How the answers frame Booking.com
This snapshot covers 25 answers from ChatGPT, Claude, Gemini, Perplexity and Grok, measured on 2026-09-28. Booking.com was named in 7 of them. Five came from one question, "Which travel booking sites are the most trustworthy?", where every assistant named it, usually next to Expedia. The other two were a Gemini answer on hotel sustainability and a Grok answer on which company travellers should avoid.
In the trust answers, Booking.com is described as a large, established platform. Gemini credited it as "the best for hotel inventory" and for free-cancellation filters, and grouped it with Expedia and Priceline as major platforms. Perplexity listed it first and pointed to a Which? survey score of 73% among hotel booking sites. Claude described it as a place to find discounted rooms and manage reservations. Grok listed it first as well, but added "criticism on pricing consistency and support in some reviews."
ChatGPT is a partial case. Its visible answer leaned on BBB and Consumer Reports guidance and named Expedia as the trusted example. Booking.com appears only in its drafting notes, as a site that is "often mentioned."
The two other appearances point in different directions. Gemini's sustainability answer named the hotel groups Hilton, Marriott and IHG as leaders, and mentioned Booking.com only for offering sustainable-travel filters. Grok's "avoid" answer, whose first sentence said no single company should be universally avoided, still placed Booking.com first in its list of online travel agencies with frequent complaints.
Among hotel and travel names in this data, Marriott and Hilton were each named in 16 of 25 answers and Hyatt in 15. Booking.com sat below IHG (9) and above Expedia (6) and Airbnb (1). The topic tags on its mentions were mostly Customer Experience (4), with one each for Sustainability & Ethics and Not Substantive.
…appearing most consistently as the top-rated overall option…
What the press is saying
The PeakMetrics news index recorded 4,565 mentions of Booking.com over the last 30 days, with the brand in the headline of 303. That is well below Marriott's 24,774 and Hyatt's 14,355, and below Airbnb's 15,980, which barely appeared in the AI answers.
The recent headlines share a theme. Several describe fake listings: tourists sent to a Wetherspoons pub, a consumer group calling Booking.com "unfit" after a fake 10 Downing Street listing, and an accusation that it ignored fake listings. Another headline describes a guest locked out of a Belfast flat after wrong booking details were sent. One lighter item covers a "Turbulence Test" feature for travel-companion compatibility.
Frequently ranks among the most complained-about in multiple regions…
The gap
The assistants and the press are describing different things. The answers talk about Booking.com as a category default: scale, inventory, filters and cancellation flexibility. The headlines are about specific failures, namely listings that were not real and booking details that were wrong.
The answers do touch the same territory in general terms. Gemini and Grok both gave advice on spotting fake sites and verifying bookings, and Grok cited complaints about customer service. None of the answers tied that advice to the fake-listing stories in the news. In this snapshot, the recent press narrative had no visible presence in the trust answers, except in Grok's cautionary framing.
What to watch
For a Booking.com comms team, the trust question is the one that matters in this snapshot, and the brand was present in all five assistants' answers to it. The framing is not uniform, though. Perplexity called it the most consistent top-rated option. Grok placed it in both its recommended list and its avoid list, and ChatGPT's visible answer did not name it.
The recurring advice to book directly with hotels and airlines appeared in the ChatGPT, Gemini, Perplexity and Grok answers. It is a structural point for any online travel agency, whatever the press cycle.
These are single runs, and live answers vary between runs. The useful next step is to repeat the same benchmark questions and see whether the complaint framing Grok used, or the fake-listing coverage, shows up more widely.