In social emergency contexts, the ability to simulate realistic scenarios is crucial to support decision-making and improve operational intervention. This paper proposes an innovative approach based on the semantic representation of storytelling, applied to the construction of knowledge bases for conversational agents involved in simulation systems. Through the use of a storytelling ontology, action scenarios inspired by real situations are modelled, providing a structured and information-rich context to guide agent interactions. As a case study, the paper explores the role of firefighters in rescue operations, demonstrating how this methodology can improve understanding of emergency dynamics and foster more effective decision-making in multi-agent systems.