文｜品牌议题 今年5月，"什么值得买"进行了一轮产品升级，发布GEN2版本，主要依托AI技术，打造基于兴趣的内容 推荐与服务体系。结合其全新的品牌主张"我们的兴趣消费指南"，可以看出"什么值得买"的平台定位， 也正从过去的"导购工具"逐步向"AI驱动的兴趣消费平台"转变。 这个低调的电商导购平台，因拼音首字母SMZDM被网友戏谑地称为"色魔张大妈"或"张大妈"。早期积 累的一批老用户粘性极高，他们用它来比价、测评，分享优惠信息，可以说承载着一种古典互联网的实 用主义情怀。 而随着平台商业化和规模扩大，广告增多，内容质量也更加良莠不齐，而AI内容的涌入，一方面确实 极大地提升了效率，但也会让人重新思考：大家看"什么值得买"，究竟看的是什么？ 01、发家模式：以"值"为核心的导购社区 "什么值得买"始于2010年隋国栋的同名博客。 当时，电商行业的格局初步形成，并从"能买"向"买得好"转型，消费者对可信赖的导购内容，以及线上 比价的需求也随之上升。 创始人隋国栋通过博客文章的形式，推荐各种电子产品，包括最新的产品信息、优惠渠道、测评等等， 吸引了一批同样热衷于讨论数码产品的种子用户。2011年10月，"什么值得买 ...

近日，值得买科技与南开大学商学院联合研究成果《AI总结用户评论生成摘要对用户决策行为的影响》（The Impact of AI-Generated Product Summaries on the Speed and Efficiency of E-commerce Consumer Decision-Making）成功入选电子商务领域顶级会议SCECR 2025。此前，该研究 成果还入选信息技术管理领域顶级会议HICSS 2025（即夏威夷国际系统科学大会）。这是继与人大高瓴人工智能学院多项AI联合研究成果发 布之后，值得买科技与学术界的又一次创新探索。 该成果研究了AI生成摘要如何影响电商平台上消费者的决策过程，通过分析"什么值得买"平台的消费相关数据，并应用三重差分法（DDD） 来评估AI生成摘要对消费者行为的影响。研究结果表明，AI生成摘要提升了用户决策效率，使得用户因犹豫而保存、待定的行为减少了 2.8%，进一步缩短决策时间，用户购买更果断。 南开大学商学院副院长李凯介绍，现有的AIGC研究主要集中在其在个性化推荐、客户服务、产品描述和内容生成等方面的应用，强调其对消 费者参与度和满意度的积极影响， ...

Core Insights - The consumption platform "What Worth Buying" has reported a significant increase in consumer decision-making efficiency driven by AI, leading to growth in consumption during the "618" shopping festival [2] - The company launched multiple AI products, including the AI engine "Fire Eye" AIUC engine, user-facing AI products "What Worth Buying" GEN2, shopping agent "Zhang Dama," and the foundational AI infrastructure product "Haina" MCP Server [2] - The AIUC engine has enhanced product selling points, risk warnings, and purchase suggestions, with a 15-fold increase in AIUC label recognition and a reduction in content review time to one-eighth of last year's duration [2] Company Developments - The "Haina" MCP Server has gained industry attention for its integration of data and content advantages in the consumer sector, offering unique value in richness, real-time capabilities, and neutrality [3] - The "Haina" MCP Server has been made available to various partners, including large models and intelligent terminal applications, facilitating extensive adaptation for AI applications [3] - In June, the overall content and product output from the "Haina" MCP Server exceeded 8 million, marking a nearly 80% increase compared to May [3]

Core Insights - The company is enhancing its AI infrastructure with the "Haina" MCP Server, which has released two new interfaces and expanded capabilities [1] - The "What Worth Buying" GEN2 has shown significant improvements in user experience and consumer decision-making efficiency, leading to increased sales [2] Group 1: AI Infrastructure Development - The "Haina" MCP Server has introduced two major interfaces: Content Search Pro and Image Search capabilities, with the relevance of search results improving from 90% to 95% [1] - The Image Search API now supports 20 million product images across 16 major categories and 1,700 subcategories, facilitating more precise product searches for users [1] - In June, the overall content and product output from the "Haina" MCP Server exceeded 8 million, marking a nearly 80% increase compared to May [1] Group 2: Performance Metrics of GEN2 - The "What Worth Buying" GEN2 has enhanced product functionality and user experience, significantly improving consumer decision-making efficiency [2] - During the "6·18" shopping festival, the average GMV per user on the GEN2 platform increased by 15.1% year-on-year [2] - The order conversion rate achieved through the platform's AI shopping assistant rose by 162.6% year-on-year [2]

文 ｜ 阑夕 上个月月底，Veo 3的发布再次让海外社交媒体炸了锅，大量AI视频如病毒般迅速铺满了各大平台，这次受伤的不只有老面孔Ins，连TikTok也未能幸免。 这中间还不乏有反串乐子人的出现，国外有支独立乐队发了一段演奏视频称「百分百由Google Veo 3生成」，结果最后发现这段视频其实来自多年前乐队 的老歌，辨真伪游戏的难度飙升，莫说是不熟悉AI的人，恐怕连从业者想要分辨出视频来源都得费一番功夫。 包括与Veo 3同时期上线、号称「最聪明大模型」的Claude 4，在各种任务上都达到了SOTA（当前最佳）水平，只能说从视觉到能力，AI应用炸裂型驱动 的现状还远未到结束的时候。 与上述热闹的场面不同，国内互联网对炸裂这个词汇的敏感阈值在肉眼可见地下降，比如快手可灵的用量其实一直领先于Veo 3，但也不见得那些猎奇AI 视频在中国短视频平台上有多么强的病毒式传播，所有类似的内容也都被约定俗成般规范到了那一小片试验田里，不是没有，但也到不了草木皆兵的程 度。 与此同时，AI在垂类领域应用上的进程在持续加速，越来越多的消费级产品出现，这番技术迭代的光景，与移动互联网的发展规律几乎是如出一辙的， 也就是中国 ...

转自：北京日报客户端 新消费时代，面对形形色色的信息和消费选择，不少消费者陷入信息过载的陷阱，精准、高效地提升用 户决策效率成为消费科技的发展方向。5月28日，值得买科技发布了"火眼"AIUC引擎（AIUC即AI Understand Content，人工智能内容理解技术）、"什么值得买"GEN2和购物智能体——AI购物管家"张大 妈"。 "当前用户的消费习惯正在快速变化，用户对于搜索产品的期待，从'找信息'过渡成了'解决问题'。"平 台负责人说。 据了解，通过新一代推理引擎，AI购物管家"张大妈"能将意图解析与任务执行相结合，实现购物流程一 站式管理，提升用户决策效率与消费体验。它可以根据人们的语音输入消费诉求并搜寻内容、给出结 果，未来还将追加下单、订票等一系列产品功能。目前，张大妈安卓尝鲜版已经可以扫码下载。 在发布购物智能体的同时，"火眼"AIUC引擎也同步亮相。值得买科技CTO王云峰介绍，相比AIGC（人 工智能内容生成）或Agent（智能体），AIUC并非行业最关注的概念，但在消费领域中极为重要，它能 充分利用目前互联网上的海量消费内容，运用理解、归纳、总结的能力，提炼出对消费者更有决策价值 的关 ...

Core Insights - The company held an AI progress sharing conference to showcase the results and advancements of its "comprehensive AI strategy" over the past year [1] - Key highlights included the introduction of the "Fire Eye" AIUC engine, the upgraded "What Worth Buying" GEN2, the shopping assistant "Zhang Dama," and the foundational "Haina" MCP Server [1] Group 1 - The AIUC engine leverages extensive internet content understanding to drive a new paradigm of intelligent decision-making [1] - The "What Worth Buying" GEN2 is AI-driven and focuses on interest-based consumer content, aiming to create a trusted guide for users [1] - The AI shopping assistant "Zhang Dama" facilitates a one-stop management of the shopping process, enhancing user decision-making efficiency and consumer experience [1] Group 2 - The "Haina" MCP Server is a standardized consumer data service platform based on the Model Context Protocol (MCP), which connects large models directly to data sources [1] - The CTO highlighted that the MCP protocol enhances the efficiency of AI ecosystems by improving the connection between agents and tools, allowing for focused optimization in various domains [2] - The company has launched the Haina MCP Server on its official website and integrated it into platforms like Alibaba Cloud, utilizing a vast database of 12 billion consumer content entries and 1 billion product entries to support AI applications [2]

Core Insights - Worth Buying Technology Group has launched multiple AI products, including the "Fire Eye" AIUC engine, two user-facing AI products ("What is Worth Buying" GEN2 and shopping agent "Zhang Dama"), and an AI infrastructure product ("Haina" MCP Server) [2] - The AIUC engine is designed to deeply understand and extract value from various content types, enhancing content production and business decision-making efficiency [2] - The company emphasizes the importance of AIUC in the consumer sector, stating it has improved content quality and process efficiency across its diverse business operations [2][3] Group 1 - The "Fire Eye" product focuses on helping consumers discern the authenticity of content, filter quality information, summarize genuine opinions, and enhance decision-making efficiency in consumption [3] - Worth Buying Technology Group is committed to a comprehensive AI strategy, prioritizing AI application research to transform cutting-edge AI technology into practical productivity [3] - The company aims to explore more possibilities of "AI + consumption" and actively share and exchange AI capabilities to contribute to the AI era [3]

5月9日下午，值得买（300785.SZ）2024年度业绩网上说明会成功在全景网举办。 交流会中，问及25年业绩增长点，值得买董事长兼首席执行官隋国栋强调，2025年，公司的整体战略方 针是"聚焦AI"和"全面提效"，推动AI技术与公司业务的融合、创新型AI产品 的研发以及内部管理和运 营效率的全面提升，着力强化、重塑各项重点业务的市场竞争力，以实现更好的业务发展和长 期成 功。2025年年度重点经营计划，主要有以下几个方面： （4）全面推进国际化战略，拓展全球市场。当前，全球电商市场，尤其是新兴国家的电子商务领域， 仍处于快速增长阶段，消费内容及内容营销服务需求旺盛。 公司通过调研发现，"什么值得买"在国内 成功创立的"值得买"模式在国际市场同样具备强大的竞争力。基于这一洞察，2024 年，公司正式启动 了"值得买"模式进入国际市场，首站选择了泰国，迈出了国际化布局的重要一步。2025年，公司将继续 深化"值得买"模式的全球化拓展，在进一步落地泰国站的基础上，加速向更多国家落地，持续提升品牌 影响 力和市场份额，为全球用户提供优质的消费内容和服务。（全景网） 更多业绩说明会详情，请点击：https://rs ...

证券代码：300785 证券简称：值得买 北京值得买科技股份有限公司 投资者关系活动记录表 编号：2025-001 | 投资者关系活 | □特定对象调研 □分析师会议 | | --- | --- | | 动类别 | □媒体采访 □业绩说明会 | | | □新闻发布会 □路演活动 | | | □现场参观  其他：电话会议 | | 参与单位名称 | 金鹰基金 陈颖 | | 及人员姓名 | 广东天贝合资产管理有限公司 黄达 | | | 建信养老金管理有限责任公司 高月 | | | 诺安基金 王晴 | | | 中金公司 王杰睿 | | | 华鑫证券 朱珠 | | | 广发证券 黄静仪 | | | 民生证券 李瑶 | | | 国盛证券 杜玥莹等 60 余人。 | | 时间 2025 | 年 4 月 29 日 15:30-17:00 | | 地点 | 电话会议 | | 上市公司接待 | 董事长兼首席执行官 隋国栋先生 | | 人员姓名 | 联合创始人兼消费内容事业部负责人 刘峰先生 | | | 智能营销及国际化事业部负责人 许欢先生 | | | 首席技术官兼 AI 事业部负责人 王云峰先生 | | | 首席财务官 李楠 ...