数据支撑 AI 模型探讨沙龙圆满落幕 The data-supported AI model discussion salon came to a successful conclusion
当今数字化时代，数据已跃升为驱动社会经济发展的 “第四生产力”，成为企业决胜市场的关键资产。2025 年 3 月 28 日下午，一场以 “数据支撑 AI 大模型” 为主题的人工智能沙龙活动，在大家的支持下成功举办。本次活动由浙江省数字经济联合会、嘉兴数据局、嘉兴市经济和信息化局指导，嘉兴首席数据官联盟、嘉兴首席信息官联盟联合主办，吸引了嘉兴市企业的首席数据官（CDO）和首席信息官（CIO）到场，共同探索数据与 AI 深度融合的无限可能。
In today's digital age, data has risen to become the "fourth productive force" driving social and economic development and a key asset for enterprises to win in the market. On the afternoon of March 28, 2025, an AI salon event with the theme of "Data Supporting AI Large Models" was successfully held with the support of everyone. This event was guided by the Zhejiang Digital Economy Federation, the Jiaxing Data Bureau, and the Jiaxing Economic and Information Technology Bureau, and jointly organized by the Jiaxing Chief Data Officer Alliance and the Jiaxing Chief Information Officer Alliance. It attracted Chief data Officers (Cdos) and Chief information Officers (CIOs) from enterprises in Jiaxing to jointly explore the infinite possibilities of the deep integration of data and AI.
活动伊始，嘉兴市经济和信息化局四级调研员沈林中先生向大会致辞，他对本次沙龙活动的举办表示热烈的祝贺，是嘉兴首席数据官联盟成立以来的第1场活动，他指出当前数据作为企业的生产要素，越来越受到企业的重视，嘉兴市2024年，数字与数据经济方面所取得的工业产值接近400多亿，企业在数据治理数据要素和人工智能大模型的探索方面取得了重大的进展，希望嘉兴的数字化转型做出了贡献，希望联盟能够多多的举办这方面的活动，多多听取企业的呼声，了解企业的需求渠道，平台的作用。
At the beginning of the event, Mr. Shen Linzhong, a fourth-level researcher from the Economic and Information Technology Bureau of Jiaxing City, delivered a speech to the conference. He extended his warm congratulations on the holding of this salon event, which was the first event since the establishment of the Jiaxing Chief Data Officer Alliance. He pointed out that currently, data, as a production factor for enterprises, is receiving increasing attention from enterprises. In 2024, Jiaxing City The industrial output value achieved in the field of digital and data economy is close to over 40 billion yuan. Enterprises have made significant progress in the exploration of data governance, data elements and large artificial intelligence models. It is hoped that Jiaxing has contributed to its digital transformation. It is also hoped that the alliance can hold more activities in this regard, listen more to the voices of enterprises, understand their demand channels and the role of the platform.
嘉兴城市规划研究院研究员顾晓鸿，分享了”AI技术引领未来”的主题分享，他指出，在数字化浪潮中，人工智能已成为企业变革与发展的重要力量，而数据作为AI模型的机制，其价值愈发凸显。为深入探索数据AI模型在企业中的应用作用与发展，联盟举办的这次活动具有非常重要意义，他先以我们中国最古老的24个节气作为开场白，通过24个节气人类知道了一年四季的变化和其运行规律，并以科学发展观来论述当前数字经济的发展的演变规律，并就发展AI与安全、生态问题谈了自己的见解。他指出一项技术能否可以被接受，只有嘉兴城市规划研究院研究员顾晓红，分享了ai技术引领未来的主题分享，他指出，在数字化浪潮中，人工智能已成为企业变革与发展的重要力量，而数据作为ai模型的机制，其价值愈发凸显为深入探索数据ai模型在企业中的应用作用与发展。嘉兴首席信息官联盟联合举办的这次活动具有非常重要的意义，他先以我们中国最古老的24个节气作为开场白，通过24个节气人类知道了一年四季的变化和其运行规律，并以科学发展观来论述当前数字经济的发展的演变规律，并就发展ai与安全、生态问题。谈了自己的见解，他认为，一项技术能否被接受，只有两个标准：一是能否提升生产力，创造更多财富；二是其发展成果能否被相对公平地分配。同样，AI技术的发展带来生活便利的同时，也对生态环境产生了深远的影响，如对气候变化、资源短缺、生物多样性减少、环境污染等问题进行了探讨。
Gu Xiaohong, a researcher at Jiaxing Urban Planning Research Institute, shared a presentation on the theme of "AI Technology Leading the Future". He pointed out that in the digital wave, artificial intelligence has become an important force for enterprise transformation and development, and the value of data as the mechanism of AI models is becoming increasingly prominent. To deeply explore the application role and development of data AI models in enterprises, this event organized by the alliance is of great significance. He began with the 24 solar terms, the oldest in China, and through these 24 solar terms, humans have learned about the changes and operational patterns of the four seasons. Then, he discussed the evolution laws of the current digital economy with the Scientific Outlook on Development. And he/she shared his/her views on the development of AI and issues related to security and ecology. Gu Xiaohong, a researcher at Jiaxing Urban Planning Research Institute, shared the theme of "ai Technology Leading the Future". He pointed out that in the digital wave, artificial intelligence has become an important force for enterprise transformation and development, and data serves as the mechanism for ai models. Its value is increasingly prominent in the in-depth exploration of the application role and development of data ai models in enterprises. This event jointly organized by the Jiaxing Chief Information Officer Alliance holds significant importance. He began his speech with the 24 solar terms, the oldest in China. Through these 24 solar terms, humans have come to understand the changes and operational patterns of the four seasons throughout the year. He then discussed the evolution of the current digital economy based on the Scientific Outlook on Development and addressed issues related to the development of ai, security, and ecology. Sharing his own insights, he believes that there are only two criteria for a technology to be accepted: one is whether it can enhance productivity and create more wealth; The second is whether its development achievements can be distributed relatively fairly. Similarly, while the development of AI technology has brought convenience to life, it has also had a profound impact on the ecological environment, such as discussions on issues like climate change, resource shortages, reduction in biodiversity, and environmental pollution.
“制造业Al产品应用案例及实施”的技术分享是由百工汇智（嘉兴）公司的AI首席专家张春雷带给嘉宾的干货，他以自身的亲身实际操作经验分析了为什么数字化和数据驱动难落地的几方面因素，即：战略共识和决策者认知不足，数据基础设施和数据质量不佳，技术和业务脱节，缺乏真正有效的组织流程和机制，人才缺口与培养机制不足，对短期利益的过度追求，和企业对数字化转型的期待，即从战略、业务、技术和组织、经济效益层面的期待，并就多大程度上推动企业数字化转型和数据驱动等进行剖析，结合实际案例的操作分享了他的操作经验，以行业龙头制药企业的采购部门入手进行AI的模型操盘，通过内部数据、业绩报表、供应商的分析报表、采购质量合格率、供应商交货及时率、采购支出、财务的费用管控等分析、完成决策层、管理层、业务层所需要的报表，最后完成流程步骤的实施和公有云、私有云及混合云的部署，完成项目的上线。
The technical sharing on "Application Cases and Implementation of AI Products in Manufacturing" was presented by Zhang Chunlei, the chief AI expert of Baigong Huizhi (Jiaxing) Company, to the guests. Based on his own practical experience, he analyzed several factors contributing to the difficulty in implementing digitalization and data-driven approaches, namely: Insufficient strategic consensus and decision-makers' understanding, poor data infrastructure and data quality, disconnection between technology and business, lack of truly effective organizational processes and mechanisms, talent gap and insufficient training mechanisms, excessive pursuit of short-term benefits, and enterprises' expectations for digital transformation, that is, expectations from the aspects of strategy, business, technology and organization, and economic benefits He also analyzed the extent to which it promotes the digital transformation and data-driven development of enterprises, and shared his operational experience through practical cases. Starting from the purchasing department of a leading pharmaceutical company in the industry, he conducted the operation of AI models. Through the analysis of internal data, performance reports, supplier analysis reports, procurement quality pass rate, supplier delivery on-time rate, procurement expenditure, and financial cost control, etc., complete the reports required by the decision-making level, management level, and business level. Finally, implement the process steps and deploy public cloud, private cloud, and hybrid cloud to complete the project's launch.
吉网智服事业部信息总监项利松分享了“AI大模型企业实践与探索”，他从个人对大模型的理解出发，讲述了大模型发展历程与趋势，从感知走向内容生成，从简单预测走向复杂逻辑推理，从大语言模型走向世界模型的历程，探讨了大模型参数与场景分析，介绍了常见场景的DeepSeek模型规格，用一张图表列举了4种DeepSeek-R1系列模型不同参数规模和性能，适用于不同的应用场景，谈了AI应用落地的架构图、资源筹备情况，以及四条核心实施路径。最后，着重介绍了实施过程中必备的网络和硬件设备，尤其强调了网络数据安全的重要性。
Xiang Lisong, the Information Director of Jiwang Smart Service Division, shared "Enterprise Practice and Exploration of AI Large Models". Starting from his personal understanding of large models, he elaborated on the development process and trends of large models, from perception to content generation, from simple prediction to complex logical reasoning, and from large language models to world models. He also discussed the parameters and scenario analysis of large models. This paper introduces the specifications of DeepSeek models in common scenarios, lists the different parameter scales and performances of four DeepSeek R1 series models with a chart, and discusses the architecture diagram, resource preparation situation, and four core implementation paths for the implementation of AI applications. Finally, the essential network and hardware equipment during the implementation process were emphatically introduced, with particular emphasis on the importance of network data security.
用友制造行业及智能制造方案部总经理张雷鸣分享了“用友BIP重塑制造业数智生产力”，介绍了用友BIP制造行业AI整体情况，讲述了智能体有机融合了“生成式AI” 与“判别式AI”，各类智能体有机嵌入制造业供应链业务，就AI在制造行业的智能化场景进行了详细的论述，以生动的5段视频展现了其运用场景，同时介绍了用友正在做的智能化场景，即AI+库存健康诊断、DeepSeek+SPC质量分析、物料供需状态分析、订单变更模拟分析和生产执行分析。
Zhang Leiming, the general manager of the Manufacturing Industry and Intelligent Manufacturing Solution Department of Yonyou, shared "Yonyou BIP Reshaping Digital and Intelligent Productivity in Manufacturing", introducing the overall situation of AI in the manufacturing industry under Yonyou BIP, and elaborating on the organic integration of "generative AI" and "discriminative AI" by intelligent agents, with various intelligent agents organically embedded in the manufacturing supply chain business. A detailed discussion was made on the intelligent scenarios of AI in the manufacturing industry. Five vivid videos were presented to showcase its application scenarios. At the same time, the intelligent scenarios that Yonyou is currently working on were introduced, namely AI+ inventory health diagnosis, DeepSeek+SPC quality analysis, material supply and demand status analysis, order change simulation analysis, and production execution analysis.
最后，大家就AI大模型的话题进行了探讨。以中小企业在人工智能浪潮下如何运用AI分别由用友张雷鸣、吉网通讯项利松、火山引擎赵广辉、百工汇智总经理章杰艺及世源科技信息总监成涛五位嘉宾就自己的观点谈了看法。总结大家的观点，中小企业实施AI需先解决自身难题，依托数据支持，合理控制投入，利用DeepSeek、豆包等开放平台，避免盲目扩张。可先从人事招聘、售后服务、内部办公管理入手，再逐步向财务报表分析等领域拓展。正如世源科技的成涛说的他们，先听取企业内部员工的诉求，把所有诉求归纳整理，然后信息部门根据其诉求，分析出那些诉求目前可以先做，那些诉求目前技术、数据和投入成本达不到要求暂时放放，对可做的诉求采取一对一的集中精力做，目前他们已解决了10多个诉求。探讨持续了近一个多小时，大家感觉还有许多话题可探讨，因时间关系将把此话题放在下场沙龙活动继续。
Finally, everyone discussed the topic of large AI models. Five guests, namely Zhang Leiming from UFIDA, Xiang Lisong from Jiwang Communication, Zhao Guanghui from Volcano Engine, Zhang Jieyi, the general manager of Baigong Huizhi, and Cheng Tao, the information director of Shiyuan Technology, shared their views on how small and medium-sized enterprises can apply AI in the wave of artificial intelligence. Summarizing everyone's viewpoints, for small and medium-sized enterprises to implement AI, they need to first solve their own problems, rely on data support, reasonably control investment, and utilize open platforms such as DeepSeek and Doubao to avoid blind expansion. One can start with personnel recruitment, after-sales service, and internal office management, and then gradually expand into areas such as financial statement analysis. As Cheng Tao of Shiyuan Technology said, they first listen to the demands of the internal employees of the enterprise, summarize and organize all the demands, and then the information department, based on these demands, analyzes which ones can be addressed first and which ones cannot meet the current technical, data and investment cost requirements and are temporarily put on hold. For the demands that can be addressed, they adopt a one-to-one concentrated approach to handle them. So far, they have resolved more than ten demands. The discussion lasted for nearly an hour and a half. Everyone felt that there were still many topics to discuss. Due to time constraints, this topic will be carried over to the next salon event.
本次沙龙活动给大家提供了一个交流探讨的平台，起到了应有的效果，联盟将持续在当今流行的前沿技术方面与各位嘉宾朋友们携手共进，多多提供技术分享和交流，为企业在数字化改造和人工智能方面尽微薄之力。联盟是大家的，希望嘉宾朋友们多为联盟建设提出更好的建议，我们期待着下次活动再见。
This salon event provided a platform for communication and discussion, achieving the expected results. The alliance will continue to work hand in hand with all the guests and friends in the cutting-edge technologies that are popular today, offer more technical sharing and exchanges, and contribute its humble efforts to enterprises in digital transformation and artificial intelligence. The alliance belongs to all of us. We hope that our distinguished guests and friends can offer better suggestions for its development. We look forward to seeing you again at the next event.