Enablers and barriers to AI adoption: evidence from the heavy machinery industry

Abstract Artificial intelligence (AI) holds significant potential for heavy machinery manufacturing, yet adoption in this safety–critical and highly customized industry remains limited and insufficiently understood. This study examines enablers and barriers to AI adoption through a multiple-case study of heavy machinery manufacturers, analyzing dynamics across external, organizational, and individual levels. The findings show that AI adoption is shaped less by technological maturity than by safety requirements, regulatory complexity, organizational capabilities, and human expertise. Safety emerges as a central lens guiding adoption decisions across all levels. Simulation plays a key enabling role by supporting safe development, validation, training, and coordination, while reducing uncertainty about AI reliability. Adoption follows a hybrid and incremental logic, with firms retaining humans in the loop and expanding AI-supported decision-making as reliability and confidence increase. Organizational orchestration capability is critical for aligning technological possibilities with regulatory, organizational, and human constraints. By focusing on heavy machinery manufacturing, this study extends AI adoption research to an underexplored industrial context and clarifies how enablers and barriers shape AI adoption pathways.