Yilianzhong Launches 'ZhiDing YunFan' Large Model Platform to Enhance Public Service Efficiency and Quality

SelfSince the breakthrough of GPT, a wave of exploration centered on large language models has swept across various industries. In late 2023, the National Data Administration also released the “Three-Year Action Plan for ‘Data Element ×’ (2024–2026) (Draft for Comments),” proposing to support the development of large models with scientific data, build high-quality corpora and basic scientific datasets, and accelerate the practical application of both general-purpose and vertical large models.

In the realm of public welfare services, sectors such as human resources and social security, medical insurance, and health care involve high-frequency communication in daily operations, making them highly compatible with the capabilities of large language models (LLMs). Recently, Zhiding Technology, a subsidiary of YLZ Information Technology Co., Ltd., launched the “Zhiding Yunfan” industry-specific large model—the YLZ Public Welfare Information Service Large Model Platform. This initiative marks the first introduction of LLMs into the public welfare sector, aiming to meet the scalable demand for such technology in this field.

Over the past two decades, the rapid advancement of internet technology has facilitated a comprehensive upgrade in sectors pertaining to public welfare. Under service models characterized by paperless operations, standardization, and interconnectivity, both staff work efficiency and user service experience have witnessed tangible improvements.

However, service models have become relatively rigid, and incremental updates centered on fixed applications are no longer able to significantly optimize the efficiency of public livelihood services. According to Zhuang Guoqiang, Deputy General Manager and Technical Director of YLZ Zhiding Technology Co., Ltd., achieving a qualitative improvement in service quality within the established paradigm of public livelihood systems requires reforms driven by more human-centric and intelligent objectives.

“To ensure accuracy and authority, policies typically provide comprehensive and detailed descriptions of services and the various scenarios users may encounter when utilizing them. However, the public tends to focus on their own needs when interpreting these policies, and their expression of such needs may lack directness and standardization. This places significant demands on the service capabilities of both staff members and existing application systems.”

Furthermore, policies themselves are subject to dynamic adjustments, and provisions and procedures vary across different regions. This necessitates continuous differentiation between old and new policies as well as among regional policies, thereby further increasing the complexity of knowledge management.

In theory, artificial intelligence (AI) technology can address most of the challenges encountered by the aforementioned service providers. Compared with ordinary individuals or traditional application systems, algorithms can more accurately match user needs, policies, and business requirements based on prompts. The recently released “Zhi Ding Yun Fan” industry-specific large language model by YLZ Information Technology Co., Ltd. further enhances the system’s ability to understand user requests and improves the management of policy and business knowledge, thereby comprehensively strengthening the capabilities of service providers.

According to Zhuang Guoqiang: YLZ Information Technology Co., Ltd.The “ZhiDing YunFan” industry-specific large language model currently possesses foundational capabilities, including intent recognition, industry knowledge organization and comprehension, intelligent Q&A, content generation, logical reasoning, and business system integration. It can develop comprehensive solutions tailored to the specific operational needs of sectors such as human resources and social security, medical insurance, and healthcare. This empowers more intelligent implementation models for applications such as smart transaction processing, intelligent consultation, data-driven decision-making, smart services, intelligent regulation, and virtual digital humans.

In the process of developing its large language model (LLM), the R&D team at YLZ Information Technology Co., Ltd. started from specific business scenarios in human resources and social security, medical insurance, and healthcare. They incorporated domain-specific knowledge and business logic into large-scale self-supervised training to automatically extract industry insights, thereby enhancing the accuracy and professionalism of the LLM. Furthermore, YLZ’s “Zhiding Yunfan” industry-specific LLM integrates existing AI atomic capabilities from the “Zhiding Tiangong AI Platform,” supporting the deployment of the LLM through features such as industry plugins, prompt engineering, and compliance management.

Compared with existing AI products in the public welfare sector, the advantages of YLZ’s “Zhiding Yunfan” industry-specific large model can be broadly categorized into four aspects.

First, YLZ Information Technology Co., Ltd.’s more than 20 years of experience in the public welfare sector enable its “Zhi Ding Yun Fan” industry-specific large language model to accurately comprehend domain knowledge and business logic in public welfare information services, while also addressing interoperability challenges within public welfare systems. Currently, this large language model supports deployment models such as on-premises privatization and SaaS, allowing it to integrate seamlessly with YLZ’s proprietary systems as well as interoperate with systems from other vendors.

Secondly, livelihood-related businesses place greater emphasis on safety and compliance. The “Zhi Ding Yun Fan” industry-specific large language model can further refine and customize exclusive models based on domain-specific foundation models to meet specific business needs, providing functionalities such as knowledge provenance, compliance management, and security control, thereby ensuring legal and regulatory compliance as well as information security.

Furthermore, to address the diverse and dynamic nature of public welfare services, YLZ Information Technology Co., Ltd. has endowed the “Zhiding Yunfan” industry-specific large language model with autonomous learning capabilities. This enables the model to independently extract, summarize, and generate knowledge corpora from massive volumes of industry data, facilitating continuous iterative evolution and reducing manual operational and maintenance costs.

Finally, the “ZhiDing YunFan” industry-specific large model breaks through the limitations of traditional AI in single-modality data processing. It supports multiple user interaction modes—including voice, text, documents, images, and video—simultaneously, delivering a more intelligent, user-friendly, and efficient interactive experience.

To date, YLZ Information Technology Co., Ltd.“ZhiDing YunFan” industry large model has had multiple capabilities widely applied in the field of public welfare.

The “Smart Service” application launched by the medical insurance department serves as a typical example. Traditional systems require users to fill out multiple forms and repeatedly confirm the entered information when serving users. In contrast, this application adopts an innovative “one-sentence” service model, allowing users to initiate business processes through natural language conversations. The system intelligently parses key information, completes form filling, and submits it to the business system for intelligent approval, thereby achieving immediate processing and completion or instant acceptance of services.

In this process, AI-powered intelligent voice technology accurately captures and parses user voice commands, while natural language processing (NLP) and large language model (LLM) technologies efficiently process data in the background to ensure accurate information collection and organization. Meanwhile, virtual digital human technology provides a friendly interactive experience, making service procedures smoother and more user-friendly.

For healthcare security authorities, intelligent administrative services can enhance the efficiency of government services, optimize resource allocation, and reduce the waste of human and material resources. For the public, 24/7 intelligent services mean they can easily handle healthcare security matters via mobile devices anytime and anywhere, avoiding the inconvenience of queuing at service counters and significantly saving time and effort.

Intelligent recognition systems are also widely applied in sectors pertaining to public welfare. As a text processing tool that integrates Optical Character Recognition (OCR) with large language models, the system leverages OCR technology to rapidly and accurately convert paper documents into digital text. It then utilizes large model technology to extract key information from the text and transform this data into a unified standard format.

Large language models play a core role in this system. They not only improve the accuracy of information extraction but also enable the identification and understanding of complex information from unstructured sample data. This capability is particularly important for processing diverse documents and forms, such as invoices, contracts, and reports.

Compared to the non-AI systems currently in widespread use, intelligent recognition systems address the time-consuming and error-prone nature of manual data entry, while also overcoming the limitations of traditional OCR technology in information extraction and standardization. In practical applications, this system has enabled government agencies and enterprises to automate data processing, significantly improving work efficiency and data accuracy, with notable success in areas such as sporadic medical insurance reimbursements. According to Zhuang Guoqiang, declaration tasks that previously took over an hour can now be completed in approximately five minutes, boosting the work efficiency of relevant departments by around 90%.

Overall, the value of the “ZhiDing YunFan” large language model lies not only in improving people’s quality of life, but also in helping governments and enterprises better respond to public needs, optimize the allocation of social resources, and even provide data support and insights to enable more scientific and rational decision-making.

Furthermore, in an era that pursues the universality and equity of services, large language models offer a new perspective. By empowering service providers with enhanced capabilities, YLZ Information Technology Co., Ltd. aims to ensure that everyone can access high-quality, efficient public livelihood services, thereby fulfilling its corporate vision of “reshaping a new ecosystem for public health and well-being through digitalization.”