House fires are an all-too-common real-world problem threatening life and property. However, what if we had the tools to predict the likelihood of future house fires? We use Activity Theory (Vygotsky 1978; Leontiev 1981; Engeström 1987), as the lens to examine how outputs of an information system created using design science principles (Dreschler, 2015) help us predict the likelihood of future house fire instances. We use two different AI models to present data through six tools in the information system, that include tables, lists, maps and numeric indicators that could be used by households and emergency management practitioners to inform decision making in how and when to prepare for and respond to potential future instances of house fires. This paper presents the results of 10 practitioner interviews and 200 survey responses on how these tools would be useful, and their value in decision making.