Penny, AI Travel Assistant

Customers had last-minute questions at key moments in the booking and post-booking journey, such as hotel questions at checkout. Without quick answers they hesitated or stalled, while traditional customer support carried the load.

Penny, Priceline’s proprietary generative AI travel assistant, started as a simple test on the checkout page for answering last-minute hotel questions. I set a testing-led approach: rather than design the full assistant up front, we learned through incremental testing and iterative design, expanding Penny one step at a time.

Alongside that, my team and I designed a North Star for Penny, a future vision of how AI could support discovery, browsing, and trip management, that we could test into. It gave every iteration a clear direction. I led a small team of two designers and a researcher and worked closely with engineering as the experience scaled quickly.

The testing approach. I set how we would grow Penny: start with one high-intent moment, test, learn, and expand step by step toward the North Star.

Design execution. I led design and research for Penny across the homepage, details, checkout, and post-booking experiences, with a team of two designers and a researcher.

Conversation that simplifies shopping without alienating anyone.

The hardest design challenge was finding the right balance: a conversational interface that works as a simple way to shop, while making sure we did not alienate our older customers.

Penny expanded to the homepage, details, checkout, and post-booking experiences, helping customers get answers, book with confidence, and manage trips. It was designed as more than a support chatbot, adding AI assistance at high-intent moments and aiming to reduce reliance on traditional customer support.