Understanding the drivers and barriers of human–AI knowledge orchestration: insights from social exchange theory

Purpose Despite the application of generative artificial intelligence (GenAI) in organizational workflows and business processes, many firms have not achieved improvements in innovation capabilities and value creation. This suggests that capability development depends not only on artificial intelligence (AI) but also on the effective orchestration human and artificial knowledge in work routines and business activities. This study explores the mechanisms underlying human–AI knowledge orchestration (KO) and identifies drivers, inhibitors and boundary conditions shaping managers’ participation. Design/methodology/approach Integrating knowledge orchestration theory and social exchange theory, the study conceptualizes knowledge growth, AI self-efficacy (AISE), perceived organizational support as benefits, execution cost (EC) and knowledge power loss as costs and trust in AI as the moderator. The model is tested using structural equation model with data from 341 managers in AI adopting firms. Findings Knowledge growth, AISE and perceived organizational support positively affect human–AI KO, whereas EC and loss of knowledge power (LKP) have negative impact. Trust in AI strengthens the positive effect of knowledge growth and AISE but unexpectedly exacerbates negative impact of EC and LKP. Social implications This study highlights the importance of fostering trust and addressing barriers in human–AI collaboration, offering insights into the responsible adoption of AI in the workplace. The findings may contribute to more inclusive digital transformation and AI-enabled knowledge management practices. Originality/value By conceptualizing human–AI KO as a dynamic knowledge management process embedded in workflows and business processes, this study advances research at the intersection of AI, knowledge management and business process management. It responds to calls to embed knowledge management in work routines and examine individual-level factors in AI-supported business processes. It reveals the benefit–cost trade-off underlying managers’ participation and provides implications for AI-enabled knowledge management, business process improvement and capability development.