Our work is a redesign of a clinical trial AI platform Altrovia built for data scientists, principal investigators, and regulatory teams. The core design challenge was trust: how do you make AI-generated outputs feel credible and actionable in a high-stakes medical context?
Design Process: We started with competitor analysis, studying Palantir, Medidata, Veeva, and Linear to understand what visual systems communicate authority without overwhelming users. From that research, we defined three principles: surface AI outputs visibly, give users control at every step, and use structure over decoration. From there we built wireframes directly grounded in those patterns, then moved into high-fidelity screens covering the full trial workflow, from uploading a synopsis to generating a final protocol.
Key Design Decisions: We moved away from rounded corners and soft surfaces toward sharper borders and denser information hierarchy, directly informed by feedback from a Palantir product designer. We introduced collapsible section rows, a muted label vs. dark value pattern for scannable parameters, and AI extraction badges to make the source of every data point transparent.
The dashboard was redesigned as a portfolio view with visual trial status, amendment risk grids, and phase breakdown charts, responding to stakeholder feedback that the product needed to feel like an analyst tool, not a marketing site.
How We Used Adobe Express: We used Adobe Express to structure our competitive analysis findings into a visual pros and cons chart. That chart became a direct reference during wireframing, surfacing the clearest pattern across all the products we studied: the most trusted tools show their reasoning. Every layout decision we made in Figma traces back to that one insight from the Express artifact.
Built With
- adobeexpress
- claude
- cursor
- figma
Log in or sign up for Devpost to join the conversation.