Project Salus is an AI tool that ingests and displays critical supply chain data to predict COVID-19 hotspots and related logistics issues before they happen, owned by the Joint Artificial Intelligence Center (JAIC), United States Department of Defense.
Our role as the design partner was to formulate and teach a reproducible design methodology for building world-class technology products inside the Department of Defense, create a flexible, intuitive design system, and assist in creating usable, functional dashboards/interfaces for military leaders and analysts alike.
What is detailed here in this case study is a publicly-available representation of my role in the project. All unclassified, for official use only (FOUO), or otherwise proprietary information, has been withheld.
“Project Salus is operating as a dashboard platform that is fed into the Common Operational Picture (COP) interphases for both NORTHCOM and the National Guard.”
— Lt. Cdr. Arlo Abrahamson, JAIC
Over the course of our engagement, we interviewed personnel from various combatant commands. Specifically, users from the Air Force, USNORTHCOM, INDOPACOM, and National Guard were interviewed over 9 months culminating in over 150+ hours of research sessions. These working sessions were instrumental in ascertaining the situations, environments, and capabilities that needed to be addressed using the Project Salus platform.
Demo at the Eisenhower Executive Office Building, The White House
In addition to end users, we also conducted user research and design feedback sessions with Project Salus technology partners including, but not limited to, IBM, ECS Federal, Lucd.ai, and Microsoft. Most interviews were conducted via recorded video conference, given the COVID-19 situation in the United States at the time. We produced two comprehensive research reports detailing our findings, including recommendations for the project's next steps.
Working closely with members of the JAIC and USNORTHCOM, we designed in weekly sprints for quick feedback loops, sometimes meeting multiple times a week in order to nail an interaction or gain deeper understanding of the problem at hand.
The design process was unlike any I’ve ever been a part of, given the system complexity and interdependent nature of the various end-users we needed to account for. Designing and building a platform that can be interpreted by both deeply technical analysts and senior military officials was one of the most rewarding design challenges of my career thus far, and I’m extremely proud of how our team rose to the challenge at the height of COVID-19 making its way stateside.
We designed an end-to-end workflow for analysts, military officials, and private sector technologists to collaborate and share data intelligence in realtime. In addition to this work, an extensible design system was created and built in Figma for Project Salus and for the JAIC to use across all of its future products.
Once we established design patterns and a shared language, it became clear that Salus could easily be plugged into other critical DoD initiatives, such as Joint All-Domain Command and Control (JADC2). Salus had the flexibility to be used in disaster response management, scenario planning, and even predictive wargaming.
To demonstrate these enhanced capabilities, we built a prototype with real and historical data to demonstrate how the system fluidly responds in a situation where a hypothetically adverse weather event affects virus spread, and showed the chain reaction that might necessitate mass evacuations.
This prototype was demoed to The White House, FDA, NORTHCOM, INDOPACOM, 1-Star and 2-Star Generals across the military services.
Due to the sensitive nature of this project, all design work will only be shown on a case-by-case basis.
As always, it takes a village and none of this work would have been possible without the huge contributions from the following talented people that I had the pleasure of working with (in no particular order):
Matthew Miller, Shaun Modi, Katherine Tobin, Andrew Vilksac, Dave Augustine, Kirsten Berken, Lauren Rhode.