Exploring paradoxical barriers to AI adoption through the TOE framework

This study examines the barriers hindering the adoption of artificial intelligence (AI) in small- and medium-sized enterprises (SMEs). Drawing on the Technology–Organization–Environment (TOE) framework, the study employs a qualitative research design based on 12 semi-structured interviews with owners and managers from 12 Finnish manufacturing SMEs, selected through purposive and snowball sampling. The data were analyzed using a hybrid inductive–deductive thematic analysis to capture both emergent insights and theoretically grounded patterns. The findings identify key technological, organizational, and environmental constraints, including fragmented data, legacy IT infrastructure, skill shortages, employee resistance, weak institutional support, and regulatory uncertainty. Importantly, the study reveals paradoxical elements, such as strategic caution and employee resistance, that function as adaptive mechanisms regulating the pace of AI adoption and helping firms avoid premature or misaligned implementation. This highlights the dynamic and interdependent nature of TOE dimensions in resource-constrained contexts. The study contributes to the literature by extending the TOE framework beyond static interpretations, conceptualizing barriers as context-sensitive and adaptive elements in early-stage AI adoption. It also provides practical insights for SME managers and policymakers, emphasizing the need for data readiness, workforce development, and targeted institutional support.