Shennong Smart Guard：An AI-Native Spatio-temporal Paradigm for End-to-End Agri-Risk Management. Qorus-NTT DATA Innovation in Insurance Awards 2026

Concept and Objectives

Shennong Smart Guard is the industry’s first domain-specific "Agri-Disaster Large Model" designed to transform agricultural insurance from a passive "claim-pay" model into an active "predict-prevent-protect" ecosystem. By integrating multi-source data—Satellite Remote Sensing (RS), real-time meteorology, and IoT—we have built a digital twin of the agricultural landscape. Our objective is to achieve "Precision Underwriting & Precision Claims" while building climate resilience for the global food supply chain.

Reasons Behind the Innovation

While China represents the world’s largest agricultural insurance market (~$21B USD) with a 16.2% CAGR, the industry remains plagued by three systemic pain points:

• Information Asymmetry: High rates of fraudulent claims and "ghost" policies due to manual verification.

• Operational Inefficiency: Fragmented data and labor-intensive damage assessment.

• Reactive Nature: Traditional insurance only reacts after the loss has occurred, offering no value in disaster mitigation.

With a strategic $70M+ (500M RMB) R&D investment since 2019, Ping An P&C identified the need to bridge this "protection gap" through deep-tech integration.

State of Competition & Sources of Inspiration

While the global InsurTech market has seen various digital tools, most existing solutions hit a "technical ceiling" by acting as passive data repositories. Shennong Smart Guard defines a new generation of risk management by addressing three systemic limitations of current competitors:

• From Fragmented Data to Holistic Fusion:

Most industry peers rely on generic weather APIs or basic satellite imagery. We have moved beyond this by engineering a multi-dimensional data fabric. By fusing meteorology, remote sensing, soil science, agronomy, and multi-source data, our system understands the biological "ground truth" of the crop, not just the overhead weather patterns.

• From Descriptive Models to our Agri-Disaster LLM:

Current market solutions typically use traditional statistical models that describe what happened. Our proprietary Agri-Disaster Large Language Model (LLM) shifts the focus to Generative Risk Logic. It processes complex, non-linear environmental variables to provide prescriptive advice, transforming raw data into actionable intelligence.

• From Macro-Estimation to Plot-Level Precision:

Where traditional systems operate at a coarse regional or county level (500m+ resolution), we provide hyper-local precision at a 10m-plot scale. This allows for the identification of micro-climates and specific soil stress, enabling "Precision Underwriting" that was previously considered technically impossible at scale.

• From Passive Indemnity to Active Resilience:

The traditional competition remains rooted in a "Pay-for-Loss" reactive cycle. Shennong Smart Guard reclaims the initiative with a 72-hour early warning window (87.1% accuracy). We are not just an insurance provider; we are a provider of "Climate Resilience-as-a-Service," actively reducing the physical loss before the financial claim even exists.

Departments involved

Our organizational framework follows a "Full-stack Internal Integration + High-tier Academic Synergy" model, ensuring that cutting-edge R&D is seamlessly translated into field-level operations.

• Cross-Functional Full-Stack Integration (Ping An P&C Headquarters):

The project is strategically orchestrated by the Agricultural Insurance Department, achieving deep vertical integration with the Claims Operations Center, the IT R&D Team, and the Data Intelligence Platform Team. This unified structure eliminates data silos and enables end-to-end lifecycle management—from scenario definition and data governance to rapid model and product iteration .

• Industrial-Grade Infrastructure & Remote Sensing (Ping An P&C Tech):

Our technology arm provides the core computational powerhouse. This includes a robust MLOps (Machine Learning Operations) framework and a proprietary Industrial-grade Remote Sensing Platform. These assets support the distributed training of the "Shennong" large model, the processing of PB-level spatio-temporal data, and millisecond-level real-time inference.

• Scalable Operational & Validation Network (42 Regional Branches):

The system is deployed across a nationwide network of 42 branches, with 11 key pilot regions (including Hunan, Shandong, Guangdong, and Heilongjiang) serving as "Vertical Living Labs." For instance, our Camellia Oleifera warning project in Hunan provides high-fidelity ground-truth data that continuously calibrates our algorithms, ensuring the model's adaptability across diverse climates and crop types.

• Multi-Source Heterogeneous Data Governance:

We have established a sophisticated data governance ecosystem involving 10+ authoritative entities, including Changguang Satellite, Caiyun Weather, the Bureau of Agriculture and Rural Affairs, and national financial regulatory bodies. By unifying data standards across satellite imagery, real-time meteorological streams, and government records, we have constructed a dynamic "Digital Twin" of the agricultural risk landscape.

• Academic-Industry Innovation Matrix:

We collaborate with top-tier global institutions to define industry standards and push the boundaries of disaster science:

Tsinghua University: Jointly published the White Paper on Climate Change Adaptation and Disaster Governance, establishing international benchmarks for insurance-led climate resilience.

Southern University of Science and Technology (SUSTech): Co-developed a high-resolution disaster simulation system, utilizing reinforcement learning to enhance our generative capabilities for extreme events like typhoons, rainstorms, and floods.

Tsinghua University Fintech Institute: Collaboratively authored the 2022 White Paper on High-Quality Development of Tech-Enabled Agri-Insurance, leading the strategic roadmap for the industry.

Main Results So Far:

Ping An P&C has pioneered a "Predict-Prevent-Protect" model, providing over $105 Billion (7,419B RMB) in risk coverage for the agricultural sector over the past three years. We have successfully translated deep-tech innovation into measurable economic, social, and ecological returns.

• Technical Excellence: From Passive Indemnity to Proactive Governance

We have engineered the industry’s first Generative AI & Spatio-Temporal Large Model dedicated to agricultural risk. This system creates a seamless digital loop:

Pre-Disaster (Predictive AI): Achieved 87.1% accuracy in disaster forecasting with a 72-hour lead time, enabling farmers to pivot strategies before the event.

During-Disaster (Dynamic Simulation): Real-time IoT and RS data allow for dynamic loss inference with 85% estimation accuracy, supporting government resource dispatching.

Post-Disaster (Smart Settlement): Using 10m-resolution LAI (Leaf Area Index) monitoring and Reinforcement Learning, we automated damage assessment with 85% accuracy, slashing settlement times.

Recognition: 5 Wins in International AI Competitions; 80+ Patents; and a State Council Commendation for our "Double Precision" model, marking a historic paradigm shift in the insurance industry.

• Economic Impact: High-Precision Loss Mitigation

Over 3 years, we covered 25 million farming households. By investing $40M (285M RMB) in prevention technology, we directly reduced agricultural losses by $107M (760M RMB).

Case Study (Drought & Typhoons): In 2025, our "Digital Heatmaps" helped Henan and Suzhou farmers save over $1M in wheat yields. During Typhoon Wipha, our model locked down risk zones 7 days in advance, deploying 161 drone sorties to prevent $1.8M in potential losses.

Precision Response: In Xiangyang, our early-warning system and donated resources saved $3.3M in autumn grain, demonstrating the direct bottom-line impact of tech-enabled prevention.

• Social Resilience: Operational Integrity & Efficiency

We deployed an intelligent risk control network covering 591 insurance products and 3,459 crop growth stages:

Governance & Anti-Fraud: Executed 250,000 monitoring tasks, successfully identifying 2,163 fraudulent claims and correcting 358,400 acres of false reporting.

Efficiency Gains: In Liaoning, claim settlement speed increased by 74.32% (from investigation to payment).

Customer Trust: Farmer satisfaction rose by 22%, and our methodology has been adopted by 8 industry peers.

• Ecological Innovation: The Global ESG Benchmark

We have integrated ecological resilience into our core underwriting:

Resource Optimization: Our system (utilizing TVDI drought indices) boosted irrigation efficiency by 23% and reduced chemical pesticide usage by 35%.

Global Scalability: As a premier ESG digital benchmark, our technical solutions have been exported to Vietnam and Thailand, supporting climate resilience across the Asia-Pacific region.