Travel Through the Lens of AI with Booking.com CEO Glenn Fogel

Glenn Fogel (guest); No Priors / Sarah Guo (host) · podcast · 2026-07-11 Reader 🖥️

practical · expository interview / conversational podcast dialogue — A single expert (Booking Holdings CEO Glenn Fogel) is interviewed by a VC host (No Priors, Sarah Guo). Propositions emerge through Q&A and are often restated; the thrust is prescriptive (how to run a business, deploy AI, and choose a career in the AI era) rather than a formal theoretical argument, though it carries macro observations about bubbles and labor.

Structure

Key terms

• no moat / 'there is no such thing as a moat'
  No competitive position is permanently protected from innovation; a company may hold a real 'competitive advantage in areas' today, but it 'can go away tomorrow', so the only durable strategy is perpetual re-invention.
• agent / personalized agentic travel assistant (Penny)
  An AI system that knows everything about a traveler, evaluates an infinite set of permutations no human could, plans and re-plans complex multi-leg trips (including miles-vs-cash trade-offs), and—crucially—serves as the single point of contact that anticipates and repairs cascading failures.
• agency (customer agency) Agency
  The traveler's desire to make, or at least confirm, the final decision; on complex choices the agent should defer and double-check with the human, while acting more autonomously on routine ones.
• token economics AI Unit Economics
  The live cost-management discipline of deciding which model to use, for which purpose, and when, and of tracking how many tokens a task consumes—weighed against an as-yet-unknown ROI and customer lifetime value.
• complexity (of the travel business)
  The operational, supply-side and regulatory intricacy of actually serving travelers and partners—loading and managing inventory, helping property managers, meeting merchant-of-record rules worldwide—that outsiders underestimate and that constitutes the incumbent's real (if temporary) advantage.

Propositions

• explicit⚡ disputed & corroboratedThere is no such thing as a moat: no position protects a company against innovation; a firm may hold a competitive advantage in some areas today, but it can disappear tomorrow.conf: high
  contradictsWe've Stopped Arguing About Whether AI Changes Software · p13 — Fogel denies that any advantage is permanent ('there is no such thing as a moat'); p13 warns the AI-adoption flywheel can harden the laggard gap into 'a permanent moat rather than a lag'—exactly the permanence Fogel rejects.
  supports ←The New Rules for Becoming an Enterprise Leader · p9 — Both argue no fixed position protects a company against a fast-moving environment, converging on adaptive capacity as the only durable advantage.
  supports ←Transcript for Cursor Team: Future of Programming with AI | Lex Fridman Podcast #447 · p26 — Michael's 'if you stop innovating on the product, you will lose' — moats don't protect you in this space — matches the later, more general claim that no competitive position is secure against innovation in the AI era.
• explicitThe only way to win over the long term is to continuously develop new services and new ways of doing things for customers.conf: high
  refinesThe Founder's Playbook: Building An AI-Native Startup · p15 — p15 makes moat-building against copy-cat incumbents the central imperative; Fogel refines it—advantages exist but are never permanent, so the real defense is perpetual re-invention rather than a static moat.
• explicitA business exists to do something of value for its customers; Booking's two customers are travelers and partners, and AI—especially LLMs—is a tool to serve both better, cheaper and more valuably, not a replacement for the marketplace.conf: high
• explicitAI will make trip planning dramatically easier through personalized agents that know the traveler, evaluate an infinite set of permutations, and can out-serve any human travel concierge.conf: high
• explicitThe most important value of an AI travel agent is being the single point of contact that predicts and handles cascading failures—'travel is like dominoes'—when things go wrong.conf: high
• implicitCustomers always want agency—to make or at least confirm the final decision—so an agent should confirm with the human on complex choices while acting more autonomously on routine ones.conf: medium
• explicitThe AI wave rhymes with the late-1990s Internet boom: enormous real value will be created alongside massive speculative failure—of roughly 1,500 dot-com IPOs only about two dozen survived.conf: high
• explicitBooking's agentic tool (Penny) is producing real gains—adoption doubling monthly, higher conversion, faster search, shorter path to booking, lower cancellations—but is still too small in absolute terms to move the ~$186B / 1B+ room-night business.conf: high
• explicitcorroboratedThe binding management questions for AI in production are its economics: token economics (which model, for which purpose, when), cost per outcome, and an as-yet-unknown ROI and customer lifetime value.conf: high
  supports ←When Your Buyer Is an AI Agent · p3 — Concrete outcome-based pricing implementations (Sierra, Intercom's $0.99-per-resolved-conversation Fin AI) corroborate that cost-per-outcome, not seat/session count, is becoming the operative unit of AI commercial economics.
• explicitAI customer service already lowers cost per contact (about 10%) while raising customer satisfaction, but some customers still want a human, so the mix must be balanced around what is best for the customer.conf: high
• explicitExcess cash flow should be allocated by priority: reinvest in the company if the ROI justifies it, otherwise pursue acquisitions, otherwise return capital to shareholders, who can then invest it better than the company can.conf: high
• explicitThe incumbent's real, underappreciated defense against an AI-native entrant is hard-to-replicate scale and connectivity—8.6M alternative-accommodation listings, deep partner relationships, and the operational work of actually serving supply.conf: high
  supportsBaseten CEO Tuhin Srivastava on the AI Inference Crunch, Custom Models · p2 — Fogel's claim that supply connectivity and listings an AI-native entrant cannot easily obtain protect the incumbent parallels baseten's thesis that the independent application layer survives against the labs via unique assets/user signal the labs cannot access.
• explicitTravel's heavy and increasing worldwide regulation (merchant-of-record rules) is a barrier only large-scale players can afford to meet, so newcomers underestimate what it takes to displace incumbents.conf: high
  supports#36 Thinking big from day 1 to Unicorn | Pål Næss (Gelato, a Norwegian Unicorn) · p12 — Both treat the operational and regulatory complexity of operating at global scale as a genuine competitive barrier that outsiders underestimate.
• explicitTechnology-driven job replacement is not new (agrarian-to-urban shifts; machine translation eliminated Booking.com's ~40-language human translation and customer-service jobs), but the novel danger is the speed mismatch between job destruction and new-job creation, and the plight of those who cannot retrain (the 50-something truck driver).conf: high
• explicitCompanies and society have an obligation to continuously upskill people toward AI literacy, because government retraining has historically failed and a fear-driven rejection of technology would leave a society disadvantaged relative to others (e.g., China).conf: high
• explicitClaims that the public is broadly anti-AI are largely an artifact of framing—people love using AI tools but fear job and career loss, so questions are asked 'a certain way on purpose'; the honest response is level-headed calibration, not doomerism.conf: medium
• explicitYou only get one life, so—for those with the freedom to choose—choose your path wisely by asking what you are really trying to achieve, rather than defaulting to a comfortable, prestigious track.conf: high

Arguments

Narratives

• Fogel joined Priceline in February 2000, taking his 1999 bonus first—then the Nasdaq peaked the week he went long Internet. Priceline's market cap collapsed from ~$30B toward a couple hundred million, the stock fell below $1, and a reverse split (to ~$6) staved off delisting. He stayed 27 years as it grew to a ~$180B peak—up roughly a thousandfold.
  → Embodies the 'fight for the customer every day / you never stop' ethos and the survival-through-reinvention thesis; also grounds the dot-com/AI parallel in lived experience.
• Fogel recounts using Priceline's agent 'Penny' the other night to plan a complex family trip to Europe—two cabins (adults up front, young-adult kids in back), one member returning to a different city, hotel-on-arrival vs. travel-that-night decisions, restaurants—then adding frequent-flyer miles and having it work out, back and forth, that he'd pay miles for the adults and cash for the kids.
  → Concrete proof-of-value for the personalized agentic-assistant vision.
• Around 2005 Booking.com handled hotel reservations, content, scripts and customer service in over forty languages—all translated by human beings, since there was no machine translation. Machine translation later erased all those jobs entirely.
  → Makes the abstract labor-disruption argument tangible from an operator who lived it, and sets up the speed-of-change concern.
• Fogel fears too many of his law-school classmates chose law simply because it was a well-paid, comfortable default path, and later in life found themselves wondering 'what did I do?'.
  → Humanizes the 'choose wisely / you only get one life' values message.

Questions addressed

• ✅ How should people think about travel through the lens of AI?
  AI is a beneficial tool that lets Booking serve travelers and partners better, cheaper and more valuably—not a force that makes travel companies worthless; the market's swings on OpenAI's checkout feature overread a shallow view of the business.
• ✅ Will agents transacting on users' behalf disintermediate Booking, and does its scale (e.g., 8.6M listings) form a durable moat?
  Booking is itself building the agents; supply connectivity, operational depth and regulatory complexity are hard for a pure-AI entrant to replicate—but Fogel insists this is a temporary advantage, not a permanent moat.
• ✅ Should founders and CEOs keep going or sell in this environment?
  No general rule; it depends entirely on the specifics—conviction about future value, options available, and one's actual motivation (build something that matters vs. make money).
• ✅ Will AI cause a jobs apocalypse?
  Tech-driven displacement is old news; the real risks are the speed mismatch between destruction and creation and the non-retrainable worker, met by continuous upskilling and honest debate rather than fear.
• ⭕ What is the ROI and lifetime-value impact of the agentic tools, and are they net-accretive at scale?
  Explicitly left open—'things that we don't know yet that we're going to have to work on'; cost, ROI, LTV and repeat/loyalty effects are still being measured.

Gaps

• incompleteFogel claims the personalized agent will do 'so much more than any human travel agent could' and that no moat protects incumbents, yet never reconciles the tension: if the agent's value is itself replicable AI and supply can be reached via APIs, what specifically stops OpenAI's or Google's agents from disintermediating Booking beyond the assertion of 'complexity'?
  Read next: Read frontier-lab agentic-commerce plays (e.g., Google's Universal Cart in extract-google-may-2026-ai-recap) and disintermediation analyses of marketplaces to test how defensible supply connectivity really is.
• incompleteThe unit economics are raised but left unresolved: Fogel states cost per booking, ROI and lifetime value are not yet known, so whether the agentic tools are net-accretive at scale is undetermined even as adoption doubles monthly.
  Read next: Read ai-unit-economics treatments of agent cost-per-outcome and whether per-contact efficiency gains survive at scale (e.g., extract-baseten-ceo-inference-crunch, extract-building-ai-agents-for-the-enterprise cost-structure gap).
• incompleteFogel names the labor speed-mismatch and the non-retrainable worker as the central problem but offers no mechanism beyond corporate upskilling—while conceding that fifty years of government retraining has largely failed—so the policy and social response is left open.
  Read next: Read ai-labor-disruption and ai-regulation material on displaced-worker transitions, safety nets and the political economy of dislocation.
• incompleteThe 'customers still want a human / want agency' balance is asserted but never operationalized: no criteria are given for when an agent should defer to a human, insert a confirmation step, or hand off to a live person.
  Read next: Read human-in-the-loop design work on judgment points, approval steps and confirmation UX to specify the deferral criteria Fogel leaves implicit.

Closes gaps elsewhere

• AI Companies Are Hiring More · g4 — Fogel supplies the analysis g4 flags as missing: net job creation does not help the specific displaced worker, the destruction-vs-creation rates likely diverge, and the non-retrainable case (the 50-something truck driver) is the real problem.
• Building AI Agents For The Enterprise · g5 — Fogel brings the omitted cost dimension g5 flags—per-contact cost (down ~10%), model-price exposure and 'which model for which purpose and when', token consumption, and ROI/LTV—though by his own words the ROI/LTV itself remains unresolved ('things we don't know yet'), so he surfaces rather than settles the economics.

Related documents

• The Founder's Playbook: Building An AI-Native Startup
• #36 Thinking big from day 1 to Unicorn | Pål Næss (Gelato, a Norwegian Unicorn)
• Agents Over Bubbles
  Agentic AI AI Bubble Debate Agency Moats & Defensibility
• Arvind Narayanan on Why AI Isn't All That Revolutionary
  Adaptive Advantage / Adaptive Capacity AI Labor Disruption Agency Human-in-the-Loop
• AI's biggest critic has lost the plot
  AI Backlash AI Bubble Debate AI Unit Economics
• Howard Yu's Post — American AI models went from 70% of usage to 30% in 12 months
  Agentic AI AI Unit Economics Moats & Defensibility
• Why Enterprise AI Has a Leadership Problem
  Agentic AI AI Bubble Debate AI Labor Disruption
• Everyone I recommend on AI is optimistic. Except Cory Doctorow.
  AI Backlash AI Bubble Debate Human-in-the-Loop
• The latest AI news we announced in May 2026
  Agentic AI Agentic Commerce Moats & Defensibility
• The New Rules for Becoming an Enterprise Leader
  Adaptive Advantage / Adaptive Capacity Agentic AI AI Labor Disruption