Artificial intelligence in sustainable supply chain resilience with predictive analytics as a mediator and environmental uncertainty as a moderator

Recent disruptions caused by pandemics, climate change, and geopolitical conflicts have exposed the vulnerability of supply chains, particularly in developing economies with limited digital infrastructure. This study investigates how artificial intelligence can increase sustainable supply chain resilience through predictive analytics. It explores the influence of environmental uncertainty on sustainable supply chain resilience in Bangladesh’s manufacturing and pharmaceutical industries. Data were collected from 372 managers of manufacturing and pharmaceutical industries in the Dhaka and Chattogram divisions. PLS-SEM was used to analyze the proposed model in SmartPLS 4. MGA was also performed to compare the structures in the two sectors. Results showed that the use of AI has a positive impact on predictive analytics (β = 0.36, p < 0.001) and sustainable supply chain resilience (β = 0.17, p < 0.001). Predictive analytics also has a significant positive effect on sustainable supply chain resilience (β = 0.36, p < 0.001) and partially mediates the relationship between AI and resilience (β = 0.13, p < 0.001). There is a negative moderating effect of environmental uncertainty on the connection between AI and resilience (β = − 0.13, p = 0.001). The model explains 13% of the variance in predictive analytics and 58% in sustainable supply chain resilience. A second analysis of the results across the different groups indicates that the negative moderating impact of environmental uncertainty is significantly greater in the manufacturing sector than in the pharmaceutical sector. This study extends the Resource-Based View and Contingency Theory by conceptualizing AI as a strategic technological resource. Predictive analytics is conceptualized as an organizational capability. It illustrates that analytical power is the key component of AI for creating resilience, but certain environmental variables may determine its success. This study combines mediation, moderation, and sectoral comparisons, providing realistic and useful insights for managers and policymakers in emerging markets.