Goal: Predicting the evolution of commodities price to improve anticipation to supply-chain disruptions is hard. We propose an approach based on an event detection model on data stream to assist experts in such task. The final goal is to report to experts a meaningful description of the most impactful events occurring around the world, to help them in their daily decision-making. Design / Methodology / Approach: This work results from a cross-fertilization between business management, Information Technology and Computer science. This work relies on an expert analysis and advanced AI engines, including a case study on a specific raw material and a literature review to define the parameters to supervise. Results: We propose a general architecture based on IT and business synergy. We conduce a general study on the factors influencing raw materials price fluctuations, namely events influencing supply and demand of the commodity. Finally, we present a case study of the events, which historically affected phosphates prices. Limitations of the investigation: An in-depth knowledge of the domain is needed to analyze and quantify the events impact on the supply chain. Practical implications: This approach was first designed for assisting raw material purchasers but it can potentially be reproduced to assist other decision-makers. Originality / Value: We propose a new approach on how to anticipate the implications of external events on supply chain disruption and raw materials price evolution. This method is multidisciplinary, involving expert domain knowledge and state-of-the-art artificial intelligence.