摘要：场景驱动创新已成为数智化背景下制造企业价值创造的新范式，而AI可供性被认为是衡量技术能否有效赋能创新的关键变量。然而，在制造情境下，AI技术向创新绩效转化的内在机理仍存在“黑箱”，尤其是在不同场景创新模式下，这种转化路径的差异性尚不明确。基于动态能力理论，本文探讨了AI可供性对制造企业创新绩效的影响，以及数字敏捷性和场景驱动创新的关键作用。具体而言，本文开发了场景驱动创新模式的测量量表。遵循“技术→能力→绩效”的研究逻辑，本文全面阐明了AI可供性、数字敏捷性、场景驱动创新与创新绩效之间的关系，并提出了研究假设。随后，本文开展了实证研究以检验假设并得出结论。研究发现:场景驱动创新包括主动创造型场景驱动创新和被动适应型场景驱动创新; AI可供性对制造企业创新绩效具有显著的正向影响;数字感知敏捷性和数字适应敏捷性在AI可供性转化为创新绩效的过程中发挥了显著的中介作用;主动创造型和被动适应型场景驱动创新模式分别有效调节了上述数字敏捷性的中介作用。异质性分析进一步显示，AI可供性的促进作用在非国有企业及数字化基础较弱的传统制造行业中表现出更强的边际提升效应。在理论方面，本文通过数字敏捷性的中介效应打开了AI可供性影响企业创新绩效路径的理论“黑箱”，同时引入场景视角，丰富了不同场景驱动创新模式下AI技术应用机制的理论探讨。在实践方面，本文认为制造企业应重视并培育数字敏捷性，以实现对AI技术的有效利用;同时，企业应根据外部动态环境和业务需求灵活选择适合的场景驱动创新模式，从而最大化发挥AI可供性在推动高水平创新绩效中的驱动作用。  

Abstract: AI technology is an important technical guarantee for enterprises to carry out scene driven innovation. Availability is considered to be an important variable to measure whether AI technology can effectively enable innovation. However,the mechanism of AI affordance adapting to scene innovation and improving innovation performance is still not clear. Following the research logic of “technology capability performance”,this paper takes digital agility as the intermediary variable and scenario driven innovation mode as the moderator variable to explore the impact of AI affordance on innovation performance of manufacturing enterprises. The results show that AI affordance has a significant positive impact on enterprise innovation performance,in which digital perceptual agility and digital adaptive agility play a mediating role,and active innovation and passive adaptive scenario driven innovation modes adjust the above intermediate roles respectively. Heterogeneity analysis further shows that the promoting effect of AI affordance on innovation performance is relatively stronger in non-state-owned enterprises and other manufacturing industries. This paper takes manufacturing enterprises as the research object,enriches the discussion on the role mechanism of AI technology in the scenario driven innovation mode, and has important implications for Chinese manufacturing enterprises on how to use AI technology to achieve high innovation performance at the practical level.