Case study - I
chapterit
"Can he drive results?"
From Models to Outcomes: Designing a Unified AI Experience
Cloudhands is a multi-modal AI platform designed to simplify how users interact with AI tools.
Instead of navigating separate products for text, images, video, and code, the platform brings everything into a single, unified experience, enabling users to focus on outcomes rather than underlying models.
Role
Founding Product Designer& Product Manager
Team
1PM, 2 ML Engineers, 1 Full-stack
Timeline
2026 - present
Focus
AI UX, ambiguous product definition
Outcome
Private beta → paid pilot in 4 months
01 - Context
ChapterIt is a digital platform designed to help athletes capture, structure, and reflect on their journey over time.
Rather than focusing solely on performance metrics, the product reframes athlete development as a combination of performance and narrative—bringing together stats, milestones, and personal moments into a single, cohesive system.
02 - My Role
Head of Product & Design
Defined product vision and MVP scope
Led roadmap planning and feature prioritization
Designed core user flows and system architecture
Built high-fidelity product experiences using AI-assisted development (Claude Code)
03 - The Problem
AI capabilities are powerful but fragmented.
Users often need:
Multiple tools
Multiple subscriptions
Different interaction patterns
To accomplish even simple tasks, users must understand which model to use and how to use it—creating unnecessary complexity.
04 - Key Insight
Users don’t care about models—they care about outcomes.
This insight reframed the product from a collection of tools into a cohesive system designed around user intent.
06 - Concept Exploration
Early design exploration focused on how to unify diverse AI capabilities into a single, understandable experience.
Key questions included:
Should the experience be tool-based or intent-based?
How much should users control vs. automate?
Can different modalities share a common interaction model?
05 - Defining the Core System
The platform evolved into a multi-layered system that balances flexibility and consistency:
Multi-Modal Generation
Chat
Image
Speech
Audio
Code
All generation types were designed to feel cohesive while supporting their unique outputs.
Assistant Layer
A conversational interface that acts as a central entry point for interacting with different capabilities.
Recipes System
v1: Pre-filled prompts for quick access
v2: Structured forms for guided generation
This allowed users to move from exploration to repeatable workflows.
Model Comparison (Arena)
A feature enabling users to compare outputs across different models, helping them understand tradeoffs without requiring deep technical knowledge.
User Layer
Profiles
Preferences (including light/dark mode)
Supporting personalization and continuity across sessions.
07 - Tradeoffs
From Figma to Claude Code
To accelerate development, I transitioned from traditional design workflows into AI-assisted implementation using Claude Code.
Benefits:
Rapid iteration from concept to production
Reduced friction between design and development
Faster validation of ideas
Challenges:
Maintaining visual and interaction consistency
Establishing tighter design constraints within code
This shift required a more integrated approach to design, where thinking and execution happened in parallel.
08 - Outcome
The result was a functional MVP that transforms athlete tracking into a cohesive, story-driven system.
By combining performance data with narrative context, the platform enables athletes to better understand their progress, reflect on their journey, and present a more complete representation of their development.
09 - Whats Next
Future iterations focus on expanding intelligence and automation within the platform:
AI-generated summaries of athlete progress
Automated highlight creation from moments and events
Deeper performance insights connected to narrative milestones
These directions continue to push the platform toward a more intelligent, personalized experience.