DESK: DESIGN + FRONTEND
HANDOFF: NOT REQUIRED
STATUS: OPEN TO OFFERS + FREELANCE
THURSDAY, 15 OCTOBER 2026
EDITION: WEB
UX CASE STUDY · MEDECRO.AI · 2026
BEHANCE CASE STUDYHow redesigning a single patient-facing card moved the AI X-ray Analyser feature from 20% to 43% adoption — and convinced a room of sceptical VCs with one slide.
Adoption lift
20% → 43%
Drop-off eliminated
78%
Sprint duration
3 weeks
Doctors interviewed
11
02 — THE CONTEXT
THE CHALLENGE
Medecro's AI X-ray Analyser was clinically validated, CDSCO-compliant, and completely invisible to its users. Session recordings showed a painful loop: doctor opens patient card → realises X-rays need AI screening → navigates away to the X-ray tab → loses their place → gives up.
MY ROLE
Embedded as the sole designer + design engineer for a 3-week sprint. My deliverables spanned the full design-to-code pipeline: user interviews, information architecture, Figma prototypes, and production-ready Next.js components using shadcn/ui.
03 — USER RESEARCH
Eleven contextual interviews across 4 clinic archetypes in tier-2 Indian cities. Moderated remotely via Zoom with screen share, supplemented with Hotjar session recordings. Four archetypal personas emerged.
Dr. Navneet
General Dentist
"I know the AI can help me catch things faster, but I just don't have time to dig for it."
Dr. Parag
Oral Surgeon
"The feature exists. I just don't trust it enough yet to build a workflow around it."
Rima
Front Desk Staff
"Half the time I don't even know the patient had an X-ray until the doctor asks me."
Dr. Aisha
Practice Manager
"We paid for AI. I need to see our team actually using it before I report to the board."
04 — THE IMPACT
The redesign didn't add features — it removed distance. The AI X-ray panel moved from a buried tab to a contextual card action.
BEFORE — Original Dashboard
AI feature buried in a separate tab · 78% drop-off
AFTER — Redesigned Dashboard
AI surfaced inline on the patient card · 43% adoption
05 — RETROSPECTIVE
Three weeks, eleven interviews, one pivotal metric. Five takeaways I'd carry into every healthcare UX engagement from here.
01 — THE HARD PART
78% drop-off → design sprint approved
The founding team pushed back hard on a full redesign. The data changed the conversation. When I mapped the 78% drop-off against monthly active users, a single slide shifted the room.
VCs respond to metrics, not mockups. The 78% abandonment rate on the AI X-ray flow became the north-star argument that unlocked the design sprint — and taught me to lead every future brief with a business-case number before opening Figma.
02 — CORE METRIC
Feature adoption of the AI X-ray Analyser — the metric that defined sprint success.
BEFORE
18%
AFTER
43%
03 — ROOT CAUSE
Dentists operate in a 7-minute consultation window. Any feature requiring a context switch — even a tab change — is effectively invisible. The fix wasn't better onboarding. It was bringing the feature to where the doctor already was.
KEY INSIGHT
“Proximity beats discoverability. Put the tool where the moment of need is.”
04 — DESIGN SYSTEM
Every component shipped with a shadcn/ui primitive underneath — Sheet, Card, Badge, Tooltip. Zero custom component debt. The AI feature panel is a single Sheet component with a custom trigger surfaced on the patient card.
05 — ENGINEERING
Every Figma frame maps 1:1 to a component in the Next.js codebase. The design file links directly to the Tailwind token. Developers shipped the new AI panel in two days — no spec translation needed.