The rise of agentic AI—systems capable of autonomous decision-making and task execution—marks a pivotal moment in the evolution of risk management and security. Most recently, Anthropic’s Claude October 2024 demo of agentic AI “taking over” a virtual machine demonstrated the following, all from a single prompt: finding the best spot for viewing the sunset over the Golden Gate Bridge, mapping the route, and scheduling the time. Anthropic’s series of demonstrations showed how agentic AI can conduct tasks autonomously, and a host of new use cases began to proliferate. Within days, a Wharton researcher showed how agentic AI can review a video of a construction site for potential safety concerns, finalizing in a prioritized list of action items.
While most security and risk organizations are only just beginning to explore this potential, the technology’s rapid advancements signal a near future in which AI systems will play a central role in detecting, adapting to, and mitigating threats. This shift has the potential to redefine physical security, crisis management, and travel risk management, offering unprecedented levels of customization, precision, recommendations, and efficiency.
What is Agentic AI?
Agentic AI represents a new frontier in artificial intelligence, combining the analytical power of traditional AI with the autonomy to make decisions and execute tasks independently. Unlike earlier models reliant on humans, agentic AI adapts to changing conditions, learns from interactions, and takes proactive actions to achieve objectives. While currently guided by human oversight in critical applications, agentic AI’s ability to synthesize data, predict outcomes, and coordinate actions across domains sets it apart as a transformative tool for dynamic and complex environments.
Current and Near-Term Applications of Agentic AI in Security and Risk Management
1. Physical Security: Proactive Threat Detection and Mitigation
Agentic AI will transform physical security by shifting from passive monitoring to active threat management. These systems will integrate data from environmental sensors, historical trends, and live video feeds to anticipate and neutralize risks.
Example Use Cases
- Behavioral Anomaly Detection:
Detects unusual patterns without predefined rules.
Example: At a major sporting event, the AI identifies someone loitering near restricted areas and triggers the deployment of security personnel.
- Autonomous Resource Deployment:
AI agents lock doors, activate alarms, or redirect personnel in response to evolving threats.
Example: During an active shooter event in a hospital, the AI secures emergency room entrances and directs staff to safety, in concert with PA systems, HRIS systems, and mass notification technologies.
- Dynamic Dispatching:
Agents autonomously deploy drones or robots for real-time hazard investigation.
Example: A facility’s AI dispatches a drone to investigate a suspicious package, minimizing human risk.
It’s important to clarify: humans will remain “in the loop” for some time, given both the critical nature of the tasks herein, as well as the ongoing risks of previously discussed challenges in Large Language Models or LLMs (i.e., hallucinations, overconfidence, etc). However, as commonly stated, “this technology is the worst it will ever be.” Agentic AI will continue to rapidly improve, and the frequency of hallucinations and overconfidence should diminish over time.
2. Crisis Management: Smarter Coordination and Decision Support
During crises, agentic AI will accelerate response times by synthesizing data, generating scenarios and recommendations, and automating communication.
- Dynamic Scenario Modeling:
Adapts crisis response plans based on evolving data.
Example: During a wildfire near a manufacturing plant, AI predicts wind shifts and devises an evacuation route in real time.
- Integrated Communication:
Tailors crisis messaging for different stakeholders to ensure clarity.
Example: During a blackout, the AI system sends customized alerts to store managers, staff, and customers (each with a different interest in the blackout; each with tailored messaging) to coordinate closures effectively.
Autonomous Decision Trees:
Creates and executes multi-step plans without human intervention.
Example: A university AI agent identifies dormitories to evacuate during a gas leak, prioritizing those closest to the source.
3. Travel Risk Management: Personalized Safety and Itinerary Optimization
In the future, agentic AI will redefine travel security by acting as a proactive risk advisor and crisis mitigator for travelers.
- Personalized Risk Analysis:
Evaluates individual travel plans against dynamic risk factors like local protests or natural disasters.
Example: An AI agent warns an executive about a high-risk area, recommending alternate accommodations. The agent “understands” the profile of the executive, what they care about, their home country, their risk profile, and the needs of the business — and tailors recommendations based on these characteristics.
- Autonomous Itinerary Adjustments:
Proactively rebooks flights or arranges alternative transport during disruptions.
Example: During a rail strike, in coordination with the corporate TMC, the AI rebooks hotels and cars for employees, ensuring minimal disruption.
- Traveler Monitoring and Response:
Continuously tracks employee locations and issues safety updates.
Example: A corporate AI tracks employees in conflict zones, issuing immediate evacuation orders when violence erupts.
Note: Just as described above, humans (in this case, the TMC) will remain “in the loop” due to the same risks described above: hallucinations, etc. The use cases above will enable TMC agents to operate with greater efficiency and scale.
What Sets Agentic AI Apart?
Agentic AI distinguishes itself by combining advanced learning, automation, and integration capabilities to address complex challenges.
- Self-Learning and Adaptability: Unlike traditional systems, agentic AI refines its understanding of threats and responses over time, improving with each iteration. For example, an AI system might learn to better predict evacuation needs after analyzing post-crisis outcomes.
- Task Automation at Scale: By autonomously handling routine yet critical tasks—like monitoring, alerting, and resource allocation—agentic AI reduces human error and ensures continuous, reliable operation.
- Seamless Cross-Domain Integration: These systems connect diverse tools and teams, streamlining communication and decision-making across security, operations, and crisis management for cohesive action.
- Refinement through Deep Customization: In each of the use cases above, understanding “What would the right set of responses be for my (specific) organization?”
This combination enables organizations to tackle risks more efficiently and effectively, making agentic AI a transformative asset.
The Role for the Innovative Security Leader
We frequently hear this sentiment during Q&A sessions after speaking on AI: “So, the technology is evolving rapidly…and agentic AI could be a step-change…what then, for us?” The rise of agentic AI provides security leaders with an unprecedented chance to elevate their personal and professional impact. By first exploring — and then championing — this technology, security leaders can establish themselves as forward-thinking pioneers, strengthening their credibility and influence within their organizations and across the industry. As a practical matter, we typically see some of the best exploration from the “bottom of the org chart,” whereby junior members of security teams explore new use cases to support their day-to-day tasks. Still, we encourage strong guardrails to prevent harm — again, from hallucinations, overconfidence, and particularly in systems to ensure information security.
More generally, leaders who skillfully integrate agentic AI into operations will not only drive measurable improvements in resilience and efficiency, but also demonstrate how small, agile teams equipped with AI can achieve levels of security once thought possible only with larger resources. Ultimately, this is about driving business outcomes. This ability to amplify outcomes positions innovative leaders as indispensable architects of the future of security — and in their organizations.
The Road Ahead
Agentic AI offers a glimpse into a future where organizations can better anticipate and respond to evolving risks. While its adoption is still in its early stages, its transformative potential is massive, particularly in domains like physical security, crisis management, and travel risk management. Organizations that begin exploring these technologies today will position themselves to respond more effectively to an unpredictable world. This period of rapid evolution presents a powerful opportunity to enhance resilience and preparedness for the challenges ahead.
At Crisis24, we remain at the forefront of agentic AI’s evolution, helping organizations unlock its full potential. Let us guide you in integrating these groundbreaking systems into your security and risk management frameworks.
Author(s)
