Synthesising Organisational AI Adoption: Topic Modellingand Thematic Coding through an Engineer Lens

Abstract

Artificial intelligence (AI) is increasingly embedded in organisations, offering efficiency and innovation but also introducing risks, making adoption a continuing challenge. Governance remains fragmented, with limited understanding of how stakeholders perceive trust, risk, and benefit. Engineers, as both builders and users of AI systems, are particularly influential yet remain under-explored in adoption research. This study addresses this gap by integrating the Extended Valence Framework and the Dual Trust Model through a stakeholder lens. Using a three-stage design, we constructed a corpus of organisational AI adoption literature with BERTopic modelling, applied coding and thematic analysis to identify and classify 75 factors across cognitive and affective dimensions, consolidated them into three themes, and conducted a stakeholder-specific review. The study contributes a unified research model for stakeholder-specific AI adoption, demonstrates a mixed quantitative-qualitative approach, and establishes a scalable foundation beginning with engineers and extending to other stakeholders through future empirical validation.

Ähnliche Arbeiten