
Core Insights - The AI industry is transitioning from a parameter competition to a focus on value return, emphasizing the importance of ROI as a key metric for businesses [1][2] Group 1: ROI and Application Value - This year marks the fifth anniversary of Lenovo's SSG, witnessing the shift of generative AI from a "fever phase" to a rational implementation stage, with ROI becoming the critical standard for measuring input-output [2] - Companies need to identify suitable tools and scenarios to effectively leverage AI technology, focusing on cost-effectiveness rather than merely pursuing larger model parameters [2] - In the B2B sector, the emphasis is on whether AI can deliver value in specific scenarios rather than the model's inherent capabilities [2][3] Group 2: Challenges and Solutions - The issue of hallucination rates remains a significant challenge in the application of large models, which cannot be completely eliminated in the short term [4] - Solutions to hallucination rates require a systematic approach, optimizing the entire process from model engineering to multi-modal interaction [4] - Lenovo's iChain supply chain AI demonstrates the effectiveness of a multi-agent collaboration approach, achieving a 90% accuracy in risk identification and reducing response time by four times [4] Group 3: Strategic Insights - The advancement of AI is not solely about technological breakthroughs but also involves a comprehensive system engineering approach, requiring a full-stack layout from infrastructure to industry-specific practices [6] - There is a notable difference in market preferences, with overseas clients favoring SaaS models for AI consumption, while domestic clients prioritize localization and hybrid deployment for data security and flexibility [6] - Lenovo's strategy of "internalizing and externalizing" its practices across 180 countries allows for tailored solutions that meet local demands [6]