The rapid growth of global supply chains has brought major environmental challenges, particularly through rising transport-related emissions. In response, businesses and policymakers increasingly view artificial intelligence (AI) as a tool for achieving both operational efficiency and sustainability. This paper analyses the potential of AI-driven route optimisation within the context of green supply chains. The theoretical framework contrasts AI with classical optimisation methods, highlighting how machine learning, heuristics, and reinforcement learning introduce adaptability, scalability, and multi-objective optimisation that traditional models lack. The empirical part reviews three areas of AI application. Transport and logistics illustrate direct contributions through route optimisation, last mile efficiency, and multimodal planning. Predictive analytics improves demand forecasting and risk anticipation, reducing unnecessary trips and inventories. Digital twins and simulation offer advanced modelling for complex logistics systems, supporting resilience and sustainability under uncertainty. Practical illustrations demonstrate the tangible impact of these technologies. The paper concludes that AI-driven optimisation is not merely a technical innovation but a structural enabler of the green transition, aligning operational efficiency with sustainability objectives and paving the way for more resilient global value chains.