Public Area

Members Portal

by German Vargas and Dario Melo

The Use of Generative Artificial Intelligence as a Simulation Tool for Advanced Training in Crisis Management

"Crisis simulation is the testing ground where strategic ingenuity in decision-making is developed, and organizational resilience is put to the test without posing real risk to the organization."

German Vargas



In an era where businesses face poly-crisis challenges ranging from data breaches and impacts from natural disasters to global pandemics, the ability to respond swiftly and effectively during a crisis has never been more essential.

Historically, crisis management training and preparation have relied on scripted simulations where predictable outcomes emerge from static scenarios. While effective on their own, these traditional methods often lack the dynamism and unpredictability of crises, as seen in the real world.

This is an invitation for those of us working in these disciplines to harness the transformative power of generative AI (GAI), a tool we are experimenting with to reshape the landscape of crisis management training. This advanced form of artificial intelligence is not merely an enhanced simulator but a revolutionary element that creates hyper-realistic and responsive scenarios, challenging even the most experienced professionals in ways previously unattainable.

As companies operate in an increasingly volatile and changing environment, leaders must be prepared to navigate the murky waters of unforeseen challenges. By using GAI, we have found that there is an opportunity to train in a test environment that closely mirrors the unpredictability and complexity of the real world.

In this article, we delve into the capabilities of GAI, explore its revolutionary approach to crisis simulations, and provide insights into how modern businesses can leverage this technology to strengthen their crisis response strategies.


1. A New Dawn for Crisis Simulations:

Traditionally, crisis management simulations functioned like a well-rehearsed play: predefined characters, established narratives, and predictable outcomes. Participants would play their roles within the confines of a known story, preparing for what was essentially a list of predetermined challenges. While these practices honed skills and improved team coordination, limitations are evident. The real world, with its chaotic interplay of variables, rarely adheres to a script.

With the advent of generative AI, the landscape of crisis management training has been radically transformed. Instead of a predictable play, it is conceived as a dynamic and improvised theater, where the story evolves in real time, adapting to decisions made on the fly.

Now, with generative AI, the realm of crisis management training has taken an exhilarating turn. Take, for instance, our experimentation with NovaTech, a fictional leading smart home device manufacturer. In our simulated scenario, they faced a whirlwind of public concern after their top product, HomeGuard, was found sending data to an undisclosed overseas server. This wasn't a real-world crisis but a fictional situation, crafted with such precision by the AI that it felt eerily plausible.


a) Evolving Real-World Scenarios:

Generative AI enables us to create scenarios that are not just static snapshots of potential crises but living situations that change and evolve. If a senior executive makes a decision in the simulation, we evaluate that choice using AI and subsequently recalibrate the scenario. This mirrors the cascading effects a decision can have in a real- world crisis, teaching participants to think several steps ahead.

In another of our simulated tests, a fictional biotech leader, NexaBioTech, had to navigate a crisis when a report emerged about their anti-aging drug, RevitaGen, having concealed harmful side effects. As we made decisions, the AI recalibrated the scenario in real-time, mirroring the multifaceted repercussions one might expect in the real world.

b) Multifaceted Context in Multi-Scenarios:

Unlike the singular, one-dimensional scenarios of traditional simulations, GAI-based simulations can integrate a multitude of variables: economic, social, technological, ecological, and more. For instance, while managing a crisis stemming from product defects, with AI, we could introduce an unexpected twist related to stock market fluctuations or a negative viral campaign on social media, requiring participants to juggle multiple challenges simultaneously.

In a particularly challenging test, participants at a fictional drone manufacturer had to respond when one of its drones malfunctioned during a school demo, causing injuries. Simultaneously, they faced stock market reactions, a viral social media campaign, and internal company disputes. The complexity layered into these scenarios pushed participants to multitask and strategize on the fly.

c) Customized Complexity:

One standout feature of the simulations we can power with GAI is the ability to tailor complexity to match participants' experience levels. Whether it's a novice manager trying crisis management for the first time or an experienced executive seeking an advanced challenge, we can modulate scenario complexities to match the learner's experience. For instance, with NovaTech's situation, we didn't just set the bar high; we actively requested the AI to increase the complexity and reduce the probability of success. It was akin to walking a tightrope with strong winds against you.

The CEO's role became pivotal; if they didn't adeptly navigate the intricate maze of challenges, the AI would simulate the ramifications of their dismissal. This kind of high- stakes environment truly tested the mettle of the team, pushing them to make nuanced decisions under immense pressure.

In essence, the fusion of generative AI with crisis management training signifies a “seismic” shift from static to dynamic, from predictable to unpredictable. By simulating the unpredictable nature of real-world crises, it trains and develops professionals in the adaptive skills necessary in today's volatile business landscape.


2. Dynamic Response Testing and Comprehensive Feedback: A Dual Power

As we intensify scenarios within the simulation, we also induce the need for immediate and informed responses, even in recreated low-data decision-making situations. Amidst a crisis, every decision carries the weight of potential consequences. Here lies the unparalleled strength of generative AI: With it, we not only recreate adaptive scenarios but also monitor, evaluate, and provide real-time feedback on participants' responses.


a) Real-Time Adaptability:

We employ generative AI not as a passive observer. When a participant chooses a particular course of action, we don't limit ourselves to reaching a predetermined outcome. Instead, we dynamically alter the course of the simulation based on that decision, effectively lobbing the ball back into the participant's court. This iterative back-and-forth mirrors real-world scenarios where one action can lead to a cascade of unexpected results.

b) In-Depth Analytical Insights:

Beyond the simulated scenario, AI continuously analyzes participants' decisions against a vast database of best practices, historical precedents, and potential outcomes. It's not just about "winning" or "losing" the simulation but understanding the nuances of decision-making in poly-crisis contexts.

c) Post-Simulation Reporting:

Perhaps one of the most valuable aspects of GAI-driven crisis simulation is the post- scenario feedback. Generative AI enables us to provide detailed insights into participants' performance, highlighting strengths, identifying areas for improvement, and offering practical recommendations. For instance, if, during a data breach scenario, a participant prioritized public relations and communication over immediate technical rectifications, AI can provide insights and recommendations on the importance of addressing vulnerabilities before and/or in tandem with managing brand image and even generate viable alternatives on how to do so.

d) Continuous Learning Loop:

Unlike traditional simulations that conclude once the scenario unfolds, GAI-driven simulations can be seen as an ongoing learning loop. Based on results and feedback received, we can encourage participants to re-enter the simulation armed with new insights and strategies and face a fresh set of challenges.

By incorporating dynamic response testing and comprehensive feedback, GAI ensures that the learning derived from each simulation is deeply ingrained, actionable, and tailored to individual needs. It's not just about surviving a crisis but mastering the art of effective strategic response, ensuring that when real-world challenges arise, leaders are not just trained but truly poised for success and the generation of well-being and prosperity.


3. A Paradigm Shift: From Preparedness to Mastery

Historically, crisis management has been grounded in preparedness: organizations develop protocols, train their employees through drills, and prepare responses to potential threats. However, as the business landscape evolves, with its intricate web of global relationships and rapid technological advancements, mere preparedness often falls short. It is in this new era that experts are orchestrating AI-driven simulations, marking the onset of a profound shift: from mere preparedness to true mastery.


a) Mastery through Repetition:

One of the fundamental principles of skill acquisition is repetition. The ability to engage in a scenario, receive feedback, adjust strategies, and then confront new challenges allows for deeply ingrained learning. Generative AI enables us to engage participants in this cycle multiple times, enhancing their decision-making capabilities with each iteration.

b) A Safe Space for High-Risk Decisions:

In the real world, a misguided decision during a crisis can have dire consequences, ranging from financial losses to tarnished reputation, and even jeopardizing the organization's survival. Simulations using GAI provide a secure environment where leaders can make high-risk decisions without real-world repercussions. This safe environment fosters audacity, innovation, and the willingness to explore unconventional solutions.

c) Bridging the Knowledge Gap:

In practice, each leader brings a unique set of experiences and knowledge to the table. However, blind spots always exist. By leveraging generative AI, with its extensive database of scenarios, strategies, and outcomes, we can identify these gaps and tailor simulations to address them. Over time, this results in a more holistic understanding of crisis management, reducing vulnerabilities arising from individual biases or limited experiences.



As the boundary between digital and physical worlds continues to blur, the challenges faced by businesses will only grow in complexity. While traditional methods of crisis management preparation remain valuable, they must be supplemented with more advanced, adaptable, and comprehensive tools.

We have conducted numerous practical exercises, combining GIA tools (ChatGPT, PI, and Bard). With the support of experts, we can confidently state that AI-driven generative crisis simulations represent the next frontier: a frontier where leaders will not only be prepared for crises but will truly master the art of resilience.