Employee-AI collaboration errors and service innovation: dual pathways of error reflection and negative affective rumination

Employee-AI collaboration models are increasingly implemented in service organizations. Although such models enhance employee efficiency and service quality, errors in the collaboration process remain frequent. However, the effects of errors during human-AI collaboration on employees’ service innovation remain unclear. Drawing on the transactional theory of stress, this study investigates how employee-AI collaboration errors influence employees’ AI-enabled service innovation behavior. Based on a three-wave survey and an experiment, this study finds that when AI technology-task fit is high, employee-AI collaboration errors are more likely to trigger employees’ error reflection, which in turn facilitates their AI-enabled service innovation. In contrast, when AI technology-task fit is low, employee-AI collaboration errors tend to induce negative affective rumination, thereby undermining employees’ AI-enabled service innovation. Overall, this research extends prior literature on human-machine collaboration and service innovation while offering actionable insights for service organizations seeking to foster employees’ innovation behavior in the context of human-AI cooperation.