



Project at a glance
2 min read
Problem
Trip details are scattered across different channels, resulting in inefficient and time-consuming trip planning.
Solution
Developed an AI-powered travel app that parses travel links, auto-generates itineraries, and provides interactive map navigation, helping users plan trips faster and more efficiently.
Impact
Currently in active coding phase, with an anticipated launch in December 2025.
Role
Product Designer
Contribution
Feature Scoping
UX/UI design
A/B Testing
Accessibility Design
Timeline
Jan 2025 - Present
Team
2 UX designer
1 UX researcher
1 Product Manager
3 Software Developers
Overview
Designing smart link recognition
We use AI to analyze uploaded screenshots, links, or short videos and automatically extract key details like destinations and bookings. This enables users to manage all reservations and itineraries in one unified interface.
Visualizing travel destination and itinerary
Designed a flexible itinerary system that integrates location, transportation, and trip types. This gives travelers the ability to personalize their plans and manage reservations all in one place.
Adding AI-powered realtime travel suggestions
Developed an AI-driven feature that provides instant itinerary updates based on weather and trip context, helping travelers adapt their plans seamlessly and discover new opportunities.
Problem
Initial Problem Discovery
Trip details are scattered across different channels, resulting in inefficient and time-consuming trip planning.
User research
Explore Problem Space
We used a mixed-methods approach to identify product-market fit, conducting 20 interviews and collecting 350+ survey responses to understand user needs.
While I won’t go into every detail here, I’ve documented my findings and analysis elsewhere. Feel free to reach out if you’d like to discuss my process further
20+
Interviews
350+
Surveys
100+
Articles
Understanding users
Dynamic User Journey Map
To understand the complete travel experience from the user's perspective, we mapped out the journey from idea to post-trip sharing.
My Design Process
Prioritize displaying the map, with the itinerary fixed in a dedicated window
Prioritize displaying the itinerary
AI Exploration
Exploration in Leveraging AI
Even with thorough planning, unexpected disruptions can ruin expectations and lead to frustration.
Opportunity 1
AI Travel Tips
I designed the AI-generated Travel Tips module to proactively surface context-aware recommendations, combining:
Location-specific advice(e.g., NY tap water is cold, dry air in March)
Seasonal considerations(e.g., layering in winter)
User intent recognition(e.g., camera gear for content creators)
Packing psychology(essentials vs bonus)
Instead of giving generic tips, we deliver personalized and purposeful suggestions that align with user mindset: “What will I actually need there?”


Opportunity 2
Adaptive Planning: Weather-Aware Itinerary Updates
AI + crowd insights (“80% of users recommend this”) builds trust
Minimal interruption via one-screen, two-option UI
Flexible control — no forced changes, user decides
Map-based visualization grounds the recommendation in real context
Design Validation
During implementation, we ran into technical issues with our first infinite scroll design, where the itinerary kept loading as users scrolled—like a social media feed. This setup caused performance problems because each scroll triggered multiple real-time API calls, leading to higher server costs and laggy interactions.
To fix this, we switched to a paginated tab view, loading each day’s content separately. This change reduced API calls, eased server load, and made the interface much smoother and faster for users.
Final Impact
Currently in active coding phase, with an anticipated launch in Dec 2025
Reflections
One-canvas cross-functional collaboration
The project touched three sub-teams (ML, frontend, business). Any misalignment amplified risk. I therefore built a “visual decision canvas” that merged user journeys, tech constraints, business goals, and design tenets in a single FigJam board, color-coded and updated live. Stakeholders could now lock priorities in a 30-minute weekly “sync-and-decide” slot, while potential trade-offs (e.g., model latency vs. interaction smoothness) surfaced early. This transparency taught me to use visual language to accelerate decisions, and to wield a shared component system to keep execution coherent—showing that a UX designer is not just a “pixel pusher” but a translator and tempo-setter for the whole team.
Considering development effort and Revenue-Generating Design
As designers, we aim for aesthetics, but in B2B companies, design must align with value creation. It’s crucial to balance front-end work with business impact and find opportunities to drive revenue. We must also recognize limitations, like data gaps or designs slowing down the website.

















