Sorino

← All guides

A Sorino report

The AI Fit Gap

Hospitality has the lowest AI adoption of any UK sector, and most who tried it say it barely helped. What that actually means for hotels, restaurants and bars

Sorino · Hotels, restaurants and bars · 2026

Foreword from the founders

I keep hearing the same line from hotel owners, restaurant managers and bar operators: "we tried an AI thing, it didn't really do much, so we stopped."

That usually means the tool did not fit the way the venue actually works.

Hospitality has the lowest AI adoption of any sector in the UK right now, and most of the businesses that did try something say it made little or no difference. Read that as proof AI has nothing to offer this industry and you will draw the wrong conclusion. Read it properly and it says something much more useful: many off-the-shelf tools are built for desk work, drafting an email, summarising a document, not for a venue where the phone, the DMs, the emails and the walk-ins never stop, and where every one of them is a real guest waiting for a real answer.

This report sets out the numbers, what they actually mean for a hotel, a restaurant and a bar, and what a piece of AI has to do differently to be worth having in a venue at all. No hype, no made-up results, just what we found and what we would tell a friend running a venue who asked us straight.

Chanel, founder, Sorino

The headline numbers

18% AI adoption in hospitality and leisure, one of the lowest of any UK sector measured, behind IT and telecoms (56%) and marketing (53%). Real estate (11%) and transport (15%) are the only sectors lower.

31% of UK SMEs currently use AI at all, with another 15% planning to. Most businesses, across every sector, are still on the sidelines.

60% of hospitality SMEs say AI has had little or no positive impact on their business so far, even counting the ones who have already tried it.

Sources · YouGov B2B Omnibus survey, August 2025. The Caterer, 2026.

Put those three together and the honest reading is that most of the hospitality businesses who tried something found it did not fit how a venue actually runs, so adoption stalled and results disappointed. That is a fit gap. And fit gaps can be closed.

Why generic AI stalls in a venue, and admin AI does not

Most AI tools sold to small businesses are built for the same handful of jobs: drafting an email, summarising a document, answering a question from a knowledge base someone uploaded once. Useful work, but back-office work. It happens on someone's own time, at a desk, when they choose to open the tool.

A hotel reception, a restaurant floor or a bar on a Friday night does not work like that. The phone rings while a guest is checking in. A booking enquiry lands on Instagram while the kitchen is mid-service. An email asking about a dietary requirement arrives at 11pm, after the last person who could answer it went home. None of it waits for someone to be free, and every one of those moments is a guest deciding whether to book, or whether to try somewhere else instead.

That is the gap a generic AI tool cannot close. It was never built to sit on the phone line, the inbox, the social DMs and the chat widget all day and night, all grounded in the same facts about your rooms, your menu, your policies, your availability. It was built to help one person do one task faster. A venue does not have one task. It has every channel, open all the time, and a guest on the other end of each one.

What "fit" actually looks like

The businesses getting real value from AI in hospitality are the ones whose AI does the specific job a venue needs done: answer every enquiry, on every channel, correctly, at any hour, using the venue's own information, and hand a person the booking or the query the moment it needs a human touch.

That is four channels working from one source of truth, not four separate logins nobody has time to check. It is coverage for the hours nobody is on the desk, not a tool that only helps during office hours. It is built for a hotel, a restaurant or a bar specifically, not adapted from a generic customer-service script that has never heard of a room-only rate or a tasting menu.

Sorino's agents work across four channels, voice, email, chat and social, all pulling from one knowledge base built from your venue's own rooms, menu, hours and policies. Live in two to six weeks. Plans from £299 a month, first month free.

1. Hotels: the reception that never actually stops

An 18% sector-wide adoption figure hides a specific hotel problem: reception is one desk covering every channel a guest might use, and it is not staffed for all of them at once. A late booking enquiry that lands by email while the guest in front of the desk is checking in does not get answered in the moment it arrives. It waits, and a guest choosing between your hotel and the next one does not always wait back.

The businesses in that 60% who say AI has not moved the needle mostly tried a chatbot bolted onto their website, answering only the questions someone thought to programme in, blind to the phone and the inbox next to it. An agent that actually helps a hotel needs to be on the phone line, the email, the chat widget and the socials at once, working from the same room information reception already has, so a guest gets the same accurate answer whichever way they reach out, at 3pm or 3am.

hotels.sorino.ai · Plans from £299 a month, first month free.

2. Restaurants: the booking that arrives mid-service

A restaurant's busiest moment for taking new bookings is often the exact moment nobody in the room can pick up the phone. The kitchen is firing, the floor is full, and a call about a table for six on Saturday goes to voicemail because answering it properly would mean stepping away from a guest already sitting in front of you.

That is precisely the always-on, multi-channel gap the 60% figure is describing. A generic tool answering website FAQs does nothing for a call at 8pm on a Friday. What actually helps is an agent that takes the call, the DM and the email the same way a well-briefed member of front of house would, checking real availability and writing the booking straight into the diary, so the team keeps their attention on the table in front of them instead of the phone beside it.

restaurants.sorino.ai · Plans from £299 a month, first month free.

3. Bars: the enquiry that lands after last orders

Bars live on social. A DM asking about a private hire, a table for a birthday, or whether the venue is open on a bank holiday can land at any hour, and a bar team working the floor has no spare hand for Instagram at 11pm. Left unanswered until the next morning, that enquiry has often already gone to whichever venue answered first.

This is the same fit gap again: a tool that only handles the website chat widget misses where bar enquiries actually happen. An agent built for a bar needs to be watching the DMs and the phone as closely as the till, replying with real information about the venue, the hours and what is on, so nothing sits unanswered until it is too late to matter.

bars.sorino.ai · Plans from £299 a month, first month free.

What this means if you run a venue

If you tried an AI tool once and it did not change much, that experience was real and there is no reason to pretend otherwise. But it says more about the tool than about whether AI has anything to offer a venue like yours. The better question is whether that AI was actually built to run a venue's front door.

Coverage across every channel a guest uses. Answers grounded in your own information, not a generic script. Live in weeks, not a project that sits half-finished. That is what closes the fit gap, and it is what we build.

Sources

Sorino · the AI operations and guest-experience layer for hospitality. Hotels, restaurants and bars. One knowledge base, every channel answered, on the systems you already run.