Hyatt’s chief executive says the travelers who reach Hyatt through its apps on ChatGPT spend more but rarely finish a booking there. “They are younger and spending a lot more on travel”, Mark Hoplamazian said on Hyatt’s third-quarter call on 29 October, according to The Motley Fool’s transcript. He added that “the booking activity where it terminates or concludes rather in that environment is still very low”.
Other disclosures show the same gap. Uber counted four agent rides and around 500 agent Uber Eats orders in a Halloween week of about 320 million trips and orders, or 0.00016%, by our arithmetic. Booking said LLM referrals are under 1% of Booking.com’s room nights. In Forter’s data, ChatGPT Instant Checkout bought at 24% of the stores that received any agentic order. On 13 November OpenAI launched ChatGPT Wallet, saying “we’re now helping people complete purchases”.
Business Pill · Agentic Commerce Explained
An evergreen Business Engineer explainer on what changes when the shopper is an agent. It teaches the concept only and does not discuss Hyatt, Uber or ChatGPT Wallet.
The key insight: As we read the disclosures, agents are getting shoppers to the point of purchase more often than they are completing it: Hyatt’s CEO describes higher-spending ChatGPT travelers but bookings concluding there as “still very low”, and Uber’s agent orders were 0.00016% of a 320-million-order week, by our arithmetic.
The remarks came in answer to Bank of America analyst Shaun Kelley, who asked how agents change the booking journey and how Hyatt can steer them to its direct channels. Hoplamazian opened with a level set: “the actual bookings that are being completed that originate through LLM platforms is very, very small”. He added: “It’s coming, and it will grow fast”.
On ChatGPT specifically, he said: “Our learnings from being very early in representing Hyatt directly in the apps that you can find through ChatGPT’s platform have taught us a lot”. Hyatt now has “a much better handle on the people who are coming through those channels”, he said, before the line on age and spending. The transcript gives no figures for those bookings, their value or the travelers’ ages.
Hoplamazian said Hyatt will have a presence “across Gemini, OpenAI” and others and will “remain relatively agnostic”, because the platforms “don’t all present your product the same way”. He called himself “a believer that agentic booking will be a significant portion of our total distribution picture”, and said the shift matters most in resorts, luxury and lifestyle hotels, not in economy brands.
PhocusWire reported the same remarks on 12 November in a roundup of travel executives on AI assistants, and The Transcript posted the quote on X early on 14 November. Hyatt’s transcript lists World of Hyatt at 72 million members, up 19% from a year earlier.
Five disclosures on agent commerce, each on its own base: Visa’s shopping benchmark (runs passed), Forter’s store reach for ChatGPT Instant Checkout, Booking.com’s LLM-referral share of room nights (a ceiling), Uber’s agent rides and Eats orders in Halloween week, and Hyatt’s CEO, who gave no figure. 57.2% and 0.00016% are our arithmetic. Not a funnel.
The Same Gap in Four Other Disclosures
Uber’s chief financial officer said on its 4 November call that of about 320 million trips and orders in Halloween week, around 500 Uber Eats orders and four rides came through third-party AI agents. Our piece on the Uber and Amazon figures has the details. That is about 1.6 per million, by our arithmetic.
Booking Holdings’ chief financial officer Ewout Steenbergen said on 3 November: “LLM referrals have accelerated over the last several weeks but still account for significantly less than 1% of Booking.com’s room nights”. Chief executive Glenn Fogel said AI travel discovery “doesn’t end up with a lot of bookings”. Booking’s own assistants converted at about twice the rate of non-users, as our Booking piece sets out.
Visa’s benchmark is a controlled test, not live orders. Visa put 22 AI models from 12 labs through 15 shopping challenges; they failed 42.8% of runs, and in 54% of the failed runs the AI still made a purchase, as our Visa benchmark piece reports. That leaves 57.2% of runs passed, by our arithmetic.
On the merchant side, Forter’s agent order data for 1 August to 7 October shows ChatGPT Instant Checkout bought at 24% of the stores that received any agentic order, so 76% of those stores had none, by our arithmetic. OpenAI’s browsing agent, ChatGPT Agent, bought at 45%, and Meta’s Muse at 74%. Forter’s explanation: “They can shop anywhere, while the protocols can only reach those stores that have integrated them”.
Share of stores on Forter’s network with any agentic order that each agent bought at, 1 August to 7 October 2026, as published in Forter’s blog on 2 November. Muse, Instinct, ChatGPT Agent and Grok drive stores’ own checkout pages; ChatGPT Instant Checkout and Google use commerce protocols.
What OpenAI Says ChatGPT Wallet Addresses
OpenAI’s post on ChatGPT Wallet says “many of them turn to it to research products, compare options, and decide what to buy”, and that Wallet gives agents “a way to pay that businesses can accept with their existing systems”. The payment network issues a token in place of the card number, and the user approves every purchase with a security check such as Face ID. It starts with ChatGPT Pro subscribers in the US with eligible US-issued cards, as our launch piece and explainer detail.
OpenAI names the failure points itself. “Wallet’s payment tokens work with existing checkout flows, but risk systems may still flag legitimate purchases made by agents, leading to declines or cancellations”, OpenAI’s post says in its section for businesses, adding: “Some sites may unintentionally block purchases through ChatGPT Wallet or other agentic payment products”.
The same day, Sierra released version 1.0 of its Personal Agent Protocol, with Shopify enabling it for every merchant and more than 100 design partners, as our Poppy piece covers. Meta’s Muse arrived in the Chrome Web Store showing 126 users against 22,000,000 for Anthropic’s Claude extension, per our listing read. And Amazon’s robots.txt now bars 99 named crawlers and AI agents, including ChatGPT-User, by our count of the file.
The Structural Read
As we read OpenAI’s post, ChatGPT Wallet is aimed at the last step: OpenAI says it is “now helping people complete purchases”, and lists blocked sites and risk-system declines as the frictions businesses need to fix.
As we read Forter’s data, reach depends on how the agent buys: OpenAI’s browsing agent bought at 45% of stores with agentic orders against 24% for its checkout protocol, and OpenAI says Wallet’s tokens work with existing checkout flows.
Mark Hoplamazian, Hyatt chief executive, Q3 2026 earnings call, 29 October 2026 (The Motley Fool transcript)
“Our learnings from being very early in representing Hyatt directly in the apps that you can find through ChatGPT’s platform have taught us a lot.”
Three Implications
FOR HOTELS AND TRAVEL SELLERS Hyatt’s CEO said the shift matters most in resorts, luxury and lifestyle hotels and that Hyatt will stay “relatively agnostic” across platforms; Booking said its own assistants converted at about twice the rate of non-users.
FOR MERCHANTS OpenAI’s post asks businesses to make sure agents are not blocked and to work with payment and risk providers to reduce declines; our Amazon robots.txt count shows the opposite policy at one large retailer.
FOR READERS OF AGENT METRICS The figures here have different bases, and Hyatt gave no number; read each against its own denominator, as our Booking piece does for room nights.
The framework says of what a technology can do and what users can actually do with it: “They are not the same curve, and the gap between them — what the platform makes possible versus what the product surfaces — has a name.”
As we read it, agent commerce shows that gap at checkout: agents that can browse and compare reach the point of purchase, while the completed bookings Hyatt and Uber describe stay small. The framework does not discuss Hyatt, ChatGPT Wallet or agent bookings.
What Is Not Established
The Hyatt quotes come from The Motley Fool’s transcript, which marks one word on the call as indiscernible; we did not hear the webcast. Hyatt’s CEO gave no booking count, spending figure or age data for ChatGPT users on the call, in the transcript we read. Uber’s figures come from a machine transcript of its call, as our earlier piece notes.
The five figures in our first chart have different bases: test runs, stores, room nights, trips and orders, and a description with no number. They are not a funnel and should not be read as one. Forter’s data covers its own merchant network only, and no completion data for ChatGPT Wallet has been published, since it began rolling out on 13 November. We did not contact Hyatt, OpenAI, Uber, Booking or Forter for this piece.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Hyatt’s CEO says ChatGPT brings younger travelers who spend more, but that bookings completed there are “still very low”. Uber’s 0.00016% and Booking’s under 1% put numbers on that gap, and OpenAI’s own post lists blocked sites and declined payments among the frictions. OpenAI says it will “learn from early use and expand availability”, starting from US Pro subscribers; our Forter piece has the store-reach baseline.
A note on sourcing. Hyatt’s CEO is quoted from The Motley Fool’s transcript of Hyatt’s third-quarter call of 29 October 2026, which we read in full on 14 November; PhocusWire’s 12 November roundup reports the same remarks. OpenAI’s ChatGPT Wallet post and Forter’s agent order data blog were read in full on 13 November. The Uber, Booking, Visa, Muse, Sierra and Amazon figures come from our own earlier pieces, each linked, which cite their primaries. Shares, rates and ratios are our arithmetic. We did not contact Hyatt, OpenAI, Uber, Booking or Forter for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
