Aster
A 0→1 AI assistant that helps people think, plan and follow through on everyday life, built to compete in the U.S. market.
- Role
- Product design lead
- Duration
- 1.5 years
- Focus
- AI conversational design, modular system, product strategy
- Platform
- Native mobile
Situation
SK Telecom's Global AI Platform set out to enter the U.S. market with an ambitious goal: build an AI assistant that could compete globally.
Early research reframed the brief. People weren't looking for more automation. They needed help thinking, planning and following through on everyday life.
Task
Shape aster* from 0→1 for a U.S. launch: find the key use cases, design against them, test and refine, then guide our production team in South Korea.
I partnered directly with the CEO and ML engineering leadership on product strategy, balancing what users needed with what was technically and commercially feasible.
Action
From generate to accelerate
Users didn't need more content, they needed faster, smarter starting points. Prompt suggestions helped people break big goals into smaller steps and get to a first plan sooner.
A hybrid interaction model
Aster combines the flexibility of chat with the clarity of structured UI, so people can speak naturally in some moments and use guided inputs when entering lots of detail.
Recommendations people can trust
Clarifying questions capture clear intent, in-chat suggestions encourage discovery, and contextual reasoning explains why each option was suggested.
Learning about users over time
Onboarding, preference capture and memory help Aster understand each person and grow more personalized with every interaction.
A modular plan framework
Plans are built block by block, flexing from travel to workouts to meal prep while reducing engineering effort and keeping users in control.
Raising the bar
I defined domain guidelines to improve LLM accuracy and consistency, and coached the Korean design team on accessibility and U.S. design patterns.
Result
Time to value dropped and more plans were completed, with users saying Aster understood them from the start.
Explaining the AI's reasoning turned output into guided decisions and improved trust and satisfaction, while revealing information progressively reduced decision fatigue. The work shipped as a working MVP for beta testing.
AI is changing our world at a rapid pace. A fail-fast mindset is crucial to keep momentum going.