Breaking Data Silos in Demand Planning in Humanitarian Medical Supply Chains using AI

Humanitarian medical supply chains operate under extreme uncertainty, long lead times and fragmented data. Over 300 million people in 72 countries need humanitarian assistance, and supply chains are central to delivering life-saving supplies (Nieto, Jaschke & Piperno, 2025). This poster examines how data silos constrain demand planning, using Médecins Sans Frontières (MSF) as a case study. Interviews with field demand-planning professionals were used to map the data needed for needs forecasting and to rate each source for reliability and criticality to planning. Barriers and enablers to AI adoption were then analysed across financial, technical, ethical and cultural/organisational dimensions. The findings show that data fragmentation is a structural constraint on decision performance. Planning relies on manual consolidation and spreadsheets, planning cycles run 6–8 months, and the expert forecaster becomes the bottleneck. The poster proposes an AI-enabled data flow in which an agent harmonises country-level variation in Ministry of Health standard definitions before data are integrated. A hybrid integration layer then supports forecasting, scenario analysis and supply decisions, with humans kept in the loop. The aim is to improve organisational sense-making, not to automate decisions. The poster also outlines an implementation roadmap and future directions: digital twinning, external risk signals and wider multi-agency use. Keywords: humanitarian logistics; demand planning; medical supply chains; data integration; data silos; agentic AI; decision support; Médecins Sans Frontières