A collaborative and sustainable engineering design framework enabled by agentic AI

This study introduces an agentic-AI-enabled S5 collaborative framework for sustainable engineering design that integrates five concurrent dimensions: Sensing, Smart, Sustainability, Social, and Safe within a unified virtual environment. The framework employs agentic AI to coordinate optimisation, risk assessment, and environmental performance in real time, enhancing traceability, transparency, and informed decision-making across digital and physical domains. A case study of a mobile robot for greenhouse operations demonstrates the implementation and lifecycle traceability of the framework. Each agent performs domain-specific optimisation: the Sensing agent ensures perception accuracy and simulation fidelity, the Smart agent reduces path length and energy consumption, the Sustainability agent evaluates life cycle indicators aligned with circular economy principles, the Social agent enables human-AI co-design through immersive visualisation, and the Safe agent validates operational reliability through continuous monitoring. Quantitative results show a 15% reduction in total mass, 12% improvement in energy efficiency, 16.7% increase in recyclability, 92% task validation rate, and a 25% decrease in failure occurrence. These outcomes demonstrate that the S5 framework provides a traceable, data-driven, and human-centric methodology that unifies sustainability, intelligence, and safety. Its modular and scalable architecture supports adoption across domains, such as smart manufacturing, logistics robotics, and sustainable infrastructure.