AI-Native Safety Inspections
for the U.S. Army
A greenfield AI-native platform vision — designed from scratch in 10 days — that opened a real enterprise door at the Pentagon.
US Army's existing safety platform.
RMIS — Risk Management Information System, U.S. Army Combat Readiness/Safety Center
The U.S. Army reached out to ServiceNow for a modernisation vision — what AI-native safety management could look like on a commercial platform. Their existing system, RMIS, is a reactive accident database. Data goes in after harm occurs. No pattern detection, no prevention loop. With 40,000 safety officer roles open, the timing to propose something better couldn't have been clearer.
Research chose the focus. We didn't wait to be told.
The Army brief was open-ended: show us what AI-native safety management could look like on ServiceNow. Rather than defaulting to a broad overview, the team ran rapid research synthesis to identify the single highest-impact scenario to design for. Safety Inspections emerged as the clearest opportunity — the most frequent safety officer interaction, and the one most underserved by RMIS.
We started with a textual narrative first — writing the story before designing a single screen, then iterated the narrative until the logic was airtight.
This direction explored a soldier reporting an incident in the immediate aftermath — voice input, injury classification, scene documentation. It was technically compelling but fundamentally flawed. A soldier in distress doesn't open an app. They seek medical attention. Designing for that moment meant designing for a scenario where no one would actually use the product.
Proactive inspection puts the safety officer in control before an incident happens. AI earns its place here — not as a form-filler, but as a pattern-recognition layer that surfaces what a human cannot see across hundreds of incidents and locations. This is where intelligence changes outcomes.
The human insight that killed Narrative 1: a soldier in distress seeks medical attention first. Any design that depends on that person opening an app is a design that will never be used when it matters most. The right AI story is prevention, not documentation after the fact.
10 days. Two sprints. One story.
My contribution was strategic and present throughout — vision framing, narrative direction, team configuration, and hands-on design reviews two to three times a day. The direction I gave the team at the start shaped everything that followed.
"Forget ServiceNow. If you could bring anything in from the world today — any AI, any interface, any interaction model — what would you build?"
Direction given to the design team, Week 1 kickoff
5 designers + design manager · Narrative building, research synthesis, direction lock
Questions & Research Synthesis
We framed all the questions, met with cross-functional partners to get the answers.
Textual Flow
We created a textual flow that made sense for the demo before jumping into lofi concepts.
Lofi Scene Prototypes
We broke the narrative into scenes and designed lofi prototypes for each — validating the story before committing to high-fidelity execution.
By the end of week one, the team had a locked narrative direction — proactive inspection, not reactive reporting. One clear story. No parallel tracks held open. That clarity is what made week two possible.
2 focused designers · Mobile flow + web flow · 2–3 design reviews daily
Command View
AI-native home screen surfacing the officer's tasks, priorities, and inspections.
Conversational Inspection
Voice-first inspection — AI structures the report in real time as the officer speaks.
AI Camera
Point and detect. AI identifies hazards on sight and surfaces corrective actions instantly.
Dynamic Inspection
AI adjusts the task flow by context. Moving away from rigid form filling — entirely.
In ten days, the team went from a blank canvas to a Pentagon-ready demo — built on an entirely new component system, shaped by the Army's own design language, and told as a single coherent story from field hazard to pattern prevention.
This project is under NDA. I'm happy to walk through the full designs in person or on a Zoom call.
Design work that opened real doors.
A 10-day design sprint became the proof of concept that started a real enterprise conversation — and travelled further than most production work does.
Presented to senior Army officers at the Pentagon. The AI-native, mobile-first vision was well received.
The demo was presented to the entire ServiceNow vertical — used as a flagship example of AI-native platform thinking.
The Army's technical team initiated a follow-up with ServiceNow to discuss a migration plan and build roadmap.