CAPTRS Announces the Winners of 2025 AI Challenge

CAPTRS is pleased to announce the winners of the 2025 CAPTRS Challenge: AI-powered Human Augmentation for Emergency & Crisis Games and Exercises. Developed in partnership with the University of L’Aquila and the Institute for Experimental AI at Northeastern University, this international competition called on innovators around the world to use AI to reimagine how organizations design and run emergency preparedness exercises.

The challenge was born out of a simple but urgent problem: designing high-quality tabletop exercises and crisis simulations takes enormous time and specialized expertise resources that many smaller organizations simply don’t have. Smaller hospitals, municipalities, and regional public health agencies often need to run exercises but lack dedicated in-house teams to build them. CAPTRS wanted to know: could AI change that?

The answer, based on the submissions we received, is a resounding yes.

After a rigorous evaluation process, CAPTRS is proud to recognize two award-winning teams.

First Place: ArcManifold
Eric Ezenwanne - United States

Eric Ezenwanne’s ArcManifold takes one of the most resource-intensive parts of emergency preparedness, the tabletop exercise  and transforms it into a continuous, AI-driven process. Rather than relying on static, pre-scripted drills, ArcManifold uses a neurosymbolic Multi-Agent System to dynamically simulate crisis scenarios in real time, adapting to decisions as they unfold and stress-testing organizational resilience at every step.

Planners begin in a Foundation Wizard that guides them through defining their organization type, hazard profile, and operational frameworks such as HSEEP and ISO 22398. From there, an AI Architect Assistant helps orchestrate participant networks and map complex inter-agency relationships. A TTX Exchange feature enables community-driven template adaptation, allowing organizations to build on existing exercise designs rather than starting from scratch. A multi-phase adversarial AI then stress-tests the exercise design under pressure. When it’s over, the system automatically produces SITMANs, After Action Reviews, and Master Scenario Events Lists along with granular telemetry on decision-making under stress  removing hours of post-exercise documentation work.

Second Place: SmartEx - Emergency Exercise Assistant Platform
Roberto Pizzi - Italy

Roberto Pizzi built SmartEx with a clear principle at its core: AI should augment the facilitator, not replace them. SmartEx is a fully functional prototype of an AI-assisted platform that walks facilitators through every phase of exercise design and evaluation from the first structural decisions through content development, inject sequencing, and material generation.

What makes SmartEx stand out is how it handles the full lifecycle. Most tools help with one part of the process. SmartEx covers all of it, using agentic AI logic and a large language model backbone that works alongside the user, offering guidance and generating outputs while keeping human judgment in the driver’s seat throughout.

“A consistent theme across the submissions was the use of AI to structure reasoning, surface trade-offs, and support decision-making under uncertainty. That is a meaningful step forward. For CAPTRS, it validates our focus on integrating AI into simulation design and execution in a way that strengthens, not abstracts, human expertise.”
- Georgia Thomsen, Executive Vice President and Chief of Staff, CAPTRS.

CAPTRS extends its congratulations to both award-winners and its gratitude to every innovator who submitted a solution this year. The depth of creativity and technical rigor on display reflects a growing global community dedicated to making crisis preparedness more accessible, scalable, and effective and that is exactly the future CAPTRS is working to build.

Visit 2025 CAPTRS Challenge results to learn more.