TLDR: Leidos and NVIDIA have partnered to develop Command-and-Control AI (C2AI), an agentic AI system designed to drastically cut down emergency response times during disasters. C2AI uses a network of specialized AI agents to automate information flows, enhance situational awareness, and provide rapid, data-driven recommendations to human responders, thereby easing cognitive load and accelerating critical decision-making.
The article, “Agentic AI Aims to Cut Down Emergency Response Time When Disasters Strike,” published on Leidos.com on October 22, 2025, highlights a groundbreaking collaboration between Leidos and NVIDIA. Their joint effort focuses on developing Command-and-Control AI (C2AI), an advanced agentic AI system engineered to significantly reduce emergency response times during critical disaster events.
The Challenge of Disaster Response: Disaster scenarios are characterized by high stakes, intense manual tasks, and immense mental pressure on responders and management teams. Traditional command and control processes can be time-consuming, with critical information flow from 911 operators to incident commanders often experiencing friction. Corey Hendricks, VP and Chief Engineer of Leidos’ Commercial and International Sector, emphasizes this, stating, “Command and control in disaster management can take time when seconds matter, getting information from 911 operators to the incident commander and determining an action plan.”
C2AI: A Multi-Agent Approach: C2AI is built upon a network of task-specific autonomous agents designed to transform emergency response into a more efficient and automated process, fostering collaboration between humans and AI. The system aims to serve as “nimble eyes and ears” for responders, automating crucial information flows and processing multiple data streams to offer situational insights for more effective decisions.
Key Functions and Workflow:
911 Call Transcription: When a disaster strikes, such as a gas explosion, 911 operators are overwhelmed. An AI agent transcribes incoming distress calls, alleviating the multitasking burden on operators and allowing them to remain focused on callers.
Incident Report Generation: An AI orchestration agent routes these transcripts to a third agent, which parses them to create detailed incident reports. These reports, including incident locations, injury numbers, and severity, are then displayed to operators via a chat interface.
Real-time Radio Monitoring: The transcription agent also monitors first-responder and EMS radio communications, transmitting vital details to the orchestration agent.
Visual Alerts and Event Detection: Concurrently, other AI agents flag and tag events from municipal video camera feeds. These “visual alerts agents” provide real-time updates through chat. Hendricks notes the potential: “There are cameras located all across most cities, so imagine if you have AI agents monitoring all those camera feeds 24/7. And instead of static image classification, they can do event detection.” This machine vision capability can detect events humans might miss, correlating visual data from various feeds to build a comprehensive situational awareness picture.
Augmented Decision-Making: The system is designed to augment human operators. “The intent is to augment operators and for the AI to start building the course of action,” Hendricks explains. For emergency operations center personnel, who often monitor multiple screens, C2AI can use color-coded alerts (red, orange, yellow) to draw attention to critical events.
Dynamic Response Planning: In a simulated scenario, visual alerts agents detect a building collapse. An EMS planning agent, trained to identify cues like smoke and structural damage, displays initial assessments. When embers appear, the agent rapidly adjusts the recommended response plan, dispatching a fire truck and requesting human affirmation or revision. An experienced incident commander can then refine the plan, for example, by ordering additional fire engines and a ladder truck.
Integration with Existing Tools: C2AI is being integrated with the Team Awareness Kit (TAK), a communications and collaboration application co-developed by the Air Force Research Laboratory and Leidos, currently used by 70,000 DOD, DHS, state, and local users. This ensures seamless flow of descriptive and visual map updates, 911 transcripts, and approved action plans to all first responders.
Technological Foundation and Deployment: Leidos built these agents using NVIDIA’s foundational speech AI, large language models (LLMs), and vision-language models, enabling natural communication between agents and humans. The agents are modular and scalable, relying on real-time retrieval of specified documentation for decision-making, allowing quick adaptation to changing procedures without extensive retraining.
C2AI is developed as lightweight microservices, eliminating the need for large computing infrastructures. This allows it to run on servers at the “tactical edge,” processing data with minimal delay. NVIDIA’s established AI models are packaged in containers with necessary computing resources for streamlined deployment, enabling agencies to adopt C2AI incrementally.
Also Read:
- Ascendo AI Agents Leverage Google Gemini A2A Protocol to Drastically Reduce Enterprise Escalations
- NVIDIA Unveils DoMINO NIM: Accelerating AI-Powered Simulation for Breakthroughs in Computational Engineering
The partnership between Leidos and NVIDIA, established earlier this year, aims to accelerate the operationalization of AI capabilities for government agencies, combining Leidos’ mission expertise with NVIDIA’s AI technologies. “With C2AI, we are focused on applying agentic AI to solve complex problems and augment decision-making in high-consequence situations,” Hendricks concludes. “In disaster scenarios, let’s reduce the time to critical care.”
