在人工智能（AI）掀起全球产业变革浪潮的当下，制造业作为国民经济的命脉，正处于从"数字化"迈 向"智能化"的关键节点。然而，在大模型落地应用的过程中，企业却普遍面临"业务用不起来、结果不 敢用、IT管不住"的困境。 1月份，美的集团股份有限公司（以下简称"美的集团"）旗下美云智数科技有限公司（以下简称"美云智 数"）正式发布新一代工业AI平台——美擎AIGC3.1平台（以下简称"美擎AIGC平台"）暨智能体工厂解 决方案。 近日，美云智数研发中心总经理兼CTO宋云报接受了《证券日报》记者采访，深度解析制造业如何通过 打造"统一底座"，破解AI落地难题。 本报记者 王镜茹 将智能体融入关键业务流程 谈及大模型在工业领域落地的最大障碍，宋云报认为，"幻觉"是主要挑战之一。能否将智能体融入关键 业务流程，取决于知识是否可控。"许多企业的知识库往往局限于简单的文档管理或向量索引，这种不 可治理的知识反而增加了大模型的不确定性。"他表示。 对此，美擎AIGC平台将知识视作一项"工程能力"来打造，强调从采集、加工、应用到治理全过程。知 识经过治理，智能体给出的结论才能得到业务人员和管理层的信任，这是企业级智能体和通用智能 ...

在人工智能（AI）掀起全球产业变革浪潮的当下，制造业作为国民经济的命脉，正处于从"数字化"迈 向"智能化"的关键节点。然而，在大模型落地应用的过程中，企业却普遍面临"业务用不起来、结果不 敢用、IT管不住"的困境。 在宋云报看来，企业AI应用的成功，最终不取决于某一个具体的模型或单一的应用，而是能否建立起 一个支撑智能体（Agent）长期稳定运行的平台。"智能体是入口，而平台决定了上限"。 在大量的企业实践中，美云智数发现，缺乏平台支撑的智能体通常会陷入三个主要困境：其一，业务端 无法应用，因为开发门槛过高，过度依赖少数经验丰富的工程师，致使AI难以融入日常工作流程；其 二，AI给出的结果无法追溯、不可解释，难以进入关键决策流程；其三，IT部门无法有效管控，数据权 限和运行状态不可控，导致应用场景受限。 针对这些痛点，宋云报表示，美擎AIGC平台从建立之初就有着明确的定位，它并非一个新的业务系统 或单一的模型，而是一个统一底座，能够连接企业内部的ERP、PLM、MES等现有系统的数据和流程， 并对外开放算力、算法模型及生态支持，为企业未来的持续发展预留空间。 那么，如何实现AI的规模化应用？宋云报进一步阐释了 ...

Core Insights - The AI agent market is expected to accelerate its commercialization and application in various sectors by 2026, with a focus on unlocking the next wave of AI business value [3][4] - Midea Group has launched the Meiqing AIGC 3.1 platform and intelligent factory solutions, emphasizing the importance of AI strategy and the development of AI agents [2][5] Group 1: Midea Group's AI Strategy - Midea Group aims to enhance its AI strategy significantly by 2025, focusing on building AI agent capabilities [2] - The company has developed over 13,000 AI agents in the past three years, with 158 agents contributing 95% of the value output, although only 0.2% have been commercialized [2][5] - Midea plans to achieve cost reductions of 600 million yuan (approximately 84 million USD) through AI by 2025 and over 900 million yuan (approximately 126 million USD) by 2026 [2][5] Group 2: Market Trends and Predictions - The AI agent industry is projected to grow from 5.1 billion USD in 2024 to 47.1 billion USD by 2030, with a compound annual growth rate of 44.8% [5] - Predictions for 2026 include a shift in AI competition from parameter comparison to practical implementation, with the year being termed as the "Year of the Billion AI Agents" [3][4] Group 3: Product Development and Market Positioning - Midea Group plans to release 20 to 30 new AI agent products, expanding its offerings to over 40 agents for external sales [3][4] - The newly launched AI products will target general scenarios as well as specific applications in manufacturing, research and development, marketing, and supply chain [4] - Midea aims to establish a competitive edge by focusing on domain-specific AI agents, recognizing the barriers to understanding application scenarios [5]

Core Insights - The core viewpoint of the articles emphasizes the accelerating commercialization of AI agents, with a focus on how companies are leveraging AI to unlock new business value and enhance operational efficiency in various sectors [1][2][4]. Group 1: AI Agent Development and Commercialization - In 2026, the competition in AI agents will shift towards unlocking the next wave of commercial value, with a significant increase in the number of AI agents being developed and deployed [2][4]. - Midea Group has launched the Meiqing AIGC 3.1 platform and intelligent agent factory solutions, highlighting the importance of AI strategy and the construction of agent capabilities [1][3]. - Midea has created over 13,000 AI agents in the past three years, with 158 agents contributing 95% of the value output, although only 0.2% have been commercialized [1][4]. Group 2: Financial Impact and Projections - Midea Group aims to reduce costs by 600 million yuan in 2025 and over 900 million yuan in 2026 through AI initiatives [1][4]. - The AI agent market is projected to grow from $5.1 billion in 2024 to $47.1 billion by 2030, with a compound annual growth rate of 44.8% [4]. Group 3: Strategic Focus and Future Plans - Midea's strategy involves a phased approach to AI capabilities, progressing from embedding AI in processes to collaborative work with AI, and ultimately to full automation through agents [4]. - The company plans to release 20 to 30 new AI agent products, expanding its offerings to over 40 agents for external sales, focusing on general and manufacturing-specific applications [2][3].

Core Insights - The AI agent industry is expected to mature significantly by 2026, with a focus on unlocking the next wave of AI commercial value [3] - Midea Group has launched the Meiqing AIGC 3.1 platform and intelligent agent factory solutions, emphasizing the importance of AI in its strategic development [1] - Midea aims to achieve cost reductions of 1.8 billion yuan in 2024, 6 billion yuan in 2025, and over 9 billion yuan in 2026 through AI implementation [5] Group 1 - Midea Group has developed over 13,000 AI agents in the past three years, with 158 agents contributing 95% of the value output, although only 0.2% have been commercialized [1][5] - The company plans to release 20 to 30 new AI agent products in 2026, expanding its offerings to over 40 agents for external sales [3][4] - The market for AI agents is projected to grow from $5.1 billion in 2024 to $47.1 billion by 2030, with a compound annual growth rate of 44.8% [4] Group 2 - Midea's AI strategy involves a three-step approach: from embedding AI in processes, to co-piloting with AI, and finally to full automation with AI agents [5] - The focus on domain-specific AI agents is seen as a way to build a competitive advantage, as general-purpose agents have a relatively low barrier to entry [5] - The company aims to effectively capture market share and customer mindshare by launching its AI capabilities at the beginning of 2026 [4]