Indicators of External Disruptions in Supply Chains: A Framework for Early Detection and Resilience Planning

In an era of global volatility and interconnected supply chains, the ability to detect and interpret external disruptions is essential for maintaining operational resilience. This paper presents a multimodal, AI-driven framework for the early detection of disruption indicators, designed within the M4ESTRO[1] project to support trusted, explainable, and human-centered Manufacturing-as-a-Service (MaaS) environments. The component integrates real-time sensing from diverse sources-including financial APIs, global news platforms, and social sentiment channels-using explainable AI techniques such as anomaly detection, semantic analysis, and interpretable scoring. These indicators are contextualized within a federated data infrastructure to support proactive reconfiguration and supply chain adaptation. The system promotes transparent decision-making through structured indicators, while also laying the foundation for edge-deployable AI capabilities. This work contributes to the development of trustworthy IoT/AI systems aligned with Industry 5.0 principles, emphasizing explainability, collaboration, and resilience in decentralized industrial ecosystems.