I was in Melbourne yesterday, presenting to a group of Victorian hotel General Managers at their quarterly conference. The session was called The Invisible Hotel, a look at how AI search is already deciding which hotels get seen, and which don’t.

I want to share what came out of that room.

Not because it confirms everything we’ve been saying at OmniHyper, but because the questions those GMs asked, and the genuine hunger they had for honest answers, told me something important about where this industry is right now.

These weren’t people hearing about AI search for the first time. In the last six months, something has shifted in the awareness of senior hotel leaders. They’d seen the data. They’d felt something changing in their own numbers. They came to the session wanting to understand why.

We opened with what’s actually happening to search behaviour in Melbourne. Not predictions, but real data. Keywords like “Hotels in Melbourne CBD” are down 70% year-on-year. “5-star hotel Melbourne” down 76%. “Melbourne hotel” down 64%. The searches people used to find your hotel are disappearing, not because fewer people want to stay in Melbourne, but because they’ve stopped searching that way.

Guests aren’t searching less. They’re searching differently.

And then we showed them a real example: a Perth mid-scale hotel running solid traditional SEO. Rankings were up 57% over three months. By every old metric, it was winning. But web direct bookings had fallen 56% year-on-year, and web direct revenue was down nearly 49%.

The strategy was working. The guests weren’t coming.

That’s the invisible hotel problem. Your rankings can be growing while your direct revenue quietly erodes, because the channel guests are using to discover hotels has already moved on.

One point that generated real discussion: being recommended by AI isn’t the same as capturing the booking.

When a guest asks Google AI Mode or ChatGPT to recommend a business hotel in Melbourne, your property might well be the answer. But if the citation that appears next to your name points to Booking.com or Expedia, they may still take the commission. Being the answer and being the destination are two different things.

AI Overviews are now present in 55% of searches. 60% of searches end with zero clicks. Click-through on position one drops 35% when AI Overviews appear. Hotels that aren’t visible in AI search risk becoming more dependent on OTAs, not less, even as their traditional rankings hold.

We could have gone twice as long. The questions kept coming, and if anything they intensified as the session went on. That alone tells you where the commercial urgency is.

Four stood out.

Hotels have been competing against OTA budgets in traditional SEO for years, and the ones that did it well understood something important: you don’t beat them at scale, you beat them at specificity. The real prize has always been discretionary demand, the traveller who hasn’t decided where to stay yet, and that is where specificity beats scale. The same principle carries into AI search. At an individual property level, you have the ability to own your local story in a way no aggregator can. Your proximity to specific venues, your neighbourhood, your unique amenity mix, your relationship with your local market. AI search surfaces this kind of specific, authoritative content. OTAs are broad by design. A well-optimised individual property can win at the local level precisely because it can be more specific, more credible, and more relevant than anything sitting above it in the channel hierarchy.

The way I think about it: responsibility tracks with specificity. Brand-level strategy handles national and international visibility, and it should. But by definition, that content has to be broad. It talks about the group, the network, the promise at scale. It doesn’t, and can’t, tell the story of your conference facilities, your rooftop bar, your proximity to the exhibition centre, or your spa. Those details matter enormously in AI search, where the recommendation is built on matching a guest’s specific need to a specific property. That work has to live at the property level and be owned there. Both layers are necessary. Neither replaces the other.

The signals are already in your data. Pull your Google Search Console and look at impression and click trends over the last 12 months. Check your Google Analytics for changes in organic traffic patterns. Then look at your direct booking conversion and direct revenue trajectory alongside it. If your traditional search metrics are holding or even improving, but your direct bookings are softening and your OTA mix is creeping up, you are likely already experiencing the shift. The Perth example isn’t an outlier. It’s a pattern showing up in markets across Australia.

The question I’d ask back is: where is the most discretionary demand currently occurring, and which channel gives you the most leverage over that demand right now? The honest answer is that for many properties, multiple levers still matter and budget allows you to pull more than one. But if you’re working with a constrained budget and the data is telling you that organic discovery is shifting away from the channels you’re investing in, you need to follow the demand. Guest discovery is moving. Your investment strategy should move with it, and it shouldn’t have to wait for next year’s budget cycle to do so.

These are commercially sharp, accountable people. They’re not resistant to change. They’re trying to make good decisions with real budgets and real ownership conversations to manage.

What struck me is how quickly the awareness gap is closing. The GMs in that room weren’t waiting to be convinced that AI search matters. They were trying to understand it well enough to act, and to be confident explaining it upward.

As David Bark from Mulpha put it at AHICE earlier this year: “If you’re planning AI search in your 2027 budget cycle, it’s too late. You need to be dynamic in 2026.”

That quote landed in the room. Because it’s not a prediction about the future. It’s a description of where things already are.

There were far more questions in that room than time allowed, which is the best outcome a session like this can have. A good place to start is simply testing it yourself, open ChatGPT or Google AI Mode, ask it to recommend a hotel in your market, and see whether your property appears and what source is being cited next to your name. If you’re a GM, commercial director, or leading a team that’s working through where AI search fits in your strategy, I’m genuinely happy to spend some time on it with you.

Nick brings 15+ years of experience working with hundreds of hotels worldwide, helping drive performance and direct revenue. As Partnerships Director at OmniHyper, he’s known for his strategic and commercially focused approach.