AI in hospitality: what pubs, restaurants, cafés and hotels should use it for
Hospitality uses AI less than almost any other sector, under more cost pressure than most. The opportunity is real, but only if it is judged on staff time, waste and guests, not on the tool.
Hospitality is under more pressure than most of the economy. The Insolvency Service recorded 3,208 company insolvencies in accommodation and food services in England and Wales in the 12 months to September 2026, the third highest of any industry. NIQ’s Hospitality Market Monitor counted 98,319 licensed premises in Great Britain in September 2026, 977 fewer than a year earlier.
Costs have risen on every line. Employer National Insurance went up to 15% in April 2025, on a lower threshold. The National Living Wage rose to £12.71 an hour in April 2026. Temporary business rates relief ended the same month, replaced by permanently lower multipliers for retail, hospitality and leisure properties.
At the same time, the sector is one of the slowest to use AI. That gap is either a missed opportunity or a sensible caution, depending on what the AI is for.
A late adopter, by any measure
The government’s AI adoption research, published in October 2026, found that 32% of UK businesses with five or more employees were using at least one AI technology in early 2026, double the 16% of a year before. Hotel and catering was the lowest of the sectors shown, at 23%. It was also the sector where businesses were most likely to have no plans to use AI at all: 73%, down from 88% in 2025.
Source: DSIT, Artificial intelligence adoption research 2026 (hotel and catering read from Figure 2)
The ONS tells a similar story. In its July 2026 article on AI in UK businesses, around a fifth of accommodation and food businesses reported using AI in June 2026, against nearly three in five in information and communication. The ONS also found that hospitality firms most often adopt AI through free software rather than paid tools.
Where AI is used, it is used lightly. In the DSIT research, just 4% of hotel and catering businesses said more than half their staff used AI. Its case study was a mid-sized hotel using a general chatbot to draft replies to guest reviews, which the manager said saved reception staff a little time.
The intent is there. A survey by NIQ and Sona of leaders across more than 13,000 venues, carried out in summer 2025, found 72% likely to implement AI in the next 12 months, but only 26% felt their business was equipped to do it at scale. Those businesses sell data and workforce software to the sector, so the figures are best read as direction, not measurement.
Why the gap is not simply a failure
There is a reason hospitality has been slow. The Department for Education’s 2023 analysis of the impact of AI on UK jobs found accommodation and food services to be among the industries least exposed to AI and to large language models. Most of the work is physical, social and done in the moment: cooking, serving, cleaning, welcoming. A language model does not pull a pint.
The pressure on people is real, though. The DfE Employer Skills Survey 2024 found hotels and restaurants had the highest density of skills gaps of any sector, with 6.2% of staff not fully proficient, against 4.0% across all sectors. Vacancies have fallen sharply, from 177,000 at the 2022 peak to 70,000 in July to September 2026, below the 98,000 of 2019. Fewer vacancies at a time of closures and rising wage costs is a sign of businesses holding back on hiring, not of a solved problem.
Source: ONS, UK job vacancies, accommodation and food service activities (JP9O), October 2026
What this means: the case for AI in hospitality is not about replacing staff. It is about giving time back to the people the business already has, and taking cost out of the parts of the job that guests never see.
Where it can earn its place
Three areas stand out, each with a measure that matters to the business.
The first is forecasting and waste. WRAP’s last sector-wide estimate, based on 2011 data, found that UK hospitality and food service wasted 920,000 tonnes of food a year at outlets, three quarters of it avoidable. The figure is old, but nothing published since suggests the problem has gone away. Systems that weigh and log waste, then forecast demand, are now common in large kitchens. The vendors report large reductions, Winnow for example says it can halve food waste, but those are vendor claims. The useful test is local: what did this kitchen throw away before, and what does it throw away now?
The second is rotas and back office. In the NIQ and Sona survey, 27% of leaders said they used AI for shift scheduling and 49% said it helped with forecasts and reports. Every hour a manager stops spending on spreadsheets is an hour on the floor. That is measurable, and it is the measure that matters.
The third is pricing and booking. A RoomPriceGenie survey of hotel revenue managers in eight countries, reported in October 2026, found 53% using AI to help set room prices, but only 10% willing to let it make most pricing decisions on its own. In April 2026, Google launched restaurant bookings inside AI Mode in Search in the UK, completed through booking platforms including OpenTable, SevenRooms, ResDiary and TheFork. Guests may increasingly arrive through an AI assistant before they ever see a venue’s own website.
Where it needs care
Guests are open to it, with conditions. Research for Toast, reported in October 2026, found 47% of UK diners said knowing they could easily reach a human was the biggest factor in trusting a restaurant’s AI. Some 17% said allergen and dietary questions should not be automated at all. Access Hospitality’s research found 51% of operators in the UK and Ireland worried about how their data was processed. All three are vendor surveys, but they point the same way.
The law points the same way too. Food businesses must tell customers about any of the 14 listed allergens in their food, and the Food Standards Agency’s guidance is clear that the information must be accurate. A chatbot or menu generator that gets an allergen wrong is the venue’s problem, not the software’s. The Digital Markets, Competition and Consumers Act 2024 banned fake reviews and drip pricing from April 2025, with fines of up to 10% of global turnover. The ICO’s guidance on AI and data protection is under review after the Data (Use and Access) Act. Groups with sites in the EU also face the EU AI Act, now largely in force, with rules on high-risk uses such as recruitment delayed to December 2027.
What we would count
National policy has little to say about hospitality and AI. The AI Opportunities Action Plan and the SME Digital Adoption Taskforce both call for support for small firms. Neither mentions hospitality by name. So for now the case has to be made venue by venue.
We would start with the measure, not the tool. Kilograms of food wasted per cover. Manager hours spent on rotas and ordering. Staff turnover and sickness. Covers per head of staff on a busy night. Complaints, and the time taken to answer them. Measure them for a month before anything changes, then again after, and keep the tools that move them.
The test we apply to our own work is whether anyone’s life changed, and whether we can show it. In hospitality, that is a chef who goes home on time, a manager back on the floor instead of in the office, a guest with an allergy who gets the right answer every time, and a business still open next year. An AI tool that can show any of those has earned its place. One that only shows usage has not.
- The Insolvency Service, Commentary: company insolvency statistics September 2026, October 2026
- NIQ, Hospitality loses 19 licensed premises a week ahead of crunch Christmas, October 2026
- HMRC, Rates and allowances: National Insurance contributions, April 2026
- GOV.UK, National Minimum Wage and National Living Wage rates
- HM Treasury, Budget 2025: retail, hospitality and leisure factsheet, January 2026
- DSIT, AI adoption survey report 2026, October 2026
- ONS, Artificial intelligence in UK businesses: 2023 to 2026, July 2026
- NIQ, AI beds down in hospitality but cost pressures constrain investment, December 2025
- Department for Education, The impact of AI on UK jobs and training, November 2023
- Department for Education, Employer Skills Survey 2024: UK research report, November 2025
- ONS, UK job vacancies: accommodation and food service activities (JP9O), October 2026
- WRAP, Overview of waste in the UK hospitality and food service sector, November 2013
- Winnow (vendor), company website
- Hospitality Net, More than half of hotels use AI to price rooms (RoomPriceGenie, vendor survey), October 2026
- Google, Agentic restaurant bookings in AI Mode in Search in the UK, April 2026
- IT Brief UK, UK diners want AI-driven restaurant deals, Toast finds (vendor survey), October 2026
- Workplace Journal, AI adoption grows in hospitality but data security concerns remain high (Access Hospitality, vendor survey), March 2026
- Food Standards Agency, Allergen guidance for food businesses, updated July 2026
- Competition and Markets Authority, New consumer protection regime comes into force, April 2025
- ICO, Guidance on AI and data protection
- European Commission, AI Act: regulatory framework for AI, updated October 2026
- DSIT, AI Opportunities Action Plan, January 2025
- DBT, SME Digital Adoption Taskforce: final report, July 2025
How we made this. We gather official and industry data, analyse it ourselves and write up what we think it means. AI tools help us find sources and draft; every figure is checked by us against the sources listed above. Where we have calculated a share or a change ourselves, the chart or text says so.
More insights.
The high street is a service: what small towns like Beverley and Boroughbridge do that city centres cannot
Market towns and local centres are not smaller versions of city centres. They do a different job for their communities, and councils and BIDs should judge them on footfall, dwell and spend as outcomes of that job.
ResearchPoor health is an economic measure: what the North East shows
Economic inactivity, productivity and local growth are usually treated as economic problems. In the North East of England, the data says they are health outcomes in disguise.
ResearchLeisure centres are health infrastructure: fund them on health outcomes, not memberships
Public leisure centres prevent illness, support people back from cancer and reach the places with the poorest health. They are still funded and judged as if they were gyms.
Practical ideas, once a month.
Website, growth and marketing tips for small and growing businesses. No spam, and you can unsubscribe at any time.