critter safARi
Augmenting every zoo visit to spark eco-empathy.
critter safARi integrates an LLM agent to streamline route planning and leverages GPS-based AR to deepen learning. I catalyzed and led this student project during my Master's at U of T and later evolved it via a personal design sprint — which also served as an opportunity to explore and evaluate emergent AI tools. The output is a pitch-ready concept grounded in UX research.
Role
Product Designer, Unity Developer, UX Researcher
Duration
2 months
Project
Team Project to Personal Project
Tools
Unity, Xcode, GitHub, PS, Blender, Figma, Kling AI, Midjourney, ChatGPT
CYCLE ONE
Team Project · 6 Weeks
Project Leadership
Unity Prototyping
A/B Test
PROJECT CATALYSIS
I had noticed low critter visibility at zoos and pitched AR as a solution to my small M.Sc. class — attracting two teammates.
Problem Definition
User: Zoo visitors — primarily families with children.
Frustration: Animals are far away or out of sight.
HMW: Balance animal care with the visitor experience.
Solution: AR brings critters up close to the user.
AR also conveys scale via parallax and can host other educational snippets.
North Star: A mobile AR app for our local zoo.
Primary Persona
Timmy
12 y/o · Sloth Enthusiast ·
Older Brother
"Both of our [his and his little sister's] favourite animals were missing! We had waited forever for our visit."
A/B TESTING
Testing two Unity prototypes I built, all 5 testers preferred interaction pattern B for more proximate interaction feedback and lower cognitive load.
Option A
Tapping circle button opens bottom sheet and scrolls to related text.
Option B
Tapping circle button opens pop-up.
TEAM ACHIEVEMENT
Together, we delivered a functional prototype where users scan AR markers to reveal critters.
The project earned a high grade, received support from academic faculty and early testers, and was showcased to junior cohorts!
Credits: J. Leggatt (UX/UI), B. Lazar (3D), R. Gao (Lead, Unity)
CYCLE TWO
Personal Project · 2 Weeks
LLM Feature
GenAI Video & Images
Design & Business Strategy
User Research
Branding
MORE UX RESEARCH
Because of our team's success, I evolved the project by first deepening research (5 interviews and competitive analysis) — identifying opportunities in wayfinding, education, and marketing.
Secondary Persona
Madeline
38 y/o · Mother of 3 · Busy Professional
"Planning our path, navigating, and meeting every child’s needs is no small feat."
Tertiary Persona
Mateo
66 y/o · Retired Small Business Owner
"My knees wish we took the shortest route instead." "Not enough conservation stories."
Quaternary Persona
Celine
21 y/o · Student · Blogger
"My friends' social media post of the adorable baby fox led me here." "More cool events for people my age?"
MORE HMW'S
Simplify route planning
Deepen learning
Set up a viable business plan
MORE FEATURES
To meet user needs, I introduced an LLM agent, a new AR strategy, and a partnership model.
Integrating an LLM-based agent simplifies route planning for users like Madeline and Mateo. Given development limits, preset routes can serve as a fall-back.
Changing from our initial marker-based AR to GPS-based AR prevents overcrowding near physical AR markers — allowing for more simultaneous users.
What if this was a consolidated app that partners with multiple zoos? This could increase user retention and engagement for all partners.
USER-DRIVEN ITERATION
Testing this wireflow with 4 users via think-aloud led to improvements in engagement, social media reach, and task speeds.
Boosting Zoo Engagement
Making AR critters seasonal collectibles: encourages engagement by gamification.
Marketing via Social Media
Adding selfie option: improves social media reach, targeting visitors like Celine.
Reducing Effort
Reducing interactions to select route: speeds up task for users.
Before
After
VISUAL IDENTITY
While I crafted the branding by hand, Midjourney helped me explore color schemes.
GenAI DISCLAIMER
Given this opportunity to explore and evaluate GenAI, AI-generated visual assets (Kling, ChatGPT) were integrated into this concept app. Below are examples of raw AI outputs versus my final edits.
The UX/UI of this project was otherwise created by hand.
FINAL VISUALS
The final concept brings critters up close, streamlines wayfinding, provides bite-size education, and boosts zoo engagement — to ultimately nurture ecological empathy.
LLM-Based
Route Planning
Map (Bridges
AR and Navigation)
Education via
Seasonal AR Critters
NEXT STEPS
I'd take this concept further by connecting with various zoos to identify their unique needs and constraints — crystallizing the product roadmap.
REFLECTIONS
Exploring GenAI as a Creative Partner
This project was a fun opportunity for me to test cutting-edge image/video generation tools like Kling AI, Midjourney, and ChatGPT as well as evaluate their alternatives.
However, for launch-ready projects, I'd create final visual assets using Blender and Adobe Creative Suite instead to ensure full creative control.
Next, I'd like to learn AI prototyping to enhance wireframing.
Valuing Clarity, Honesty, and Consideration
Clear, empathetic communication and documentation was invaluable for smooth cross-functional collaboration.
Sharpening Core Design Process
The rapid design sprint allowed me to reflect on my core design process and practice agile workflow principles.