Artificial Intelligence-Driven Pathways for Net-Zero Supply Chain Transformation in SMEs

Abstract Achieving net-zero emissions (NZE) in supply chains (SC) has become a strategic imperative for organisations amid increasing regulatory pressure and global sustainability commitments. Limited studies have examined how artificial intelligence (AI) can transform SC to achieve NZE by unveiling the critical success factors (CSFs) that drive sustainable operational excellence in small and medium enterprises (SMEs) within emerging economies. Addressing this gap, the present study develops a structural model for SMEs to enhance AI adoption, addressing net-zero goals while improving competitiveness. It is also tested based on data collected from executives, managers, and senior managers of SMEs. The primary methodological approach used in this study is partial least squares-based structural equation modelling (PLS-SEM) and artificial neural network. Sixteen CSFs of AI adoption were identified through a systematic literature review, and their significance was validated through a survey of 100 Indian manufacturing SMEs. The results of PLS-SEM indicate that Smart Manufacturing and Process Optimisation (SMPO), Energy Efficiency and Emissions Reduction (EEER), Decision-Making and Regulatory Compliance, and Strategic Adoption and Stakeholder Engagement (SASE) exerted a significant influence on AI adoption in the net-zero supply chain (NZSC). The SEM outcomes were taken as inputs for the ANN approach. ANN results show that SASE (normalised importance = 100%) is the most influential CSF group in predicting AI adoption for NZSC, followed by Smart SMPO (normalised importance = 72.927%) and EEER (normalised importance = 68.100%) . The findings provide both theoretical and practical contributions. Theoretically, this study advances the discourse on digital sustainability by elucidating the pathways through which AI capabilities drive net-zero transformation in SC. Methodologically, this work combines PLS-SEM to validate structural relationships and a feed-forward ANN to capture nonlinear predictive importance. In practice, the results indicate that SASE is the top-priority factor, followed by SMPO and EEER, offering SME managers and policymakers a prioritised roadmap for AI investments to achieve net-zero outcomes. This hybrid approach, combined with the SME focus, constitutes the study’s primary scientific contribution. This work advances the field by offering the first integrated linear-nonlinear modelling framework for NZSC transitions in emerging economies, with immediate relevance for researchers and practitioners.