李开复：智能体的商业价值最大化会在企业端

Core Insights - The value of AI agents for enterprises is currently underestimated, with a focus on their delivery capabilities [1][4] - The launch of the "Wanzai" super employee AI agent and the "Wanzhi" enterprise model platform 2.0 marks a significant advancement in AI capabilities for complex tasks [3] Group 1: AI Agent Development - AI agents are categorized into three types: workflow agents, reasoning agents, and multi-agents, with rapid development expected in workflow agents by 2024 [3][4] - Workflow agents will excel in executing simple tasks, while reasoning agents will enhance task breakdown and analysis capabilities by 2025 [3] Group 2: Commercial Value and Implementation - The maximum commercial value of AI agents will be realized at the enterprise level, emphasizing the importance of delivery value [4] - The development of AI agents should transition from cost reduction to efficiency enhancement, involving deep collaboration with industry partners to identify high-value scenarios and build industry knowledge bases [4] - The industry is shifting from model-driven to scenario-driven approaches, with enterprises more willing to pay for value [4]