AI IS PICKING YOUR NEXT GUEST BEFORE THEY EVER SEE YOUR WEBSITE
Ask a traveller how they planned their last trip and increasingly they didn’t start with Google. They typed a question straight into ChatGPT. That shift stopped being theoretical this month with Google confirming it has quietly begun testing agentic hotel bookings with US users, letting an AI assistant search, compare and book a room with minimal human clicking. For pub, club and hotel groups running accommodation, this isn’t a future problem to plan for later: it’s already live.
The numbers behind the shift
The scale is bigger than most operators realise. Research from Curacity and Cornell found 94 per cent of hotels are currently invisible in AI search results, while SparkToro’s clickstream analysis shows 68 per cent of US Google searches now end without a single website click. This is up from 60 per cent just two years ago, driven largely by AI Overviews cutting click-through by roughly 60 per cent. Phocuswright calls it the fastest traveller-behaviour shift it has tracked in a decade, with 56 per cent of US leisure travellers now using AI chatbots to help plan trips and adoption is even higher regionally, with Booking.com reporting 95 per cent of Asia-Pacific travellers want AI involved at every stage of their journey.
An Australian fix for an Australian-sized problem
This gap has already led to the emergence of a homegrown solution.
Sydney-founded Visaible.ai, launched by Sam Rich late last year, has raised around A$1 million to help hotels improve their visibility and recommendations across AI platforms including ChatGPT, Claude, Gemini and Perplexity. The company has already launched its service across Australia, Thailand, Singapore and Vietnam.
Its modelling estimates that a standard 150-room property could currently be losing around $30,000 (AUD) a month in missed direct-booking revenue as a result of poor AI visibility, with that figure expected to exceed $42,000AUD a month by 2027.
Rich says hoteliers increasingly recognise AI as a new direct-booking channel and are seeking cost-effective, automated solutions that can capture this demand while allowing hotel teams to focus on the many other operational priorities competing for their attention.
Who actually owns this?
Even major international hotel groups are still determining who should take responsibility for AI visibility internally. At Outrigger Hospitality Group, the issue has prompted consideration of who should ultimately own visibility across AI platforms. At Remington Hotels, the view is that responsibility extends across the website, technical infrastructure, content, reviews and the overall guest experience, making it a genuinely cross-functional function rather than the sole responsibility of marketing.
For smaller Australian pub and club groups without dedicated digital teams, this lack of clear ownership could present an even greater challenge. As AI becomes an increasingly important pathway for consumers researching and choosing where to stay, venues will need to establish who is responsible for managing their visibility across these platforms.
Why it matters to the bottom line
A booking routed through an OTA typically costs a venue 15 to 25 per cent in commission. A booking that arrives because an AI chatbot recommended the property directly keeps that margin in-house, so AI visibility isn’t a communications nicety, it’s a line item on the P&L.
What venues can do now
Make sure room types, amenities, location and pricing are accurate and consistent everywhere online, keep reviews current, and treat AI visibility the way you’d treat SEO with a long-tail investment rather than a campaign with instant ROI. Google’s live test is a signal that every pub, club and hotel group with rooms to sell should be asking now whether they’d actually show up if a guest asked an AI where to stay tonight.
Sources
Google Confirms Agentic Hotel Booking Is Now in Testing – Skift
When a Traveller Goes to ChatGPT, Is Your Hotel Recommended? – The Hotel Conversation
AI Is Deciding Which Hotels Get Considered – Skift
v





