in GMV, and has become a key cognitive infrastructure supporting the company's "Insurance + Technology" strategy.