AI is travel’s new front door. Airlines are standing outside

AI is becoming travel’s new front door. Most operators are lagging behind, but no airline appears in the research below.
McKinsey & Company and Skift published a joint report on 23 September, “Winning Hearts in an Age of Infinite Travel Choices” (a survey of 1,000+ US travellers, plus a diary study tracking real booking journeys), timed to land at Skift Global Forum. It puts hard numbers behind something the aviation marketing world has mostly been chasing anecdotally this month: AI tools now account for a quarter of initial trip-inspiration sessions, level with classic search engines, and the small slice of bookings arriving via LLM traffic already converts at meaningfully higher rates than other channels.

Being surfaced isn’t the same as being booked, though. IHG Hotels & Resorts‘ Heather Balsley calls AI “a new front door to travel planning and discovery”, one that can narrow a traveller straight down to three or four properties instead of three hundred, eliminating the rest before anyone opens a supplier site. Spain’s tourist board learned this the hard way: ask an AI model about castles in Europe and Spain barely appears, despite having more of them than almost anywhere, because tour operators package German castles as bookable products and Spain’s largely aren’t. The fix is getting into the upstream sources the models actually draw on.

The most striking line comes from Hyatt‘s own head of AI and data analytics, Pat Nestor: Hyatt.com is being restructured so its content “can be discovered, parsed, and surfaced by AI-powered search tools”, where SEO is evolving into GEO. The very next day, at the close of the same Skift Global Forum, Hyatt CEO Mark Hoplamazian said it again on stage, in his own words: making Hyatt legible to AI agents is now as important as staying visible to traditional travel advisors.

Five cases have crossed our desk in five weeks on this exact theme: Japan Airlines, Icelandair, IHG, Wyndham, Hyatt. Every one is hospitality or full-service, not a single low-cost or network carrier.
Lufthansa Innovation Hub‘s TNMT quantifies that gap from the aviation side. Citing Adobe research on “AI Citation Readability”, it found airline booking pages score just 37% machine comprehension, against 57-71% for hotel and car rental sites, dead last of any travel vertical. It’s pretty clear that, right now, aviation is the part of travel AI understands least.