Overview
KAYAK, a leading travel search platform for flights, hotels, and rental cars, is actively exploring how AI can enhance the pre-booking experience.
I designed an AI-powered moodboarding tool giving users a space for exploration before moving forward with booking. Increased User Satisfaction Score by 16%.
Timeline
3 months
Role
Lead Product Designer
Tools
Figma, Jitter,
Rotato
Team
Skills
Interaction Design
Prototyping
Product Strategy
User Research
MISSION OVERVIEW
Design a user acquisition & product growth strategy for KAYAK AI, targeted towards young adult travelers.
HIGHLIGHTS
Goals
1
Understand the limitations of Kayak AI’s current beta model
Research
Usability Testing
Market Analysis
2
Synthesize research into a product opportunity.
Brainstorm
Journey Mapping
Concept Testing
3
Design a tool for travelers to move forward with booking
Brand
Product
Interface
Research
What does travel planning look like in the age of AI?
Findings
With KAYAK AI in beta, our team conducted usability testing using a think-aloud protocol to evaluate alignment with travelers’ mental models, complemented by 25 in-depth interviews to understand real world planning behaviors.
Planning is fragmented; travelers have different methods to organize evolving ideas.
Team Ideation Session: Crazy 8s + Dot Voting
Qualitative Usability Testing
With all the user research and meta-analysis, we started thinking of ideas for our product. We took every problem we discovered and brainstormed how we could solve it.
Notes Collected during Round 1 of Concept Testing (n=5)
Participant C during round 1 of concept testing
Challenging the LLM paradigm
The stakeholder was an SME of Kayak’s click-referral commission model, where the company earns a fee each time a user clicks a travel listing and is redirected (even without completing a purchase). I kept this in mind to incorporate in mid-fidelity design iterations.
Kayak AI leverages GPT to produce its text based recommendations
LLMs are pre-trained on vast amounts of data encouraging users to have open ended prompts
Empty State UX: Designing for possibility
With all the user research and meta-analysis, we started thinking of ideas for our product. We brainstormed potential solutions that touched on nuanced pain points during the travel planning process.
"This would be super helpful for my wife. She has a hard time picturing what we're doing ahead of time..she also isn't a huge fan of putting everything in a spreadsheet."
Participant A during round 2 of concept testing
Visual Integration: Kayak's Design System
With all the user research and meta-analysis, we started thinking of ideas for our product. We brainstormed potential solutions that touched on nuanced pain points during the travel planning process.
USER COMPLETION
90%
of users successfully understood and navigated Lucid with ease
USER SATISFACTION
~83%
of participants desired to implement Lucid in their travel planning process.
