INTRODUCTION
Gen Z is spending more on travel than any previous generation their age — and AI tools are racing to serve them
Expedia and its competitors have been rapidly rolling out AI-powered features, such as smart filters, AI summaries, natural language search, and Q&A bots. While these features are for all travelers, Gen Z stands out as the segment most primed to engage; They're among the biggest travel spenders and the fastest AI adopters of any demographic.
Competitors that have salary benchmarking tools
PROBLEM
Expedia's AI Filters feature isn't capturing the users most likely to adopt it
Despite being the highest-potential audience for these tools, Gen Z hasn't been studied. This research set out to change that — investigating how they navigate the travel search experience, where friction lives, and where AI can actually earn their trust.
Expedia's current AI Filters
SOLUTION PREVIEW
An AI-prototyped redesign of Expedia's filters, backed by four research insights
1/2
Then, view in-depth details about base salary and total compensation
View information on median salary, percentile ranges, and total compensation for your selected role and market.
2/2
Then, view in-depth details about base salary and total compensation
View information on median salary, percentile ranges, and total compensation for your selected role and market.
LANDSCAPE
Mapping what AI in travel already looks like
Before talking to users, I conducted a literature review spanning six major OTAs — Booking.com, Tripadvisor, Google Travel, Airbnb, Kayak, and Priceline — plus non-travel platforms known for AI-driven discovery, including ChatGPT, Spotify, Pinterest, and Perplexity. Six AI feature categories consistently emerged across OTAs: trip planning, search and filtering, property summaries, Q&A, customer service, and smart pricing. I paired this with desk research into Gen Z travel behavior, which surfaced a tension that shaped the rest of the project: 51% of Gen Z trust AI-generated itineraries, but only 7% trust AI for actual booking. Enthusiasm for AI in ideation doesn't translate to trust where it counts.
INTERVIEWS
Six moderated interviews, end-to-end
I recruited six Gen Z travelers and ran moderated sessions covering their full trip planning journey — from initial inspiration to lodging selection — with a usability task evaluating Expedia's AI Filters feature on mobile. The interviews were open-ended enough to surface unexpected behaviors, but structured around three questions: How do you get the idea to travel? How do you find and book a place to stay? And what did you make of the AI Filters?
1/4
First, select a role and define the market
Select a role to benchmark by choosing a job title and career level, which indicates the seniority within that role. Then, specify additional details such as location, industry, company size, and other relevant criteria.
2/4
First, select a role and define the market
Select a role to benchmark by choosing a job title and career level, which indicates the seniority within that role. Then, specify additional details such as location, industry, company size, and other relevant criteria.
3/4
First, select a role and define the market
Select a role to benchmark by choosing a job title and career level, which indicates the seniority within that role. Then, specify additional details such as location, industry, company size, and other relevant criteria.
4/4
First, select a role and define the market
Select a role to benchmark by choosing a job title and career level, which indicates the seniority within that role. Then, specify additional details such as location, industry, company size, and other relevant criteria.
RECOMMENDATIONS
Rapid prototyping to turn insights into concrete recommendations
With the four insights mapped, I used AI to rapidly prototype a redesigned AI Filters experience — moving fast from concept to interactive mockup without needing to build in Figma from scratch. This let me test the core interaction pattern quickly and focus my energy on the design thinking rather than production. The central recommendation: replace the blank input box with selectable interest tags tied to travel style. Users pick from options like "Near parks and nature," "Easy street access," or "Boutique local stays" — giving them a clear entry point and letting the AI surface filters that are actually relevant to their trip before they've typed a single word.
REFLECTION
My key learnings from the internship
Special thank you to my manager, mentors, and team members who made this project possible!
Ensuring early stakeholder involvement = smoother outcomes
Without early involvement, stakeholders are less aligned, making buy-in harder to achieve
Real-world projects come with real-world limitations
Constraints are inevitable; clear priorities keep you on track
Never be afraid to ask questions
Being transparent about what I didn't know was the quickest way to gain context and deliver quality work
If I had more time…
Access for different users
Different types of users have different access levels to data
Handling jobs with limited data
Combining multiple jobs to give users an idea of a representative salary for unique roles