AI-made bookings: what it means for your hotel or restaurant

8 out of 10 travelers worldwide now use artificial intelligence to plan or book their accommodation, and in Spain that figure rises to 90% among Gen Z, according to SiteMinder’s Changing Traveller Report 2026. The same study explains why this is no longer a distant trend for a mid-sized hotel or restaurant: SiteMinder has expanded its Demand Plus and Channels Plus solutions to connect 53,000 hotels across 150 countries directly with AI environments like ChatGPT and Claude, and Anthropic has shipped connectors that let Claude search, compare and prepare accommodation bookings on Booking.com, Tripadvisor and Viator without leaving the conversation. Travelers no longer search “hotels in Barcelona” on Google — they ask an assistant, and that assistant decides which options to show them.

Why this isn’t just a big-chain problem

The underlying shift has a name: agentic commerce. Google has partnered with PayPal on agent-managed payments, Mastercard has published tooling for AI agents to act on a buyer’s behalf, and OpenAI has pushed instant checkout inside ChatGPT. In hospitality, that translates into something very concrete: the AI agent doesn’t just recommend anymore — increasingly, it also completes the booking. If your hotel or restaurant isn’t well represented in those channels, the agent recommends the competitor down the street, not because they’re better but because their data is easier to reach.

BCG’s own report, AI-First Hotels, frames it as a leadership-level decision, not an isolated digital marketing project. Its central point is control over first-party guest data — what BCG calls the guest’s “DNA” — to adjust pricing, personalize communication and win back bookings currently lost to OTAs. Chains already working this way report improvements of over 10% in RevPAR (revenue per available room). The mechanism is the same for an independent hotel or a restaurant with meaningful online-reservation volume: the more of your own data you control instead of an intermediary, the more room you have for an AI agent to recommend you with the right information.

What actually changes day to day

• Your business listing is no longer just for the traveler — it’s for the agent representing them. Hours, real availability, cancellation policy, allergens or room types need to sit in a format a system can read and verify, not just a nice photo or a PDF menu.
• Direct booking now competes with “booking made by the agent.” If the only easy channel to book is an OTA, the AI agent will book there and your business pays the usual commission. A clear direct-booking site — connected to these AI environments where it makes sense — is what lets you capture that demand without an intermediary.
• AI-handled customer service stops being a nice-to-have. Confirmations, booking changes and FAQs handled by a conversational agent free up front-of-house or reception staff for the interactions that genuinely need a person: arrival, incidents, upselling.

How to start without a chain-sized budget

You don’t need an international chain’s resources. The sequence we recommend for an independent hotel, a small restaurant group or a holiday rental is:

1. Audit what information about your business already circulates in AI channels — Google profiles, OTAs, your own site — and fix anything outdated or incomplete, the same approach we outlined in what your site needs for AI to recommend it.
2. Put availability and pricing somewhere you own and update in real time, instead of relying only on a listing managed by a third party.
3. Automate the repetitive tasks first — confirmations, reminders, FAQs — before considering more complex integrations with purchasing agents.
4. Track how much direct booking gains ground against bookings via intermediaries, month over month, to know if the effort is paying off.

It’s the same approach we apply to any applied AI project for the hospitality sector: start with the process causing the most friction today, built on your own code and infrastructure, without depending on a single platform that could change its rules tomorrow.

The takeaway

SiteMinder’s finding — 8 in 10 travelers using AI to book — and Claude’s connectors to Booking.com aren’t a technology curiosity: they’re a sign that bookings no longer always pass through a search engine or a person, but through an agent deciding based on the information it finds. A hotel or restaurant that puts that information in order and makes direct booking easy enters that conversation on equal footing; the one that doesn’t keeps paying commission for something it could solve with well-structured, first-party data.

Want to check how your hotel or restaurant shows up in the channels these AI agents already use? Get in touch: the first consultation is free and we reply within 24 hours.