Travel destination AI assistant
Duration
2+ Days
Status
User Tested
For
Internal Consideration
In a 2-day hackathon to brainstorm ways we could utilize AI in our travel booking products, my team and I came up with a feature that would help a user pinpoint where exactly their next vacation should be. View a full video demo at the bottom of this page.
The team & my role
Rough ideas and problem statements were submitted to the organizers ahead of schedule so teams could be formed around them and hit the ground running. My team consisted of 1 product owner, 3 front end engineers, and me as the designer.
Our standard landing page with a link to new special feature: "Help me decide"
Goals
The idea we pursued centered around being able to describe the essence of your next vacation and have our assistant then generate realistic location suggestions that would be a good match.
Some example prompts we pre-generated to introduce the idea to users.
Strategizing
The team spent the first morning discussing goals, scope, what kind of experience our users would find most valuable, and how we realistically might be able to actually implement the idea into our existing platform. We decided the best place to insert this feature would be at the right on the landing page (letting the user either type in their destination as normal, or use our tool which would take them on an alternative route. Each route eventually landing the user onto a hotel search results page.
Logistics
Not only was our timeframe only 2 days, but we also had to balance regular work demands throughout. To stay updated on each others progress, we had a running meeting link and slack chat for the duration of this project which included morning kickoffs and end-of-day debriefs to make sure we stayed in close communication and alignment.
A very rough wireframe my team collaborated on together during our first team meeting to start laying out our vision.
Agile teamwork
Together we laid out a very rough visual in order to make sure we were all aligned and were all able to hit the ground running. Design and engineering worked in parallel so we could all be making progress simultaneously. As I fine-tuned design details, engineers refined their code, and the product owner started putting together our story.
Overall flow of our new feature incorporated throughout the existing funnel.
Concept to completion
I worked with our product owner to come up with dummy content for a prototype walkthrough. We ideated a bit on future states and how a feature like this could continue to evolve.
A new page we created that would be populated with custom suggestions.
Leaving room for error as the technology and feature would still be new, we made sure to give the user an out by either narrowing their custom search or starting over.
Deliverables
Not only did we have a high-fidelity prototype to demo, but our devs continually improved and pushed themselves to get as close to the prototype as they could, and in the end we had a live environment ready for our colleagues and judges to test out themselves.
Market opportunity
Business traveled has dropped and remained stagnant since the pandemic, but we’ve seen major increases in leisure travel, luxury travel, and digital nomad travel. This is a multi-billion dollar industry and as it continues to quickly evolve, it’s important to consider new ways we might be able to appeal to this growing exploratory market.
Detailed informational modal where a user can read more about any of their selected destinations.
Presentation
Teams had 10 minutes to share their project to the judges consisting of our CEO and a couple of SVPs. Projects were judged on originality, sustainability, scalability, presentation, impact, teamwork, design, and technical difficulty. We were able to seal the win and demonstrate how we could seamlessly integrate this within our existing travel booking platform.
Added value
Users would be able to type in a prompt in the same way they would if they just simply used something like ChatGPT. But what would set this apart would be the per-partner customizability, the visual walkthrough, and the direct-to-platform tie-in. For example, if a user was on an American Airlines loyalty platform, then their results would be filtered only for locations served by American Airlines. Users are able to view a map, photos, and highlights of each suggested location immediately without having to open new tabs or ask for more details. Getting results of available hotels and prices directly as a next step is an additional convenience.
Video walkthrough of our product state at the end of the hackathon.
Future thinking
A couple of follow-up user interviews validated the viability of our project. I’d love to see this implemented in an experiment to get further data and use rates. There’s a lot we could iterate on including the ability to refine and save results, or a guided walkthrough vs. an input box. Either way, I think the travel booking experience will continue to evolve and require innovation to stay relevant.