The Effects of Cognitive and Organizational Factors on Human-Agentic AI Collaboration Acceptance : Moderating Role of Usage Frequency and Implications for AI Co-Creation Utility

The increasing adoption of agentic AI—autonomous AI agents capable of independent goal pursuit—requires a comprehensive understanding of the factors influencing employees’ willingness to collaborate with such systems. Grounded in technology acceptance theory, diffusion of innovations, and socio-technical systems perspectives, this study investigates the cognitive and organizational determinants of human–agentic AI collaboration, including perceived cognitive relief, agentic curiosity trigger, prompting confidence, training and guidance availability, and role compatibility perception. A survey of 300 employees from the IT, manufacturing, services, and financial sectors in Korea was analyzed using structural equation modeling. Results indicate that prompting confidence and role compatibility are the most influential drivers of human AI collaboration, while agentic curiosity trigger has no significant effect. Human–AI collaboration significantly enhances AI co-creation utility, mediating the effects of perceived cognitive relief, prompting confidence, training and guidance availability, and role compatibility perception. Multi-group analysis reveals that training support is especially critical for low-frequency AI users, whereas perceived cognitive relief is more influential among high-frequency users. These findings contribute to theory by introducing prompting confidence as a domain-specific self-efficacy construct and highlighting role compatibility as a robust predictor of AI collaboration acceptance. Practical recommendations are provided for tailoring organizational interventions according to employees’ AI usage experience to maximize the benefits of human–AI co-creation.