Episode summary
Glenn Fogel, CEO of Booking Holdings, joined hosts of No Priors to discuss his company's transformation from a near-bankruptcy during the dot-com crash to a $130 billion travel giant, and how AI is reshaping the industry. Fogel, who joined Priceline in 2000 when it was worth just a few hundred million dollars and trading at $1 per share, has witnessed the stock rise 1,000-fold over 27 years. The episode centered on Booking's aggressive AI investments and Fogel's contrarian view that no competitive moat exists in travel technology, despite the company controlling 8.6 million alternative accommodation listings globally and processing $186 billion in travel annually. Fogel revealed that Booking is investing over $700 million in AI this year, already achieving 10% reductions in customer service costs while improving satisfaction through tools like Priceline's Penny agentic assistant, which has doubled adoption monthly. He detailed using Penny himself to plan a complex family trip to Europe, demonstrating how AI can handle multi-city itineraries, frequent flyer mile optimization, and multi-cabin bookings. However, Fogel acknowledged critical unknowns around token economics, model selection, and true ROI at scale. The conversation also addressed OpenAI's abandoned travel checkout feature, regulatory complexity in global travel, and whether AI agents will disintermediate platforms like Booking. Fogel argued the company's value lies not just in inventory but in partner relationships, regulatory compliance infrastructure, and problem resolution capabilities that appear simple from outside but require thousands of employees. On broader AI implications, Fogel warned that job displacement could occur faster than retraining programs can address, citing his own company's elimination of translation jobs and expressing concern that technology rejection could disadvantage the US versus China. He emphasized continuous employee upskilling and questioned whether government retraining programs can effectively address the transition, while noting that consumer AI sentiment depends heavily on how questions are framed in surveys.
