MonkiGO
From concept to live prototype in 60 hours — using AI as a real design and development tool, not a shortcut.
By digitising a hand-mapped dataset of 88 artworks and translating UX screens directly into a working web experience, I transformed a one-time physical event into something anyone can discover. This project demonstrates how AI tools can extend a designer’s reach — from concept all the way to deployment.
Context
Self-Initiated Systems Experiment
My Role
UX + Product Designer
AI Tools Used
Cursor · Claude · Figma
Timeline
60 hours
Overview
MonkiGo started life as a physical experience — a self-designed monkey hunt across Upper Thomson MRT station, built on a handcrafted dataset of 88 hidden monkey illustrations and a carefully balanced scoring system.
After running the live event, the question was obvious: what if this lived on people’s phones? What if any visitor could experience it — without needing a coordinator, a printed sheet, or a WhatsApp group?
That question became MonkiGo. And how it got built — in 60 hours, by one designer, using AI throughout — became a more interesting story than the product itself.
”Most designers stop at the wireframe. AI tools made it possible to keep going — all the way to a live URL.
AI-Assisted Workflow
How AI supported every stage
This wasn’t about using AI to generate screens. It was about using AI as a thinking partner across the full product cycle — from structuring data to shipping code.
01
Structuring the game dataset
The original hunt relied on a manually built Excel sheet mapping all 88 monkeys — their zone, difficulty tier, adjacency, and co-location data. I used AI tools to consolidate and clean this dataset into a structured format that could actually drive game logic. What was a reference document became a working data layer.
#Data structuring #Game logic
#Claude
02
Translating UX screens into working interfaces
Rather than stopping at Figma, I used Cursor to generate and iterate on the actual interface — working from a small number of core UX screens as the source of truth. AI-assisted development handled the heavy lifting: layout generation, navigation flow, interaction logic, and component structure. The design brief became the dev brief.
#UI layout #Navigation flow
#Interaction logic #Cursor
03
Rapid prototype development
With AI assistance accelerating the build, the prototype evolved from concept to a working web experience in under 60 hours. The system included zone selection, monkey browsing with clues, an “I found it” confirmation interaction, and an integrated feedback form — all functional, all testable in the real environment.
#Zone selection #Clue system
#Feedback loop #60 hours
04
Deployment and live testing
The prototype was deployed to a live environment and made accessible to real visitors at Upper Thomson MRT. This closed the loop from design intent to user feedback — not in a Figma prototype playback session, but in the actual context the experience was designed for.
#Live deployment #Real environment testing
#monkigo.starfishlim.com
What this changes
The designer who ships
The traditional design workflow has a clear ceiling: hand off to development, wait, review, repeat. AI tools don’t just speed up that process — they remove the ceiling entirely for the right kinds of projects.
The shift isn’t about replacing developers. It’s about expanding what’s possible before you need one. Design thinking can now reach further into the product — from problem through to a working, testable, deployable experience.
MonkiGo compressed all four of those stages into 60 hours. The same design sensibility that shaped the information architecture, the interaction model, and the UX copy also drove the build — because the AI tools made that continuity possible.
reflection
A reflection on the workflow
Working with AI across the full product cycle revealed something important: the designer’s judgment matters more, not less. AI accelerates the parts that are mechanical — layout, code scaffolding, data wrangling — but the decisions about what to build, what to cut, and what the experience should feel like still require a designer in the loop.
MonkiGo is an evolving prototype. But it shows what’s possible when a designer stops treating “shipped” as someone else’s job.
Outcome
Concept.
Prototype. Live.
MonkiGo is live at Upper Thomson MRT, accessible to any visitor who finds it. It’s not a pitch deck. It’s not a Figma preview. It’s a working web product — designed by one person, shipped in 60 hours.
The feedback form built into the experience continues to collect real responses from people who’ve tried it in the field — closing a genuine design research loop.
