Discover – Make sense of the ambiguity

With no clear problem statement or structure, I began by anchoring the experience in user needs and business intent.

• Stakeholder alignment: Led a kickoff to understand business priorities—engagement and retention—and clarify what “AI coaching” meant in technical terms.

• Proto-personas & journey sketching: Mapped out first-time user mindsets—new to personal development, easily overwhelmed, curious about AI.

  

  Illustrations realized on Midjourney.

• Benchmarking: Used ChatGPT and Mobbin to analyze onboarding and challenge loops from apps like Noom and Headway.

• Visual exploration: Used Midjourney to create early visual direction that supported emotional tone—bright, hopeful, and playful.

Insight: Users don’t want more information—they want direction. A successful experience needed to focus, guide, and encourage.

Define – Reframe the problem and build a plan

I translated the ambiguity into a user-first opportunity:

Reframed Problem:
“Users feel overwhelmed by vague self-improvement tools. They need a focused, motivating entry point that builds trust with AI from day one.”

Hypothesis:
“If users are matched with a personalized challenge and introduced to a helpful AI coach right away, they’ll feel more confident and motivated to continue.”

From there, I defined key UX requirements:

• Personalized onboarding tied to user-selected growth goals

• A coach avatar that builds connection and trust

• A modular dashboard to track challenges and progress

• A frictionless experience across mobile and desktop

Design – Prototype fast and build clarity

I translated strategy into experience with rapid iterations.

Onboarding Flow:
Designed a welcoming flow where users entered via a magic code, selected an avatar, and shared key info (goals, email, preferences).

• Real-time validation minimized frustration

• Avatar selection added playfulness and emotional engagement

AI Coach Integration:
The AI coach was introduced early as a friendly guide.

• Provided challenge suggestions and feedback

• Used tone variation testing (via ChatGPT) to humanize its responses

Challenge System:

• Users could join individual or team challenges

• Progress was visualized via gamified dashboards and point systems

• A leaderboard and badges reinforced continued participation

Real-time Reflection:
After completing a challenge, users are encouraged to reflect on their learning. This self-reflection feature allows users to input thoughts about their progress, lessons learned, and areas for improvement. AI provides feedback on these reflections, suggesting new challenges or learning paths for users to explore.

Gamification:
Gamification is a central aspect of Seed-Up. The inclusion of a leaderboard motivates users to strive for better performance in challenges. There are also rewards for completing tasks, such as badges or recognition on the global leaderboard. Users can also earn streaks for completing consecutive challenges, which further fuels motivation and consistency.

Validate – Test and iterate

To derisk the design before handoff, I ran a Maze usability test with 15 users.

Findings:

• Users found the experience playful and motivating

• Most confusion came from seeing too many challenges at once

Iterations:

• Reduced visual clutter and introduced one challenge at a time

• Simplified dashboard to focus on immediate actions

Impact

Seed-Up succeeded in delivering a highly engaging platform that blends personal growth and gamification. With the AI integration, users receive personalized experiences that are both motivating and challenging. The incorporation of real-time feedback and dynamic challenge creation provides an experience that is constantly evolving, ensuring that users remain engaged and continue progressing in their professional and personal development journey.

Reflection

This project challenged me to lead through ambiguity. With no brief or user data, I built a product experience from the ground up. It taught me the power of prototyping as a tool for clarity and reinforced that AI can be a powerful coach—when designed with care.