Implementation pathways of generative artificial intelligence in healthcare SMEs

Abstract While generative AI (GAI) promises transformative gains, its organizational embedding remains a complex process. Through an inductive study of healthcare sector organizations, we develop a dual-pathway GAI adoption model that elucidates how employee-initiated (bottom-up) and management-driven (top-down) routes interact. Specifically, this co-evolutionary interaction produces three sequentially unfolding mechanisms: (a) distributed issue-selling via continuous sensemaking, (b) facilitative trickle-down dynamics that enable rather than mandate initiative, and (c) the emergence of situated AI—organization-specific routines that are an inimitable organizational capability. We explain how these mechanisms manifest across learning, HRM, and innovation, revealing a persistent governance-flexibility dilemma and a scalability threshold for enterprise-wide adoption. By specifying these bidirectional, co-evolutionary processes, we advance technology implementation theory and provide a framework for policy-makers and administrators to navigate the ‘situatedness’ of GAI without stifling the grassroots experimentation essential for long-term organizational resilience.