Twin_Transition_Mixed_Methods_Data

Employing a sequential explanatory mixed-methods design (QUAN to QUAL), this research dataset investigates the dynamics of the twin transition across 384 Thai service-sector small and medium enterprises (SMEs) spanning green agritourism, wellness and medical services, and green logistics and supply chain services. The quantitative phase utilizes partial least squares structural equation modeling (PLS-SEM) to evaluate the relationships among artificial intelligence capability (AIC), human-in-the-loop engagement (HITL), digital sustainability (DS), and Bio-Circular-Green (BCG) model alignment. The empirical findings indicate that technical AIC alone has a weak direct effect on DS, but generates substantial ecological outcomes when channeled through HITL governance acting as both a primary mediator and a positive moderator. Qualitative reflexive thematic analysis across 18 key informants further explains that unmanaged AI deployment leads to digital rebound effects, where surface-level efficiencies are negated by hidden cloud energy demands and data accumulation. Consequently, institutionalizing HITL oversight combined with national BCG scaffolding is critical to transforming foundational AI capability into meaningful and lasting digital sustainability.