24% of Holiday Travelers Now Use AI to Plan. Most GEO Tactics Still Fail

Twenty-four percent of holiday travelers planned to use generative AI for trip planning in 2025, up from 16% in 2024 and 8% in 2023, according to Deloitte's 2025 Holiday Travel Survey. That tripling in two years is the hard number behind a growing push to sell restaurants on "Generative Engine Optimization." Much of what is being sold does not work.

The pattern matters to operators for one reason: AI recommendations now connect directly to maps, reviews, transit data and booking paths. Google's 2024 travel features show how. In March 2024, the company began testing AI-generated trip itineraries in Search through its Search Generative Experience, open to U.S. users who opted in via Search Labs. An itinerary could list attractions and restaurants alongside flight and hotel overviews, then export to Gmail, Docs or Maps. Google Maps added curated recommendation lists from publishers and its community in the same period.

A traveler can now move from AI recommendation to directions in minutes, without visiting a restaurant's own website. The restaurant left off the list never gets a click.

What does the research actually support?

The academic foundation for GEO comes from Aggarwal et al. (2024), presented at the 30th ACM SIGKDD Conference. The authors reported visibility gains of up to 40% in benchmark experiments. Those are benchmark results, not commercial outcomes.

A 2026 critical survey of the GEO literature by Martinez found the picture far less encouraging. It confirmed that already-retrieved content can causally shape an AI answer. But it identified no technique with a stable, cross-platform causal effect on organic discoverability, downstream clicks or conversions. Google's own guidance is blunter. It tells site owners that many suggested AEO and GEO tactics are not effective or not supported by how Search actually works.

The gap between vendor promises and verified results is wide. Operators should treat any claim of guaranteed AI placement as unproven.

What should restaurants do instead?

Google's optimization guidance for generative AI features says foundational SEO still applies: unique, helpful content and a clear technical structure. It also states that Google Business Profiles can help local businesses appear in AI responses. The Business Profile guidelines ask businesses to keep addresses and opening hours accurate.

The practical checklist is short and unglamorous:

• Accurate hours, addresses and contact details
• Consistent listings across platforms
• Timely responses to customer reviews
• Helpful content and clean site structure

None of this is new. All of it is measurable. Accuracy of listing data was once administrative housekeeping. It is now part of how travelers discover and evaluate a business through digital intermediaries.

Is discovery the whole game?

No. The opportunity extends past appearing in AI answers. The objective is helping a traveler evaluate a recommendation, work out how to reach the restaurant, and actually visit. An AI mention without a path to the door is worth little.

One proposed framework for AI-driven local hospitality marketing frames this as connecting AI-assisted discovery with location-based decisions. Its author concedes it is a strategic proposition, not an empirically validated conversion model. That caveat deserves emphasis.

For an independent operator, the calculus is straightforward. The Deloitte adoption curve is real and accelerating. The tooling that routes AI suggestions into Maps is deployed. The tactics that guarantee placement are not. Until the evidence changes, the defensible spend is information quality — accurate data, consistent listings, responsive review management — not optimization packages sold on benchmark statistics.

Expect pressure on that position to grow as adoption passes a third of travelers and vendors refine their claims.