Embodiments disclosed herein provide techniques for dynamic response management. Weather data for a geographic region is received, and a state of the geographic region is determined based on the weather data. A first predefined rule of a plurality of predefined rules is identified based on the determined state of the geographic region. A first role of a first user is determined. A first instruction is then generated for the first user based on the first predefined rule and the first role, where the first instruction comprises one or more actions. The first instruction is then transmitted to the first user.