Organizational learning (OL) has commonly been used to explain the behavior of organizations when facing disasters. There is a growing consensus that disasters trigger OL, which enables organizations to handle the interruptions. Task performance experience is converted into knowledge that changes the organization’s context and affects its future experience (Argote and Miron-Spektor, 2011). Many scholars believe that disasters are of both high impact and relevance, which tend to produce learning that transforms organizations (Lampel et al., 2009; Desai, 2010; Madsen, 2009, Christianson et al., 2009), and that not being able to learn from disasters is dangerous (Alexander, 2012). However, the literature in this field is still divided. Disasters are relative concepts. They are hard to define and interpret. Situations that are disastrous for one organization may be less dramatic for another (Lampelet al., 2009). Disasters are also highly context based, complex and infrequent (Van Laere, 2013). Drawing scientific inferences from disasters that occurred at different periods in different contexts is challenging. Analyzing the OL at individual, group and organizational levels, this study advocates that disasters trigger the learning, and thus normalizes the learning through disasters. Disaster is the product of natural hazards and social and human vulnerability (Chmutina et al., 2017). The vulnerability of organizations combined with the natural hazard causes damages that are part of disasters. Our case study makes a valuable contribution to exploring the connection between disasters and learning. The 2012 Northern Italy earthquakes are natural hazards, which triggered disasters to the environment, local residents and organizations. AIMAG, a local multi-utility company that provides electricity, gas, water, public lighting and waste collection services to the earthquake-affected area faced major interruptions. Although it was unprepared, AIMAG resumed its services rapidly and recovered from the interruptions. It also established a schema to manage future disasters. Thus, the assumption that those involved in AIMAG learned from the disasters can be made. To investigate in detail how learning was triggered by disasters and what AIMAG learned through the interruptions, this research collected data from seven interviews with managers and from archived materials. By analyzing the data based on the multi-level learning model, this research discovered that the earthquakes caused operational, financial and administrative interruptions, which audited AIMAG’s knowledge repositories and triggered the learning process as outlined by Crossan et al. (1999). Findings support the concept of “learning through disasters,” which helps to “normalize” OL. By perceiving the earthquakes effects as a series of interruptions that may have happened in the past and may re-occur, the OL thus began long before the earthquakes struck. During the time of the disruption, knowledge was retrieved from the organization’s knowledge repositories and new knowledge was generated across all levels throughout the organization. The knowledge that proved to be effective was instituted into the organization. Routines, as the primary form of organizational knowledge, served as links to multi-level learning and connected past knowledge to the future.