Midea Group(SZ:000333) Zheng Quan Ri Bao·2026-02-01 22:44Core Viewpoint - The manufacturing industry is at a critical juncture of transitioning from "digitalization" to "intelligentization," facing challenges in AI application due to issues like usability, interpretability, and IT governance [1] Group 1: AI Platform Development - Midea Group's subsidiary, Meiyun Zhishu, launched the Meiqing AIGC 3.1 platform aimed at addressing the difficulties in AI implementation within manufacturing [1] - The success of AI applications depends on establishing a platform that supports the long-term operation of intelligent agents, rather than relying on specific models or applications [2] - The Meiqing AIGC platform is designed as a unified base that connects existing enterprise systems (ERP, PLM, MES) and provides external computational power and algorithm support [2] Group 2: Standardization and Scalability - The platform aims to transform intelligent agents from custom developments into replicable assets through standardized construction, evaluation, and operation [3] - It includes over 50 intelligent agent templates and 500 enterprise connectors, significantly lowering development barriers with no-code or low-code technologies [3] - A standardized evaluation system with over 20 metrics is established to quantify the performance of intelligent agents before deployment [3] Group 3: Knowledge Management and Decision-Making - The integration of intelligent agents into key business processes is hindered by the "hallucination" effect, which is linked to the governance of knowledge [4] - The Meiqing AIGC platform treats knowledge as an engineering capability, emphasizing the entire process from collection to governance to ensure trust in the outputs [4] - The platform boasts a semantic retrieval accuracy rate that exceeds mainstream open-source solutions by over 20% [4] Group 4: Focus on Manufacturing Scenarios - Meiyun Zhishu is committed to focusing on industrial scenarios rather than participating in generic model development, leveraging Midea Group's extensive manufacturing experience [6] - By 2025, Midea Group anticipates achieving over 600 million yuan in efficiency gains through AI applications, with cost reductions in certain scenarios reaching up to 90% [6] - The company has developed over 13,000 intelligent agents internally, identifying high-value scenarios for commercialization [6]