Key points are not available for this paper at this time.
Introduction This study examines how infrastructure readiness (IR) contributes to the development of AI-enabled circular supply chain capability (AI-CSC) in the Pakistani automotive industry. Drawing on sociotechnical systems theory and the dynamic capability view, the study investigates the mediating role of socio-technical alignment (STA) and the moderating role of human capability fit (HCF) in the relationship between IR and STA. Methods A cross-sectional survey design was adopted. Data were collected from managers and technical experts working in automotive firms in Pakistan. A total of 388 valid individual responses were obtained and aggregated at the firm level. The proposed research model was analysed using partial least squares structural equation modelling (PLS-SEM) through SmartPLS 4. Results The findings show that STA has a strong positive effect on AI-CSC. IR also has a positive influence on both STA and AI-CSC. In addition, STA mediates the relationship between IR and AI-CSC, indicating that digital infrastructure investments generate greater value when they are translated into effective human-technology routines. The results further reveal that HCF strengthens the relationship between IR and STA, suggesting that infrastructure readiness is more effective when employees possess the skills, knowledge, and role alignment required to use AI-enabled technologies. Discussion This study contributes to the literature on AI-enabled circular supply chains by clarifying the socio-technical mechanism through which infrastructure readiness supports the development of circular supply chain capabilities in an emerging economy context. The findings highlight that technical investments alone are insufficient; their impact depends on the alignment between infrastructure, employees, and organizational routines. However, the results should be interpreted within the geographical and industry-specific context of the Pakistani automotive sector. Future research should test the model across different countries, industries, and institutional settings to assess its broader generalizability.
Alahmari et al. (2026) studied this question.