How Artificial Intelligence Enhances Construction Supply Chain Resilience Through Supply Chain Integration: A Mixed-Methods Study

Abstract

Construction supply chains (CSCs) are increasingly exposed to material shortages, demand fluctuations, logistics disruptions, and inter-organizational coordination failures. Artificial intelligence (AI) offers new opportunities to improve construction supply chain resilience (CSCR) by strengthening prediction, information processing, and collaborative decision-making. However, the mechanisms through which AI capabilities enhance CSCR remain insufficiently understood. Drawing on organizational information processing theory (OIPT) and dynamic capabilities theory (DCT), this study examines whether AI capabilities affect proactive and reactive CSCR directly or indirectly through three dimensions of supply chain integration (SCI): operational, information, and relational integration. It further compares the relative strengths of these pathways. This research adopts an explanatory sequential mixed-methods design. In the quantitative phase, 353 valid questionnaires from construction professionals in China were analyzed using partial least squares structural equation modeling (PLS-SEM). In the qualitative phase, semi-structured interviews with 15 experts, alongside three real-world cases, were utilized to interpret the quantitative findings and identify contextual boundary conditions. The results demonstrate that AI capabilities have significant positive effects on both proactive CSCR (β = 0.140, p < 0.01) and reactive CSCR (β = 0.116, p < 0.05). Furthermore, AI capabilities significantly promote operational integration (β = 0.299, p < 0.001), information integration (β = 0.361, p < 0.001), and relational integration (β = 0.227, p < 0.001), which in turn enhance both resilience dimensions. Notably, information integration is an important aspect of proactive resilience (β = 0.290, p < 0.001), while operational integration is crucial for reactive resilience (β = 0.274, p < 0.001). The qualitative findings further indicate that environmental uncertainty, technical readiness, and top management support condition the effectiveness of AI-enabled SCI. Theoretically, grounded in OIPT and DCT, this study clarifies the pathways through which AI affects CSCR and the contextual conditions shaping these effects, thereby advancing the analytical framework for AI-driven resilience. Practically, it delivers tiered implementation guidance for construction stakeholders to deploy AI tools for layered integration, thereby specifically enhancing both pre-disruption proactive risk prevention and post-shock reactive recovery capacities.