Case Study
Redefining Hotel Booking with AI-Powered Travel Assistance
Voytravel Mobile App – AI + Human Travel Planning
Client Overview
Voytravel is an early-stage startup revolutionizing travel planning with an AI-powered assistant that helps users discover, plan, and book hotels through a conversational interface. It bridges the gap between AI convenience and human concierge-level curation.
The Challenge
Voytravel needed to build a mobile-first product that combines hotel search, AI interaction, and real-time booking flows—all with startup-level speed and polish. As part of a distributed remote team with rapidly shifting priorities and a tight demo deadline, the challenge was to ship a fully working, production-grade app within weeks.
Key Issues
- Complex feature scope: chat, hotel browsing, filtering, multi-currency support, and payment workflows
- High latency on AI responses requiring performance optimizations
- Multiple environments (Expo Dev Client, iOS provisioning, API mocks) slowing down integration
- Cross-functional coordination across design, AI, backend, and mobile teams
Our Process
1. Frontend Architecture & Feature Delivery
I contributed to the mobile app using Expo Router (React Native) and coordinated with backend teams on API integration:
- Built core screens: hotel listings, hotel details, Smart Picks, dynamic chat UI, user input interactions
- Implemented persistent session-based chat UI with support for structured AI tool responses (Zod + TRPC validation)
- Collaborated with backend team to debug and integrate hotel search APIs, Redis cache, AI pipeline bottlenecks
- Ensured app performance and responsiveness despite large payloads and long AI response times
2. Fast Iteration and Demo Readiness
With tight deadlines ahead of a product showcase, we focused on scope alignment and rapid iteration:
- Worked with the product team to identify and prioritize critical features for the WebX demo
- Created fallback UIs and graceful error states to mitigate incomplete or flaky AI responses
- Helped structure PR workflows and Notion tickets to keep the distributed team aligned across time zones
The Results
Voytravel successfully demoed a fully functional hotel-booking assistant at WebX Tokyo. The app showed how AI could enable smooth travel planning, from browsing to booking, all in a chat interface. My contributions were critical in making the frontend stable, performant, and presentation-ready within a matter of weeks.
- The app combined natural language AI with hotel APIs to deliver curated travel recommendations in real time.
- AI-Driven Travel Assistant
- Despite shifting scopes and delayed backend dependencies, the mobile frontend was completed on time with full functionality.
- Built for Demo Under Extreme Constraints
- A session-based, tool-integrated chat interface handled structured hotel suggestions, filters, and booking flows.
- Modular Chat System
- Synchronized design, AI, frontend, and backend efforts in a fast-paced, partially async work environment.
- Cross-Team Alignment