For more than 25 years, Priceline has helped travelers find better prices on flights, hotels and rental cars. Now, the company is applying artificial intelligence to a much broader challenge: understanding what travelers actually want before they can fully articulate it.

That vision sits at the center of the technology strategy led by Sejal Amin, Chief Technology Officer of Priceline. Part of Booking Holdings, which generated nearly $27 billion in revenue in 2025, Priceline is using AI to move beyond traditional search toward more personalized travel planning. Amin's organization oversees product engineering, infrastructure, data and technology operations, but she says much of her job extends beyond technology itself.

"A big part of it is changing how the organization works," she said. "I spend as much time thinking about organizational design as I do technical architecture."

Reorganizing Around Products

One of Amin's first priorities after joining Priceline nearly two years ago was changing how engineering teams were organized.

Historically, the company operated through functional teams, with separate organizations responsible for front-end engineering, back-end engineering, design and product management. While effective in many respects, that structure created handoffs that slowed development and diluted accountability.

Priceline has since shifted to a product operating model in which cross-functional teams own customer experiences end-to-end. Rather than handing work between departments, teams now own complete products such as hotels, flights or rental cars. "The accountability stops leaking," Amin noted.

Shared capabilities, including data platforms, finance systems and customer infrastructure, have adopted the same philosophy. Those teams now operate as product organizations with transparent roadmaps that allow business and engineering leaders to remain aligned.

Understanding Intent, Not Just Behavior

The company’s AI strategy also reflects a broader shift in how Priceline thinks about customer data. Traditional travel platforms rely heavily on behavioral signals, such as previous searches or bookings. Amin believes the more valuable information lies in understanding customer intent.

Someone may consistently avoid long layovers, prefer boutique hotels or need larger accommodations when traveling with children. Those preferences often reveal far more than historical transactions alone.

That thinking helped shape the redesign of Penny, Priceline’s AI-powered travel assistant. Penny combines behavioral history with customer preferences to generate recommendations that feel significantly more personalized. Importantly, travelers can also review and update those preferences themselves. For Amin, that transparency builds trust while improving future recommendations.

Simplifying the Booking Experience

The redesign of Penny began with a simple observation. Customers enjoyed interacting with the conversational assistant because it was often faster than navigating traditional travel pages. The experience broke down whenever users needed to leave the conversation to complete another task.

Priceline responded by redesigning the experience around a single continuous workflow that carries travelers from discovery through booking without forcing them between multiple pages. Supporting that experience required considerably more than adding a chatbot.

Behind the interface, multiple AI agents work together while sharing memory and context throughout the interaction. Penny also draws on multiple large language models, including Anthropic’s Claude. "That’s where the actual hard part is," Amin emphasized. "It's not generating the text. It's building an interface that can follow an unpredictable user."

AI Is Changing Software Development

Generative AI is also transforming how Priceline builds software internally. Engineering teams increasingly use tools such as Cursor and Claude Code to accelerate development, helping Priceline deliver new capabilities more quickly. Amin said the productivity gains became especially noticeable beginning earlier this year, contributing to the rapid rollout of Penny's new capabilities.

Still, she cautions against measuring success simply by writing code faster. Code generation represents only a small portion of an engineer's work. Requirements gathering, testing, code reviews and refining customer experiences remain equally important. As AI takes on more routine development work, Amin expects traditional boundaries separating engineers, product managers and designers to continue fading.

She also distinguishes productivity from business impact. "Performance is whether the thing you shipped actually mattered," she underscored. That philosophy has reinforced Priceline's emphasis on maintaining stable product teams that develop deep customer understanding over time.

Removing Friction from Travel

Ultimately, Amin sees AI less as a way to replace travelers than as a way to eliminate the tedious work that surrounds planning a trip.

Rather than forcing customers to compare hundreds of reviews or manually evaluate dozens of similar properties, AI can summarize information, interpret traveler preferences and surface options that better fit individual needs. Amin regularly uses Penny herself, intentionally pushing it into unfamiliar situations to identify opportunities for improvement before customers encounter them.

Looking ahead, she sees voice interfaces becoming increasingly important, particularly in customer service, and Priceline has already begun experimenting with technologies including OpenAI's GPT Realtime and Amazon's Alexa+. Even so, she believes many travel decisions will continue to benefit from visual comparison rather than purely conversational interactions.

She is equally interested in determining how much responsibility travelers ultimately want to delegate to AI. Rather than assuming customers will hand over complete control, Priceline envisions a spectrum in which travelers decide how much assistance they want, from simple recommendations to increasingly autonomous trip planning.

Peter High is President of Metis Strategy , a business and IT advisory firm. He has written three bestselling books, including his latest Getting to Nimble . He also moderates the Technovation podcast series and speaks at conferences around the world. Follow him on X @PeterAHigh .