The COVID-19 pandemic presented unprecedented challenges to public health systems and socio-economic stability worldwide, necessitating rapid, computational, and evidence-based decision-making under extreme uncertainty. This paper examines the COVID-19 AI & Simulation Project, a pioneering initiative launched by the Japanese Cabinet Secretariat in 2020, which exemplifies the integration of advanced computational modeling with public policy formation. The project developed a comprehensive framework for pandemic response that bridged the gap between scientific analysis and practical policy implementation by deploying artificial intelligence, complex network analysis, multi-agent simulations, fluid simulation, and laser optics. We present three detailed case studies demonstrating the project’s impact on critical policy decisions, including infection prevention protocols, the timing of imposing and lifting emergency measures, and vaccination strategy optimization. The project’s outcomes highlight the potential of computational approaches in crisis management while revealing essential insights about the limitations and challenges of implementing model-based policy recommendations in real-world scenarios.