How large language models and in-house ai tools are redefining how travelers discover and book trips
Jul 31, 2025
Booking.com executives, in a July 2025 earnings update, emphasised AI’s growing influence on travel discovery, positioning large language models (LLMs) as a new channel to diversify traffic and enhance user experience. Their optimism is supported by stable search performance and strategic investments in AI-powered tools.
Key takeaways
- LLMs as discovery engines: Booking.com sees generative AI not as a threat but as an opportunity—LLMs are increasingly used by travelers to get inspiration and trip ideas, and the company aims to integrate with them as partner platforms.
- Stable core traffic: Despite AI integrations, Booking.com reports no decline in traditional Google-driven or direct traffic—both continue to grow, with direct channels comprising around 60% of traffic.
- In‑development travel model: The company is working on its own proprietary “travel model,” though no further details or timeline were disclosed. It’s part of expanding into modern travel search channels.
- Existing AI applications: Booking.com leverages AI to reduce live-agent contacts, accelerate customer service responses, and improve satisfaction through voice‑enabled agents. It also contributes to ChatGPT’s Agent Mode via Booking.com data in the AI Trip Planner.
- Marketing shift to social and performance: Marketing spend is shifting from brand advertising toward social media, with a 25% YoY rise in social spending and sophisticated ROI tracking backed by data models and partnerships (e.g. with Meta).
- Strong financial performance supports AI investment: In Q2 2025 Booking.com saw revenues rise 16% to $6.8 billion and room nights booked up 8% to 309 million, underpinning continued investment in AI and digital transformation.
Get the full story at Marketing Week
