Against the backdrop of deep digital transformation and the increasing integration of AI into new product development (NPD), understanding the associations between different forms of human–AI interaction and team innovation has become an important organizational challenge. Grounded in the knowledge-based view and the Stimulus–Organism–Response (SOR) framework, this study investigates the associations between different forms of human–AI interaction and team innovative behavior, and the mediating role of knowledge domain coupling. Drawing on survey data collected from 268 NPD team managers serving as key informants for their respective teams in Chinese manufacturing firms, we examine two distinct forms of human–AI interaction: enhanced human–AI interaction (EHI) and impeded human–AI interaction (IHI). The results show that EHI is positively and significantly associated with both knowledge domain coupling and team innovative behavior, whereas IHI exhibits an inverted U-shaped relationship with knowledge domain coupling and team innovative behavior. Furthermore, complementary knowledge coupling and iterative knowledge coupling jointly serve as the knowledge-processing mechanisms linking human–AI interaction and team innovative behavior. This study contributes to the human–AI interaction and team innovation literature by explaining how different forms of human–AI interaction are associated with team innovation through distinct knowledge-processing mechanisms. The findings also offer practical guidance for managers seeking to dynamically balance different forms of human–AI interaction to foster innovation in AI-enabled NPD teams.