Piction Health
Piction Health is a clinical decision-support tool designed to help primary care physicians identify and triage dermatological conditions using AI-powered visual search and structured questionnaires. The product needed to balance clinical accuracy, speed in high-pressure environments, low cognitive load for physicians, and trust in AI-assisted recommendations. The challenge was not visual polish — it was workflow clarity under time constraints.
Fast, high-stakes decisions without specialist training
Primary care physicians often lack specialized dermatology training but must still make fast, high-stakes decisions. The existing flow was fragmented and cognitively heavy, with unclear edge cases. The key design challenge was how to structure a decision-support workflow that feels fast, intuitive, and clinically reliable even in complex scenarios.
UX/UI Design
Structuring the UX workflow. Defining interaction logic for the body model selection system. Designing edge-case behavior (multi-area selection, confirmation states). Translating clinical logic into intuitive UI patterns. Delivering high-fidelity, dev-ready design assets. Worked closely with product and engineering to ensure implementation accuracy and clarity.
Clarity under time pressure
Instead of overwhelming the user with visual and form inputs simultaneously, the flow was separated into a structured symptom questionnaire followed by interactive body model selection, reducing cognitive switching and improving decision speed.
The body selection model supports multi-area selection, clear visual feedback, confirmation states, and error prevention — optimized with tap-only actions, no zoom/rotate needed, usable one-handed.
The results interface presents the three most relevant matches based on questionnaire inputs, selected body areas, and visual analysis, with a comparison mode for visual verification before a diagnostic decision.
Reduced friction, increased confidence
The redesigned flow reduced friction in clinical workflows, improved clarity in edge-case scenarios, increased confidence in AI-assisted recommendations, and delivered a scalable interface aligned with engineering constraints.
