Luniby
Predictive petcare ecosystem connecting owners, nurses, and clinics through empathy-driven design.
- UX Research
- Multi-User Design
- AI Health
- Design & Build
- Role
- AI Design & Delivery — UX researcher, UX/UI designer, and developer
- Timeline
- 2025
- Tools
- Figma, Miro, User Interviews, Prototyping
- Where
- Auckland, New Zealand
- Chapters
- 12 sections
What this project is
Luniby is an AI-powered petcare platform designed to reduce clinic burnout, streamline triage, and improve pet wellbeing. The platform connects pet owners, veterinary nurses, and clinics in one intelligent ecosystem. I owned it end to end — the research, the interface, and the front end that shipped — so there was no handover between deciding something and building it.
- Role
- AI Design & Delivery — UX researcher, UX/UI designer, and developer
- Timeline
- 2025
- Tools
- Figma, Miro, User Interviews, Prototyping
The challenge
Clinic burnout, fragmented triage, and anxious pet owners create systemic pain points across the petcare ecosystem.
Where it hurts
- Pet owners experience high anxiety (8.2/10 avg) during symptom discovery with no immediate reassurance
- Veterinary clinics spend 40% of consultation time on triage that could be pre-screened
- Multiple disconnected touchpoints create poor continuity of care
- No unified system connecting early detection, triage, and follow-ups
Design goals
- Design clear, guided triage experience for anxious pet owners
- Create actionable Health Notes system for clinic staff to reduce triage burden
- Orchestrate multi-user experiences across owners, nurses, and clinics
- Build trust through transparent AI reasoning and empathetic design
- Establish continuity of care from triage to consultation to follow-ups
How I found out
Conducted comprehensive research across multiple stakeholder groups to understand systemic pain points:
18 pet owners (ages 24–55), 8 veterinary nurses, 6 clinic administrators in Auckland
12 hours shadowing Auckland veterinary clinic, observing triage workflows and staff pain points
Pet owners across NZ and AU on triage behavior, anxiety levels, and AI tool acceptance (82% interested, 67% would pay)
Service blueprinting across anxious owners, time-stretched nurses, and clinic operations
What we learned
“My dog started limping and I panicked — is it an emergency or can I wait? I called three clinics and they were all fully booked. I just needed someone to tell me if this was urgent.”
Key insights
- Multiple user types require orchestrated experiencesAnxious owners need clarity and reassurance, clinics need actionable data
- High anxiety during symptom discoveryPet owners experience 8.2/10 average anxiety, creating demand for immediate guidance
- 40% of clinic time spent on pre-screenable triageTime-motion study showed significant efficiency opportunity
- Early detection builds trustPilot testing showed 28% reduction in non-urgent bookings
- Empathy-driven UX reduces anxietyA/B testing showed empathetic language reduced anxiety by 34%
Journeys across three user types
Mapped complete user journeys across three user types: pet owners, veterinary nurses, and clinic administrators. Flow shows: Symptom entry → AI triage assessment → Health Notes generation → Clinic integration → Consultation booking → Follow-up care
Symptom Entry
Describes symptoms via chat
AI Triage
Receives urgency assessment
View Health Notes
Reviews AI summary
Book Appointment
Schedules with clinic
Follow-Up Care
Receives aftercare guide
Receive Notes
Pre-screened patient data
Priority Review
Confirms urgency level
Manage Queue
Schedules appointments
Post-Consult
Updates patient record
View Dashboard
Monitors clinic activity
Manage Resources
Allocates staff & time
Review Analytics
Reports & billing data
Structuring the system
One system, three portals. Each role sees only what it needs to act on, while the underlying record stays shared.
Owner Portal
- Pet profiles
- Symptom assessment
- Triage results
- Health records
- Consultation booking
Nurse Portal
- Incoming triage queue
- Health Notes review
- Priority scoring
- Clinic scheduling
Clinic Admin Portal
- Dashboard analytics
- Staff management
- Integration settings
- Billing
Low to mid fidelity
Low-fidelity exploration
Started with rough sketches to explore layout options and interaction patterns for the triage chat interface.
Symptom Chat
Triage Result
Mid-fidelity refinement
Refined information hierarchy, tested with 15+ users to validate flow and comprehension.
Pet Onboarding
Health Notes
Key design choices
Each decision below is paired with what it was tested against — rationale, not preference.
Empathy-Driven Language
Used supportive, reassuring language throughout the triage flow. A/B testing showed empathetic language reduced user anxiety by 34%.
Clear Visual Symptom Assessment
Combined visual + text reduced anxiety 34% vs text-only. Users could tap on pet illustrations to indicate symptom locations.
Transparent AI Confidence Levels
Showed confidence scores and reasoning to build trust. Prototype testing showed transparency increased trust from 6.2 to 8.7.
Mobile-First One-Handed Design
Optimized for high-stress scenarios where owners might be holding their pet. All primary actions within thumb reach.
Health Notes for Clinical Staff
Co-designed with 8 veterinary nurses through 6 workshops. Structured format showing symptoms, timeline, urgency level, and AI reasoning.
High-fidelity designs
The triage interface as shipped to beta. Select any screen to view it full size.
Inclusive design decisions
A triage tool gets used in the worst moment of someone's week, often one-handed. Accessibility here is not compliance work — it is the product working at all.
- WCAG 2.1 AA CompliantAll text meets contrast requirements, tested with screen readers
- Large Touch TargetsMinimum 44×44px for all interactive elements (tested on 5 device sizes)
- Clear Error StatesHelpful error messages with recovery actions
- Offline CapabilityService worker caching ensures access during emergencies
- Keyboard NavigationFull keyboard support for all interactions
- Simple LanguageFlesch-Kincaid Grade Level 8 for all user-facing text
- Progressive DisclosureComplex information revealed gradually to avoid overwhelming users
Impact & reflection
Platform in beta at luniby.com with 220+ registered pet owners
78% triage completion rate, 8.9/10 owner satisfaction
Conducted demos at 3 veterinary clinics (Boston, Austin, Denver area)
Key validation: Clinics requested triage as embeddable API for existing systems
Key learnings
- Multi-user design requires orchestrationEach user type needs distinct experiences while sharing data seamlessly
- Empathy reduces anxietySupportive language and clear guidance measured 34% anxiety reduction
- Co-design with stakeholders is essential6 workshops with nurses produced far better Health Notes design than solo iteration
- B2B validation shifted strategyClinics preferred embedded API over standalone platform — pivoted roadmap accordingly
- Transparency builds trust in AIShowing confidence levels increased user trust scores from 6.2 to 8.7
