Concept project · Healthcare · Diagnostics
LabSync AI — One diagnostic platform, three roles
An AI-assisted diagnostic laboratory platform connecting patients, lab technicians and operations in a single role-based experience
- Role ecosystems
- 3
- Screens designed
- 20+
- AI experiences
- 4
- WCAG 2.2 target
- AA
Overview
Patients faced confusing preparation instructions and invisible appointment status. Technicians coordinated by phone with no real-time sample visibility. Operations tracked inventory in spreadsheets and allocated technicians by hand.
LabSync AI is one platform with three purpose-built role experiences on a unified backend — patient mobile app, technician app with analytics, and a responsive operations dashboard.
Honest framing
A conceptual case study, stated plainly
Research used secondary sources, healthcare workflow analysis, industry-informed assumptions and structured AI persona agents. No real user interviews were conducted, and no metrics are claimed.
Findings were treated as directional, then evaluated through UX reasoning, healthcare workflow knowledge, accessibility requirements and product constraints.
Discovery
AI-assisted research with clear boundaries
Research questions fed structured AI persona agents for patient, technician and operations roles; themes were extracted, reviewed by hand and turned into UX insights.
AI explored the hypothesis space, generated structured responses, surfaced recurring themes and challenged assumptions. Interpretation, prioritisation, validation against constraints and every final design decision stayed human.
Strategy
Every feature traces to an insight
Patients forget appointments → smart reminders. Preparation instructions confuse → an interactive step-by-step guide. Support fields repetitive questions → an AI health assistant. Progress is invisible → real-time sample tracking. Technicians coordinate manually → a technician dashboard. Admins lack visibility → an analytics dashboard.
- Guide before users need help
- Make progress visible at every step
- Reduce avoidable manual work
- AI assists, humans remain in control
- Accessibility is built in from the start
Experience architecture
Three role ecosystems, one design system
The patient journey runs discover → book → prepare → attend → collect → process → report, with each role seeing only what matters to it: patients get booking, preparation and reports; technicians get today's collections, sample workflow and status; admins get appointments, teams, inventory and KPIs.
Designing AI for healthcare
AI that assists without overstepping
Four AI experiences — preparation assistant, patient Q&A, report explanation and operational insights — each with explicit guardrails.
AI can explain general preparation, answer routine process questions, translate terminology into plain language and surface operational patterns for review. It cannot diagnose, replace clinicians, recommend treatment, change results or override lab protocols.
Accessibility
Status is never colour alone
4.5:1 contrast for text and 3:1 for UI, visible focus states, logical focus order, 44×44px touch targets, semantic headings, ARIA labels, aria-live status updates, reduced-motion support, font scaling, high-contrast and colour-blind modes.
Every status pairs an icon, a text label and colour, so screen readers receive the full meaning.
Reflection
Clarity and safety before visual novelty
AI is most valuable when paired with structured product thinking — without clear problem framing, its outputs are noise. In healthcare, a well-labelled status indicator matters more than a beautiful animation, and human control cannot be applied retroactively.
Next case study
Enterprise Assessment Platform — Redesigning legacy search
Enterprise Healthcare SaaSContact
Let’s talk about the problem, not just the solution.
I’m open to senior product design, UX leadership, design systems, and AI-native product opportunities.
Get in touch