AI search is taking clicks from hospitality websites
Nearly seven in ten UK Google searches now end without a click. How groups keep venue facts accurate so guests, and AI tools, still find them.
Guests are still looking up your pubs, bars and restaurants. More of them now get the answer from an AI tool, without opening your website.
The click is no longer the visit
Nearly seven in ten UK Google searches end without a click. SparkToro's analysis of Similarweb data, published in June 2026, put the share at 69.5% between January and April 2026. People read an AI Overview, or ask ChatGPT, and move on.
Hospitality has felt that shift in organic traffic. Using Ahrefs data, Tank's Google AI search shift report (October 2025) found that UK hospitality organic traffic fell an estimated 6.7% in the year after Google's AI search features arrived, after 47.9% growth the year before. Of the 16 UK sectors in the study, hospitality was the hardest hit. That estimate is Tank's, not a figure 24social has measured. Many groups will have seen something similar in their own analytics: fewer sessions from search, even when the venues themselves are as busy as ever.
The visits have moved, not vanished
People have not stopped planning a night out. They have changed where they check the details.
In October 2026, Zonal and NIQ surveyed more than 5,000 British consumers. They found that 19% already use AI tools to plan a visit to a pub, bar or restaurant, and 50% are considering it. Of those using AI, 43% look up prices, 38% look up menus and 35% look up opening times.
A year earlier, CGA by NIQ and Reputation (September 2025) found that 26% of consumers in Britain use AI tools such as ChatGPT to learn about a venue. That sits on a par with Google Maps, at 27%.
Prices, menus, opening times and what a venue is like: those are exactly the facts a venue website used to answer on its own. If an assistant is now reading them out, the page still has to be the place those facts are correct.
What we publish for each venue
24social Enterprise keeps brand pages and a page for each venue. Hours, menus, events and offers sit in one editor and are published to the sites that need them, so a change made once is the change guests see.
Each venue page outputs schema.org structured data for the place type, the address, opening hours, facilities and contact details. A machine can read the same facts a person sees on the page.
Every site also serves llms.txt and llms-full.txt, in the llmstxt.org format, including a file for each venue. The files are refreshed regularly from published content, and you can add an optional description written for AI tools. Canonical URLs and sitemaps mark the right page as the source of truth, so a copied or filtered address does not stand in for the venue page. The steps are in our support guides: llms.txt for your website and set a canonical URL.
Publishing clear, consistent facts gives an assistant something accurate to work from. It does not decide whether that assistant cites you, or where you appear in search.
A short checklist for groups
- Keep hours, menus and offers accurate in one place, and publish from there. A special hour that lives only in someone's inbox will not reach the page an assistant reads.
- Give each venue a page in its own right, with its address, hours, facilities and contact details, rather than one list that tries to cover the whole estate.
- Ask the AI tools your guests actually use what they say about a few of your venues. Where the hours, menu or offer is wrong, fix it at the source.
- Keep reviews and listings current. Guests compare what an assistant says with Google, Maps and the sites they already trust.
If you want to see how this works across a multi-site estate, talk to us or read more about Enterprise.
