Case 05 · Hackasong MX 2025
4Health Habit Tracker
Build a health app concept in 48 hours — with AI at every step.
Visit demo ↗- Role
- Product Design Lead
- Event
- Hackasong MX 2025
- Timeline
- Oct 2025 · 48 hrs
- Team
- 3 designers
The challenge
Zero to award-winning prototype in two days.
Hours from blank canvas to award-winning prototype
Designers on the team across all phases
AI tools used across research, synthesis, PRD and prototype
Place at the Hackasong MX design hackathon
Context
Hackasong MX is an internal design hackathon — a 48-hour competitive sprint where teams design, prototype and pitch an app concept from zero. The 2025 edition challenged teams to combine human-centered design with AI tooling.
4Health tackles a real behavioral challenge: people struggle to build and maintain health habits. The app addresses it through habit creation, streak tracking, daily check-ins and progress visualization in a low-friction mobile experience.
What made this sprint different was applying AI deliberately across every phase — compressing weeks of discovery into hours without compromising rigor.
My contribution
- Led the team across all three phases of the 48-hour sprint
- Directed AI-assisted desk research and competitive analysis
- Generated synthetic user personas with generative AI
- Authored the product requirements document and refined the prompts that drove definition
- Built the high-fidelity interactive prototype entirely in Figma Make
- Presented the final concept to the judging panel and won the award
How it happened
The work, phase by phase.
AI-accelerated research & problem definition
Traditional discovery was not viable in 48 hours, so we ran an AI-augmented research sprint: desk research and competitive benchmarking, synthetic personas from behavioral data, then a drafted and refined PRD — a solid strategic foundation in under half a day.
The problem: habit formation is hard
Most people abandon health habits within two weeks. Root causes: no immediate feedback, unclear progress indicators, low commitment mechanisms — each addressed directly by streaks, check-ins and visual progress.
The Restarter
28–35, motivated to build healthy habits but has failed before. Needs clear early wins and gentle accountability.
The Builder
22–30, a consistent exerciser adding non-physical habits like sleep, hydration and mindfulness. Needs one low-friction tracker.
Rapid design with Figma Make
We went from wireframe sketches to a navigable high-fidelity prototype with real interactions and component consistency, using the PRD from phase one as the direct input for prompts.
Habit creation & streaks
Lightweight setup and streak mechanics that make consistency visible and rewarding.
Daily check-ins
One-tap habit completion from the home screen — friction-free logging so the behavior can become automatic.
Progress visualization
Weekly and monthly views with heatmaps and trend lines that make growth visible and motivating.
Prompt as design brief
The tight loop between research output and design generation removed handoff delay and kept the team refining and validating rather than building from scratch.
Deliverables
Presentation & award
The final stretch went to packaging, storytelling and pitching to the judging panel — showing how AI amplified each phase and demonstrating the interactive prototype live. The result was the Hackasong MX 2025 award.
- AI compresses time — discovery-to-prototype went from weeks to hours without losing rigor
- PRD as prompt — a well-structured brief becomes a design generator
- Method still leads — AI provided speed, the method provided the structure
Outcomes & impact
48 hours. Award-winning.
Place at Hackasong MX 2025
Research to prototype — blank canvas to interactive demo
AI tools integrated across every phase of the sprint
Designers — a small AI-augmented team outperforming larger squads
AI augments, not replaces, methodology
The method provided the strategic skeleton; AI filled each phase with speed and precision. Designers who master the combination become a category of one.
Great briefs make great prototypes
Output quality was directly proportional to PRD quality. Investing the first 16 hours in rigorous definition meant the design phase was generation and refinement, not exploration from zero.
Synthetic research is real research
AI-generated personas are a different research modality, not a shortcut. Grounded in desk research and behavioral principles, they give directionally valid insight at a fraction of the time cost.