Waterdrop, Inc. ADR Class A (WDH.US) has been granted three additional national invention patents for AI, empowering insurance and health service scenarios.

date
11:55 25/08/2026
avatar
GMT Eight
The acquisition of these three patents marks a new national-level recognition of Waterdrop's independent R&D capabilities in core technology areas such as intelligent voice interaction, multimodal document recognition, and social intent understanding. This provides a solid technical foundation for the intelligent upgrade of insurance technology and health service scenarios.
Recently, Waterdrop, Inc. ADR Class A (WDH.US) successfully obtained authorization for three national invention patents in the field of artificial intelligence, aimed at optimizing intelligent voice interaction, automatic classification of document images, and intelligent judgment of service timing in various business scenarios. The acquisition of these three patents marks a national-level recognition of Waterdrop, Inc. ADR Class A's independent R&D capabilities in core technology areas such as intelligent voice interaction, multimodal document recognition, and social intent understanding, providing a solid technological foundation for the intelligent upgrade of insurtech and health service scenarios. Each of the three patents has its unique breakthroughs, covering core aspects of insurance services. The first patent is for the Method, Device, Electronic Equipment, and Storage Medium for Siasun Robot&Automation Response During Outgoing Calls. This technology addresses pain points like discrepancies between the Siasun Robot&Automation's response content and reality, as well as its inability to accurately respond to users. It proposes an intelligent response solution based on a large language model. The patent captures user request audio, converts it into text, and combines it with the text information transformed from historical call records during the conversation, inputting both into the large language model to generate more precise response content. For instance, when a user asks a question, Siasun Robot&Automation not only hears the current user statement but also organizes past dialogues, providing accurate replies based on context. The second patent is for the Method, Device, Equipment, and Storage Medium for Document Image Classification. This patent integrates layout analysis technology with deep fusion of multimodal features, enabling the categorization of contents in classified document images and the fusion of feature information corresponding to layout category information for each content block. Compared to traditional document classification methods, this technology significantly improves the automatic classification accuracy of complex vouchers, medical documents, and claims materials. In scenarios such as insurance claims and medical document review, where large volumes of documents need to be processed, this technology is expected to drastically reduce manual classification time and enhance review efficiency. The third patent is for the Method and Device for Command Processing Based on Social Intent Recognition, Storage Medium, and Equipment. This patent determines social intent through conversational content and further extracts multimodal information such as voice acoustic features and text semantic features based on user conversation content to perform corresponding follow-up actions. This technology significantly enhances intent recognition accuracy through a multimodal feature fusion mechanism and counter management, while effectively avoiding unnecessary model calls, thereby reducing operational costs. This technology will be widely applied in customer operations and similar scenarios, assisting business personnel in initiating requests at appropriate moments to improve customer conversion rates. Continuing to Increase R&D Investment, AI Patent Matrix Becoming Increasingly Complete Waterdrop, Inc. ADR Class A is firmly advancing its All in AI strategy, investing nearly 300 million yuan each year in technology R&D. To date, the company has applied for nearly 200 artificial intelligence patents, including 97 related to large models, continuously exploring the application of technologies like big data and artificial intelligence in the industrial value chain. Through the self-developed Waterdrop, Inc. ADR Class A Water Guardian Large Model, as well as systematically implementing Chinas first AI underwriting expert KEYI.AI, AI intelligent customer service Bao Xiao Hui, and AI intelligent quality inspection and other innovative applications, Waterdrop, Inc. ADR Class A has achieved significant improvements in service efficiency and user experience. Not long ago, Waterdrop, Inc. ADR Class A's Beijing Water Guardian Intelligent Technology Co., Ltd., with its deep AI implementation capabilities in the insurance and healthcare verticals, successfully made it onto the Top 50 Chinese AI Technology Enterprises in 2023 Emerging AI Companies list. In the future, Waterdrop, Inc. ADR Class A will continue to deepen the application of AI technologies in insurance and service scenarios, driving the core technologies of intelligent voice interaction, multimodal recognition, and intent understanding to continuously output to the fields of insurtech and health management, working together with industry partners to promote a more efficient and intelligent development of the insurance industry.