Booking.com's strategy for scaling AI and managing token costs

Waters uses these tools to synthesize vast amounts of industry data, such as analyzing how rival platforms weight and display customer reviews. This shift extends beyond strategy; the company is embedding AI into its core operations, from engineering teams writing code more efficiently to HR and finance departments utilizing tools like Claude Code for complex forecasting. These efforts align with the firm’s broader 'Connected Trip' strategy, which includes direct partnerships with OpenAI, Anthropic, Google, and Amazon to integrate travel booking capabilities directly into conversational interfaces.

Maintaining this momentum requires a delicate cultural balance. Waters characterizes the internal climate as an 80/20 split: excitement over new capabilities tempered by the fear of losing market share to faster-moving rivals. As the company scales, it faces the challenge of managing ballooning token costs. Rather than imposing strict caps, Waters argues that spending is justified if the generated value scales proportionally. The firm is now focusing on granular visibility, ensuring product teams understand the return on their AI investments and identifying where cheaper, smaller models can handle simpler tasks without sacrificing efficacy. For leadership, the shift mirrors the transition to cloud computing, requiring discipline to prevent costs from spiraling while capturing the speed of a new technical era.