Booking.com and OpenAI personalize travel at scale
By combining its data with OpenAI’s LLMs, Booking.com enables smarter search, quicker support, and intent-driven travel
Booking.com is testing how AI can improve the early stages of trip planning. The platform has long been effective at helping users book hotels, flights, and activities, but it wasn’t as useful when travelers weren’t sure where to go or what they were looking for. With tools like ChatGPT, the company is now trying to make travel discovery more conversational - helping people explore ideas and destinations in a more open-ended, flexible way.
Key takeaways
- Shifting focus to discovery: While Booking.com has been strong at converting specific searches into bookings, it wasn’t as helpful during the early stages of planning, when people are still figuring out what they want.
- AI trip planner with OpenAI: The AI Trip Planner uses language models to respond to open-ended questions, suggest destinations, and build itineraries based on user input.
- Fast development cycle: A prototype was built in just 10 weeks by combining Booking.com's existing machine learning tools with OpenAI’s models and data.
- Blending data types: The tool brings together structured data like prices and availability with unstructured content such as user reviews and free-text descriptions to generate better recommendations.
- Natural language as input: Instead of clicking through filters, users can type questions like “Where should I go for a quiet beach vacation?” and get tailored suggestions.
- Highlighting lesser-known spots: The AI is also used to surface destinations that are often overlooked, helping steer travelers away from over-touristed areas.
- Working together across teams: The project came together quickly thanks to close collaboration between Booking.com and OpenAI, combining technical and product expertise.
Get the full story at OpenAI and Skift (subscription required)