Designing trust into an AI agent for requirements management
Meet Zelda, Stell's reliable AI agent.
I designed Zelda’s presence throughout Stell, an AI-native platform for aerospace systems engineering. Every interaction is precise, legible, and designed to earn trust when mistakes aren’t an option.
Requirements work is complex and manual. Stell changes that.
Engineers work across dense specifications, interconnected requirements, and traceability data. Stell brings this information together in one place, transforming how engineers navigate and understand complex systems.
Three views from the platform, from import to traceability.
And now, there's Zelda.
Zelda helps engineers navigate complex requirements and traceability work by surfacing relevant information and suggesting next steps. The engineer reviews each suggestion, decides what to act on, and confirms before anything changes. AI does the legwork; the engineer remains in charge.
Walk through a Zelda session
Click through the slideshow to see how Zelda rapidly surfaces relevant evidence and context, enabling engineers to make informed decisions while maintaining control over what gets linked.
Iconography
Stell links everything in a systems engineering project, from requirements to evidence records, like a constellation. That made the star the natural anchor for the product's icons. Stell wanted it to appear throughout, with each icon holding one meaning users would remember. I built on their original icons and designed the rest from scratch as the product grew. Every icon pairs a familiar object with the star, so once you learn one, you can read the rest.
.png)
.png)
.png)
.png)
.png)