Detour

Systems Design

Systems Design

AI Interaction Design

UX Research

Mobile (iOS)

Travel planning tools help you decide where to go. Booking platforms help you buy tickets. But no product manages the relationships between your bookings. The time dependencies, priorities, and flexibility that determine whether your trip adjusts gracefully or falls apart when something changes.

Detour is an AI itinerary orchestration layer. It understands the connections between every reservation in your trip. Before departure, it stress-tests your schedule for timing conflicts. On the road, it handles daily adjustments like weather changes. And when a major disruption hits, like a cancelled Shinkansen or a typhoon warning, it maps the full cascade of downstream impact and generates a recovery plan that protects what matters most to you.

This case study explores a core design challenge of the agentic AI era: how do you give users meaningful control over an AI that acts on their behalf without slowing them down when they need help most?

Methods

Research

Diary Study

Critical Incident Interviews

Competitive Experience Audit

Design

Itinerary Structure

Agent Control Model

Recovery Flow

Validation

Scenario-Based Usability Testing

IA Iteration

Trust Perception Testing

©2026 | Kevin gao

huawengao@gmail.com

©2026 | Kevin gao

huawengao@gmail.com