NextHealth
Modern Wellness Center Platform – Web & Mobile
A two-sided wellness platform where AI runs the routine and clinicians keep control of the call.
Client
Competitors
My Role
Lead Product Designer – UX, UI, design system
Industry
Healthcare · Wellness · Longevity
Deliverables
Dual-stream product architecture, AI recommendation UX, design system
Next Health is a next-generation clinic network building the “clinic of the future” – a comprehensive approach to wellness, preventive medicine, and longevity. The product needed to unify patients and medical staff into a single AI-powered ecosystem: appointment management, lab results, automated health recommendations, online consultations, and continuous monitoring – all in one interface.
The core design challenge was building a system where AI handles the routine – surfacing insights, flagging anomalies, personalising recommendations – while patients feel in control of their data and clinicians get fast, focused tools without the complexity typical of medical software.
A platform that unifies patients and medical staff in one ecosystem – appointments, lab results, AI-powered health recommendations, and automated monitoring. Designed to make complex healthcare feel effortless.
Keywords:
– Agentic UX
– AI workflow design
– Healthcare SaaS
– Multi-stream product
– Design system
– Human-in-the-loop
The challenge
Two user groups, one ecosystem. Patients wanted a calm, premium consumer experience that explained their health like a fitness app explains progress. Clinicians wanted clinical density: fast lookups, audit-friendly history, and zero ambiguity about what the AI was suggesting versus what they were prescribing. The AI layer had to make the handover between autonomous and human action invisible to patients but explicit to staff.
My Role
I led end-to-end UX and UI across both surfaces – research, IA, interaction design, design system, and usability testing – with a particular focus on where the AI agent operates autonomously, where the clinician must intervene, and how to express that boundary in the interface.
– Researched real patient and staff scenarios and mapped where AI could operate autonomously vs. where human review was required by clinical safety.
– Designed a dual-stream architecture: Client App and Nurse Dashboard – two distinct experiences sharing one design system and one data model.
– Defined the agentic UX patterns for AI-generated recommendations – confidence indicators, override controls, and audit trail visibility for clinicians.
– Architected the information structure and navigation to minimise cognitive load for both lay and expert users.
– Built a scalable, modular design system with strict medical typography and explicit state-signalling colour use.
– Reduced error rates in appointment booking and lab interpretation through three rounds of usability testing.
Impact
“In healthcare, interfaces shouldn’t impress — they should help people access critical information quickly, in plain language. AI is only useful here if the human can see exactly what it did and why.”
Key Screens
Client Dashboard
- Upcoming appointments
- Recent lab results
- Personalised recommendations
- Health trend charts
Goal: show health status as clearly as a fitness app shows progress.
Nurse Management Panel
- Quick patient record access
- Visit history by date
- Lab results timeline
- Recommendation updates
Goal: reduce routine actions and accelerate the care cycle.
Process
Discovery & research
I conducted in-depth interviews with both clinic clients and nursing staff – two groups with fundamentally different mental models. Patients consistently asked for “everything in one place” and wanted the dynamics of their health over time, not a static list of values. Nurses reported data duplication, absence of quick-action scenarios, and difficulty keeping recommendations and recent visits in focus simultaneously.
Defining where AI operates
Before any UI work, I framed the agentic boundary: which actions the system performs autonomously (anomaly flagging, recommendation drafting, appointment reminders), which require human approval (any change to a treatment plan), and which are pure information surfacing. That map became the spine of every screen.
Interface design
For patients, I introduced a “health dynamics” module – charts, key indicators, and changes over time – making health status as intuitive to read as a workout-progress screen. Smart notifications were added for new results, upcoming appointments, and consultation invitations. For clinicians, I implemented quick-action flows: marking visits, adding recommendations, and uploading results – all achievable in minimal steps without navigating deep into the system.
Design system
A modular system with a strict medical typographic scale, clear state-signalling colour use, and a unified component library spanning mobile and web. The aesthetic is calm and clean -appropriate for a premium-tier clinic without feeling cold or clinical.
Usability testing & iteration
Three critical scenarios: appointment booking, reviewing lab results, and following recommendations. The booking flow was reworked based on findings – the number of steps was reduced by nearly half. Labels and section names were revised so health information remained accessible without requiring medical literacy.
Outcome & Reflection
The result is an ecosystem that makes the medical experience simple, trustworthy, and technologically coherent. Patients can manage their wellness journey and receive recommendations that genuinely guide their decisions. Clinicians spend less time on routine tasks and more time on patient care. The clinic gained a scalable platform ready for new modules.
The result is an ecosystem that makes the medical experience simple, trustworthy, and technologically coherent. Patients can manage their wellness journey and receive recommendations that genuinely guide their decisions. Clinicians spend less time on routine tasks and more time on patient care. The clinic gained a scalable platform ready for new modules.
