Foreword from the founders
Hey, I'm Chanel, a co-founder of Sorino. Over the last few years our team has watched the same thing happen in hotel lobbies, restaurant passes and bar service stations up and down the country: the phone rings out while everyone's heads-down doing the actual job. A booking walks. A DM sits unread until Tuesday. Nobody did anything wrong, there just weren't enough hands.
So we put AI on the channels that were going unanswered, reading from one knowledge base of the venue's own menus, hours and policies. Not to replace anyone, to cover the hours your people can't pick up. This report is what we're seeing across three parts of UK hospitality, hotels, restaurants and bars, told plainly, with real numbers where solid data exists and honest plain-English observation where it does not.
What this report covers
UK hospitality is not one market, it's three that share the same problem. A guest tries to reach you, on their channel, at their moment, and too often no one is free to answer. This report looks at that shared problem across:
- Hotels: mid-market, multi-property groups juggling six channels and no single view.
- Restaurants: owner-led fine dining where the person answering the phone is also on the pass.
- Bars and food-led venues: booking-heavy, DM-first operators who lose the most at peak.
And it looks at what AI, used with discipline, is doing about it.
1. The squeeze every operator already feels
Hospitality is not a small corner of the economy. It employs around 3.5 million people, which makes it the UK's third-largest employer, it accounts for up to 12% of the workforce in some regions, and it contributes about £93 billion a year. When it is under strain, a lot of the country feels it.
And right now it is under strain. Two things have happened at once. Staffing got harder, and guests stopped waiting. Hospitality is caught in the middle.
On the supply side, the pressure is real and measured:
69% of hospitality businesses are operating at or below 85% of the capacity they need.
One in three are now operating at a loss.
70% said the April 2025 cost rises could force them to reduce the number of people they employ.
Source · UKHospitality Quarterly Members' Survey, 2025.
Read those together and you get the shape of the problem: most venues cannot fully staff the hours they are open, a third are losing money, and when cost pressure bites the first thing under threat is the headcount that answers the phones and covers the floor.
On the demand side, patience has collapsed. A guest with a question wants the answer now, not after the weekend, and not after they've already tried the place down the road. An instant reply is no longer a nice touch, it's the baseline expectation, on whatever channel the guest happens to use. The gap between "we're short of hands" and "answer me now" is exactly where bookings, covers and reputation leak away.
The cost of the gap. A missed call at a hotel is a room sold by one of the big booking sites at a commission instead of direct. A missed call at a restaurant is a table that stays empty on a Saturday. A message left unread at a bar is a party of twelve who booked the place down the road. None of it shows up as a line in the accounts, which is exactly why it goes unfixed for years.
2. Hotels: six channels, no single view
The mid-market hotel group is where fragmentation hurts most. A group of eight to fifteen properties typically runs a property system, a guest database, a booking engine and an accounting system, and none of them talk to each other. Guest conversations scatter across phone, web chat, WhatsApp, Booking.com, email and social, with no one view.
What operators tell us:
- "Guest conversations are scattered across six channels and nothing has a single view."
- "We're leaking direct bookings to the big booking sites, and paying commission on them."
- "After-hours enquiries just go to voicemail."
- "We've no consistent view across the properties, every site reports differently."
Where AI lands. Not as another dashboard to check, but as the layer that answers across every channel in one consistent voice, books from live availability, and gives operators one view of what happened, per property and across the group. The buyer that matters here is the operations director, and the pitch is unification and outcomes, not "a chatbot": occupancy lift, labour saved, direct bookings kept off the big booking sites. Every enquiry that gets answered and booked direct is a room that doesn't carry a booking-site commission, and operators feel that difference straight away.
3. Restaurants: the person answering the phone is also on the pass
UK fine dining is overwhelmingly independent and owner-led. There is no head-office marketing team. The buyer is usually the chef-patron, plating food or working the pass, who is also, somehow, meant to answer the phone, reply to the private-dining enquiry and keep the Instagram warm.
Two things make this vertical distinct:
- Allergens are safety-critical. The single biggest fear operators voice about AI is an agent getting a dietary requirement wrong. So the knowledge base has to hold the full allergen and dietary matrix per dish, and the agent has to answer from it, on the record, with a clean human hand-off the moment anything is ambiguous. This is the non-negotiable.
- Private dining is where the money leaks. A slow reply to a large-party or event enquiry is a booking that goes elsewhere. Email that reads by intent, not in order, and replies fast with the right detail, converts planners who would otherwise have moved on.
What operators tell us:
- "It has to sound like us, not like a robot."
- "I'm terrified of an agent getting allergens wrong."
- "By the time we reply to the enquiry, they've booked somewhere else."
Where AI lands. Coverage and support for the hours your people can't pick up, never a replacement for the maître d'. The reservation gets taken while the team works the floor, the dietary detail gets captured on the call and briefed to the kitchen, and the private-dining enquiry gets a fast, on-brand reply. A large share of these enquiries land in the evenings and on days the venue is closed, exactly when there's no one free to answer, so the hours the AI covers are the hours that used to lose the booking.
4. Bars and food-led venues: the phone never stops on a Friday
The booking-heavy operator, a casual-dining, bar-restaurant or premium-pub group, feels the leak most sharply at peak. Friday night is exactly when the floor cannot answer the phone, the inbox or the DMs, and it's exactly when the bookings are coming in.
The defining shift here is the channel. For this audience, the Instagram DM is now the first step in the booking journey, and it's the one most likely to go unread for two days.
What operators tell us:
- "The phone never stops on a Friday, and we miss calls when we're slammed."
- "The DMs just pile up, we see them two days later."
- "Same questions all day: are you open, gluten-free, can I bring the dog."
- "They don't turn up and don't even call."
Where AI lands. Answer the phone and the DMs while the team runs service, in the venue's own voice, from the venue's own menu, and cut no-shows with confirmations and reminders that actually go out. No-shows quietly cost these venues covers every week, and reminders that reliably reach the guest are the simplest way to claw some of them back. The language that works is covers, bookings, no-shows, the floor, the diary, the DMs, not "conversational AI engine".
5. What the AI actually does: four channels, one knowledge base
Across all three verticals the shape is the same: one shared knowledge base (the venue's menus, hours, allergens and policies) feeding four channels so every one answers the same way.
| Channel | What it does | The hour it covers |
|---|---|---|
| Voice | Answers calls 24/7 in the venue's tone, takes and confirms the booking, captures party size and dietary detail, escalates to a human when it should | Mid-service, after hours, the second everyone's hands are full |
| Reads the inbox by intent not order, replies fast to enquiries and events, sends confirmations and deposit requests | The Monday-morning backlog, the private-dining enquiry that can't wait | |
| Chat | A branded website and WhatsApp concierge that answers menu and availability questions and locks in the booking in-flow | The moment a guest is still deciding |
| Social | Answers Instagram and Facebook DMs and mentions before the lead goes cold | The DM that used to sit unread for two days |
| Knowledge Base | The shared brain all four draw on, so every channel sounds like this venue | Always, underneath everything |
The point that matters to a sceptical operator: it is not four bots. It is one source of truth, four front doors, and a clean hand-off to your team the moment a conversation needs a human.
6. What good looks like, and what we will not let AI do
The fastest way to lose a hospitality operator's trust is to over-promise. So the governance is part of the product, not an afterthought.
- Allergens answer from the record, or hand off. No guessing on anything safety-critical.
- It sounds like the venue, because it reads from the venue's own words. Not a generic script.
- A human takes over the second it gets complicated. The AI knows the edge of its competence.
- Everything is on the record. Full transcripts, per-agent permissions, clear escalation rules.
- It works alongside the team, it does not replace the maître d' or the host. Coverage, not redundancy.
This is the section that answers the doubt before the sales call, and it's the one most likely to convert a burned-before operator.
7. Getting started: a fixed-scope pilot, not a leap of faith
Pre product-market-fit, nobody should be asked to bet the business on a new vendor. The route in is a fixed-scope six-week pilot with clear measures, agreed up front:
- First-response time across the channels you choose.
- After-hours coverage, the calls and DMs that used to go unanswered.
- Enquiry capture, the bookings and events that used to leak.
Low risk is what converts the sceptic. You see it working on your own venue, on your own menu, in your own voice, before you commit to anything bigger.
Sources
- 69% of hospitality businesses are operating at or below 85% of the capacity they need. UKHospitality Quarterly Members' Survey, 2025. ukhospitality.org.uk/insight/quarterly-members-survey
- One in three hospitality businesses are now operating at a loss. UKHospitality Quarterly Members' Survey, 2025. ukhospitality.org.uk/insight/quarterly-members-survey
- 70% of businesses said the April 2025 cost rises could force them to reduce the number of people they employ. UKHospitality Quarterly Members' Survey, 2025. ukhospitality.org.uk/insight/quarterly-members-survey
- Hospitality employs around 3.5 million people, is the UK's third-largest employer, accounts for up to 12% of every regional workforce, and contributes about £93 billion a year. UKHospitality, Economic Contribution of Hospitality. ukhospitality.org.uk/insight/economic-contribution-of-hospitality
A note on the "what operators tell us" lines. Those are observed patterns in how operators describe the pain, drawn from the language we hear across hotels, restaurants and bars. They are illustrative, not survey data. This report names no venues, claims no deployment counts, and reports no customer results.
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.