Navigating the Multidimensional Misfit : A Process Study of Generative AI Adoption in Small and Medium Sized Enterprises

Sammanfattning

Generative AI marks a categorical shift in artificial intelligence, but its organisational uptake has been uneven. While large firms are moving toward strategic integration, small and medium-sized enterprises (SMEs) remain disproportionately exposed to the frictions of implementation. Existing adoption research has concentrated on antecedent conditions and treated AI as a homogeneous category, leaving the process by which SMEs navigate GenAI's distinctive implementation challenges insufficiently understood.This study examines how SMEs navigate and manage the multidimensional misfit that emerges during GenAI implementation. Drawing on Ansari et al. (2010), it develops a process-level understanding of how technical, cultural, and political tensions arise, interact, and are resolved or compounded under SME-specific constraints.A qualitative, exploratory, and abductive design grounded in social constructionism. Ten semi-structured interviews were conducted with informants from European SMEs across technical services, IT, manufacturing, construction, and cosmetics. Data were analysed in three coding cycles following Gioia et al. (2013), condensing 389 open codes into six tension dimensions.Six interrelated tensions structure the SME GenAI adoption experience: Strategic Ambiguity, Governance Paradox, Capability-Expectation Gap, Human and Organisational Readiness, Knowledge and Learning Deficit, and Regulatory and Technical Friction. They form a sequential and mutually reinforcing structure in which the individual consistently substitutes for absent organisational architecture. Contrary to adaptive-cycle assumptions, friction in resource-constrained SMEs tends to compound rather than self-correct, recasting GenAI implementation as a cumulative rather than self-balancing process.