From reactive to proactive: how artificial intelligence is transforming organizational crisis management

Purpose This paper examines how artificial intelligence (AI) can improve organizational crisis management as firms increasingly operate in a polycrisis marked by overlapping pandemics, geopolitical conflict and rapid technological change. It maps AI capabilities to the before, during and after phases of the crisis lifecycle. Design/methodology/approach This narrative/conceptual review synthesizes empirical and theoretical scholarship on AI in crisis management across management, information systems and crisis communication journals. It integrates evidence on early warning, anticipation, social media analytics, crisis communication and resilience into a lifecycle framework. Findings Before a crisis, AI-driven early warning systems can predict disruptions weeks in advance and help mitigate losses. During a crisis, natural language processing supports situational awareness, and generative AI accelerates communication. After a crisis, social listening and simulation support learning, while deepfake defense protects reputation. Benefits are conditional on governance maturity and complement, rather than replace, human judgment. Practical implications Firms should adopt AI proactively but judiciously for sensing, triage and rehearsal, guided by governance readiness rather than novelty. They should also invest in data governance and interoperability. Organizations should keep accountability and ethical judgment visibly human where trust and empathy matter most. SMEs benefit from modular, centrally supported adoption pathways. Originality/value This paper integrates AI and crisis management into a lifecycle framework centered on the governance conditions that enable AI to improve outcomes. It advances two propositions: (1) AI’s value shifts from augmenting anticipation to substituting risk as authorship becomes visible, and (2) governance maturity moderates, rather than predicts, AI’s crisis-management payoff.