Guests who find your hotel through AI tools like ChatGPT, Perplexity, or Google's AI search are more engaged than guests from almost any other source — including Google search, direct visits, and paid ads.

They're not just landing on your website. They're checking rooms, reading amenity details, looking at your location page. They behave like someone who's already decided you might be the right property and just wants to confirm it.

That's not a guess. That's what the data across our hotel portfolio showed in Q1 2026.

What we actually measured

At LOKAL, we manage marketing for hotels across Southeast Asia — boutique city properties, beachfront resorts, serviced residences, larger destination stays. Different markets, different guests, but the job is largely the same: bring in the right people and drive more direct bookings.

A big part of that work is watching what visitors actually do once they land on a hotel's website. Not just how many showed up, but whether they stayed, what they looked at, and whether they left within seconds of arriving.

Earlier this year, we pulled apart AI-referred traffic across the portfolio — guests arriving from ChatGPT, Perplexity, and Google's AI features — and compared their behavior against every other major source.

The volume is still small. Roughly 2,400 sessions out of 2.5 million in Q1. This is not a "drop everything, this is your new biggest channel" story. Not yet.



But the quality is hard to ignore.

The guests AI sends are different

Out of every 100 visitors arriving from an AI tool, around 85 actually engage — clicking through to rooms, amenities, location pages, or rates. From Google search, that number is 77. From direct traffic, it drops to 55. From paid social, it sits at 47.

The bounce rate follows the same pattern. Only 15 out of 100 AI-referred visitors leave without doing anything. For Google search it's 23, direct traffic 45, paid social 53. AI visitors also browse more — 2.7 pages per session on average, versus 2.1 for Google search, 1.5 for direct, and 1.3 for social.



It's not a clean sweep. Guests from Google search spend slightly more time per session, close to 4 minutes versus about 3 for AI-referred visitors. Organic search remains a strong, high-intent channel, and we're not dismissing it. But on engagement rate, bounce rate, and depth of visit, AI-referred traffic consistently came out ahead.

Two properties made the pattern harder to ignore

A serviced residence in a major metro saw 96.7% of its AI-referred visitors actively engaging with the site, with sessions averaging over 4 minutes. A beachfront hotel in a popular tourist destination hit 97.1% engagement, with sessions running around 3.5 minutes.

The sample sizes were small (around 150 and 175 sessions respectively), and it would be dishonest to claim statistical certainty from numbers that size. But the broader pattern held across the portfolio. Even where individual cohorts were modest, AI-referred guests consistently behaved like high-intent visitors.

To put it in context: in hospitality, 50% engagement is already considered a respectable benchmark. When you're seeing figures in the mid-80s to high-90s, it warrants attention — even on small samples.

Why these guests behave differently

It helps to think about what actually happens before an AI-referred visitor reaches your website.

A Google user searching "best beachfront hotel in Subic" is handed a list of ten links. They're in comparison mode — clicking through a few, skimming each one briefly, moving on. They're still early in the process of deciding.

An AI user is doing something fundamentally different. They're not browsing a list — they're asking for a recommendation. Sometimes something as specific as "Which boutique hotel in Makati is good for couples and has a pool?" The AI returns a short answer, maybe two or three properties. By the time that person clicks through to your website, they're not encountering you for the first time. They already believe you might be the right choice. They're just there to confirm it.

That shift in intent is what the behavioral data reflects. These are guests who are further along in the decision before they ever reach your site.

Why AI picks some hotels over others

This is where it gets practical — and where the opportunity is most clear for property owners and managers.

AI systems don't work well with the kind of language that makes up a lot of hotel marketing. Atmosphere, branding, evocative photography — these work beautifully for human readers. But when a guest asks an AI a specific question about your destination, the model needs actual information it can match against that question.

Take a query like "What's the best beachfront hotel in El Nido for families?" To answer that well, the AI needs to know how close the property is to the beach, what room types are available, whether breakfast is included, and whether the hotel is genuinely set up for families. If those details are missing, vague, or inconsistent across the web, your property becomes difficult to recommend with confidence. The model doesn't approximate — it moves on to whoever provided a clearer answer.

That's been one of the most consistent findings in our work, and it points to three things that actually matter:

Specific, extractable details on your website. The difference between "we're close to the beach" and "3-minute walk to the beach" matters more than it might seem. Same with "family-friendly" versus "interconnecting rooms available, kids' menu at the restaurant, pool with a shallow area." The more concrete and specific your content, the better the AI can match your property to a real guest query.

Consistency across all your listings. AI tools cross-reference. If your website says 42 rooms, your Google Business Profile says 38, and Booking.com shows a different room category structure, that inconsistency signals a problem. Check your GBP, your OTA listings, review platforms, and any directories your property appears in. The basics need to align.

Answering the questions guests actually ask. Most hotel websites are built around general descriptions rather than specific answers. But guests using AI tools want to know things like: Is this hotel good for couples? Can I walk to restaurants from here? Is there an airport transfer? What's the neighbourhood like at night? If your website addresses those questions clearly, the AI has more to work with when deciding whether to recommend you.

What we're doing at LOKAL — and what you can do now

We've started treating AI visibility as its own layer of work. Related to SEO, but not the same thing. That includes restructuring website content around the questions real travellers ask, cleaning up listing consistency across platforms, and what we've been calling prompt-level tracking: monitoring a fixed set of AI queries per property to see whether the hotel gets recommended, how often, and which competitors appear instead.

That last piece has already surfaced gaps we wouldn't have found any other way. One property was performing well for general destination queries but disappearing entirely when someone asked about couples' travel. That's not something Google Analytics would show you — but once you see it, you can act on it.

If you want to start checking this yourself, the method is straightforward: open ChatGPT or Perplexity and ask the kinds of questions your ideal guest would ask. "Best hotel in [your destination] for couples." "Family-friendly hotel near [landmark] with a pool." "Boutique hotel in [your city] with good breakfast." Note whether your property appears. Note who does appear instead. Then look at what information those competitors have on their websites that you don't.



That exercise alone will tell you more than most audits will.

This channel is still small. But the signal is real.

AI-referred traffic remains a fraction of most hotels' total visits. That's true across our portfolio, and it's likely true for yours.

But the guests it delivers are among the highest-quality visitors in our data — more engaged, browsing more deeply, and behaving like people who are close to booking rather than casually exploring options.

The properties appearing in AI results right now aren't necessarily the biggest names or the ones with the largest marketing budgets. They're the ones with clear, specific, consistent information that AI systems can actually use.

For hotels that can't outspend the major chains on paid advertising, that's a meaningful opportunity. The gap between showing up in AI results and not showing up is still largely an information problem — not a budget problem.

That's worth paying attention to.