AI Applications in Project-Based Supply Chain Coordination and Risk Governance

Abstract

The increasing complexity of construction projects has underscored persistent challenges in supply chain coordination and risk governance. Traditional management tools often fail to address fragmented communication, information silos, and limited predictive capacity in risk identification. Large language models (LLMs), with their advanced natural language processing and reasoning capabilities, offer a transformative potential to enhance collaboration and adaptive governance in such contexts. This paper conducts a systematic literature review of recent studies across IEEE, Springer, Elsevier, and arXiv, classifying existing research by application scenario, methodology, and outcomes. The findings reveal that LLMs are most effective in improving communication and standardizing information exchange, while their application to predictive risk governance remains largely conceptual and underexplored. This highlights an important disparity between the predictive and assistive capabilities of Large Language Models (LLMs) in this area. Integrating digital platforms, for example, Building Information Modeling (BIM) and blockchain technology, is suggested as an eventual enabler of building the resilience and openness of supply chains. Nonetheless, the review presents significant gaps, for example, the lack of domain-specific fine-tuning, the lack of empirical verification in construction contexts, and the less-than-adequate integration from coordinationand governance-based viewpoints. The paper concludes by positing that LLMs should serve as collaborative infrastructures in supply chains best as they are optimally accessed using hybrid governance structures infusing artificial intelligence-based processes and blockchain as well as statistical procedures. Future work should focus on empirical case studies, system integration, and capacity building for closing the gap between potential applications on paper and practical application.