Key takeaways
- Most UK adults now use AI tools, but general feelings about AI are still mixed, so how you introduce it matters more than whether you do.
- Guests accept automation for quick factual calls and booking requests; they resist it for complaints, allergies and anything emotional.
- Honesty, speed and an easy route to a person are what make an automated answer acceptable.
- Measure acceptance in your own venue with a few simple numbers rather than relying on national surveys.
The short answer: it depends on the call
Ask a pub landlord whether regulars will put up with an AI on the phone and you will hear a firm no. Ask the same landlord how often the phone rings out on a Friday night and the answer is usually too often. Both are true, and the gap between them is where the real question sits. Guests do not judge automated answering in general; they judge it call by call, against the alternative they would otherwise have had.
For a guest who wants to know whether the kitchen is still serving, an instant accurate answer beats a phone that rings out. For a guest who is ringing about a bad experience last night, an automated voice can feel like being brushed off. Planning around that difference is most of the work.
What national research tells us
There is little published research on AI answering in pubs specifically, so it is worth being clear about what the broader evidence does and does not show. Use of AI tools is now mainstream: six in ten UK adults had used an AI chatbot in the previous three months when the government last asked in 2024[1]. Familiarity is not the same as warmth, though. The same survey found that public associations with AI are still dominated by negative ideas, with younger people more positive[2].
Government research on AI adoption also found that face-to-face roles in sectors such as hospitality are widely seen as resilient to AI[3]. Read together, these findings suggest guests are comfortable using automated tools for simple tasks, but they still expect the hospitality itself to come from people. That is a useful line to draw on your own phone.
Venues are more cautious than guests
Operators are moving slowly. In late September 2026, 76.1% of UK accommodation and food businesses had no plans to adopt AI in the next three months[4]. Among those using or unsure about AI, improving business operations was a far more common reason than serving customers: 39.2% against 17.8%[5].
Source: Office for National Statistics, 2026[5]
Show the data
| Item | Value |
|---|---|
| To improve business operations | 39.2% |
| To provide or personalise services for customers | 17.8% |
That caution is sensible. A pub that tries automated answering on customer calls is doing something most of its competitors have not, which means its guests may not have met it before in a hospitality setting. The introduction has to be clear and the first experiences have to be good.
Which calls guests are happy to automate
| Call type | What the caller wants | Likely reaction to automation | Approach |
|---|---|---|---|
| Opening and food times | A fast, correct answer | Usually fine, often welcome | Automate, from an approved source |
| Directions, parking, dogs, garden | A quick fact | Usually fine | Automate |
| Simple booking request | Confidence the table is sorted | Fine if told clearly when it will be confirmed | Automate as a request, confirm promptly |
| Christmas or party enquiry | Options and prices from someone who knows | Fine for capturing details, not for the sales conversation | Capture details, hand to the events lead |
| Allergy or dietary question | Reassurance from someone accountable | Uneasy | Hand to trained staff every time |
| Complaint | To be heard and put right | Negative | Log briefly, hand to the duty manager |
| Regular calling the landlord | A familiar voice | Often negative | Offer a transfer or callback early |
| Sensitive bookings, such as a wake | Care and flexibility | Negative | Hand to a person |
The pattern is consistent. The more factual and routine the call, the more acceptable automation is. The more emotional, risky or personal it is, the faster it should reach a person.
What makes an automated answer acceptable
- Honesty from the first sentence. Callers who are told up front that they are speaking to an assistant react far better than those who find out halfway through.
- Speed. A short greeting and a direct answer. Long menus of options lose people.
- Accuracy. One wrong bank holiday time will do more damage than a dozen correct answers do good.
- An easy exit. Saying person or pressing a key should reach a human or arrange a callback without argument.
- No false promises. A booking request described as confirmed, when it is not, is the quickest way to lose trust.
- A sense of the place. The assistant should use the pub's name and its real details, not generic phrases.
Our guide to telling callers they are speaking to AI has sample openings that cover honesty and the exit in about fifteen seconds.
Where guests will push back
Allergy calls are the clearest case for a person. Among people with a food hypersensitivity, 71% felt comfortable asking staff for more allergy information when eating out[6]. That comfort is built on talking to someone who can check with the kitchen and take responsibility for the answer. An automated assistant should recognise these calls and hand them over, never answer them.
Complaints are the second. A guest who rings to complain wants acknowledgement first and a solution second; an assistant can take the details but should pass them to the duty manager quickly with a callback time. The third is the relationship call: the regular who always rings the landlord, or the older caller who is uneasy with automated systems. Offering a person early, rather than as a last resort, keeps those guests on side.
Introduce it before guests discover it
Guests react better to a change they were told about than to one they stumble into. A pub has more ways of talking to its regulars than most businesses, so use them before the first caller hears the new greeting.
- A short note by the bar and on the website contact page: when the team is busy or we are closed, our phone is answered by an assistant, and you can always ask for a person.
- A line in the next newsletter or social post explaining what it handles and why: so the phone gets answered even when the bar is three deep.
- A quick word from staff to the regulars who usually ring, with the direct route to the landlord or manager.
- A clear line in booking confirmations saying the request was taken by the assistant and confirmed by the team.
Keep the explanation about the guest, not the technology. Nobody needs to know which system you use. They need to know the phone will be answered, what it can help with and how to reach a person when they want one.
Measure acceptance in your own venue
National surveys are a starting point, but your guests are the only ones who matter. Track a few simple numbers from the first day.
Requests for a person
The share of calls where the caller asks for a human. A high figure on simple calls suggests the greeting or answers need work.
Early hang-ups
Calls that end in the first few seconds. Some are wrong numbers; a rising share suggests the greeting is too long or off-putting.
Completed requests
Of the callers who wanted to book, how many left a complete request with a working mobile number.
Guest comments
Ask hosts to note any remark about the phone at arrival, positive or negative, for the first month.
Staff view
A five-minute check at the weekly meeting: which calls did the team have to sort out afterwards, and why?
Review the numbers weekly for a month. Most problems show up as a cluster of one call type, which you can then move to staff or fix at the source.
Example: a four-week trial at a village pub
Example: The Wheatsheaf, a fictional village pub, switched on an assistant for calls that rang more than five times during service and all calls after closing. In week one, about one call in five asked for a person, most of them regulars who rang the landlord directly. Early hang-ups were rare, and every booking request had a complete name and number.
Example: the landlord added a line to the greeting, If you'd like to speak to Dave or the team, just say so, and set those requests to ring his mobile. By week four, requests for a person had fallen to about one in eight, the regulars knew how to reach him, and the pub stopped finding missed calls on the handset each morning. Two guests mentioned the assistant on arrival; both had been told about the booking confirmation process and had no complaint.
So, will they accept it?
Most guests will accept an AI answering the pub phone for the calls where speed matters more than conversation, provided they are told what it is and can reach a person easily. They will not accept it for complaints, allergies or calls that rely on a relationship, and they should not be asked to. Start with the routine calls, measure what happens and adjust. If you are deciding what to automate, our guide to handover rules sets out which calls should always reach staff, and the test plan shows how to check a system with your own calls first.
Frequently asked questions
Will older customers cope with an AI on the phone?
Many will, for simple questions. Some will not want to, so offer a person early in the greeting and make the transfer or callback work. Never make a person the last option after a long menu.
Should we tell regulars before we switch it on?
It helps. A note on the bar, a line in your newsletter or a quick word from staff means regulars are not surprised, and you can tell them how to reach a person directly.
Do guests mind AI taking booking requests?
Generally not, as long as it is clear the booking is a request and the team confirms it quickly. Uncertainty about whether the table is really booked is what annoys people.
What is a sign that guests are not accepting it?
A high share of callers asking for a person on simple calls, early hang-ups, or comments from guests on arrival. Check which call types they cluster around and fix those first.
References
Department for Science, Innovation and Technology (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (opens in a new tab) ↑
Department for Science, Innovation and Technology (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (associations with AI) (opens in a new tab) ↑
Department for Science, Innovation and Technology (2026). AI Adoption Research (opens in a new tab) ↑
Office for National Statistics (2026). Business insights and impact on the UK economy, wave 165 dataset (AI adoption plans) (opens in a new tab) ↑
Office for National Statistics (2026). Business insights and impact on the UK economy, wave 165 dataset (reasons for AI use) (opens in a new tab) ↑
Food Standards Agency (2025). Food and You 2: wave 11 key findings (opens in a new tab) ↑