You can learn a lot when three hospitality and tourism business leaders sit down to openly discuss how they use artificial intelligence (AI).
Unfiltered conversations are becoming more common as travel professionals experiment with AI. Smaller, curious companies are embracing general-purpose AI platforms like ChatGPT directly or building their own workflows. They are also test-driving AI features built into existing systems, such as property management systems or off-the-shelf platforms.
The results can be hit and miss.
What’s working
Hotels in particular show significantly higher adoption of general AI tools, like ChatGPT, than features offered by their existing technology providers, according to research conducted for Destination AI. Its survey of hotel leaders found 72% of brands get their AI capabilities from tools such as ChatGPT, Gemini and Copilot.
It seems to be the case for Simon Hunt, owner of Kilmartin Castle in Scotland.
Speaking at the recent Independent Hotel Show in London, he highlighted where AI was saving him time.
“Definitely in our communication,” he said. “The time saving is really good for me. It used to take a couple of hours a day to respond to all the emails, now it’s 45 minutes.”
Hunt was speaking during the “Bots and beds: How far should hospitality automate?” panel alongside Ben Murphy, founder of short-term rental and holiday property management company Hosting Holidays, and Morgan Gulland, general manager of U.K. glamping network Wigwam Holidays.
Murphy noted one of the biggest AI impacts emerged after using AI to scan reviews to better understand the guest experience.
“That's helpful because when you use a third party, you take out the human element and you find out quite quickly about certain things,” he said.
Examples include identifying problems with accessing the hotel, issues with on-site staff or the check-in process. Hosting Holidays also uses AI to cross-reference its performance and benchmark competition, which helps the company assess whether it needs to improve any areas, such as average length of stay.
Search was also discussed, and “counterintuitive” advice to come from the panel was a call for properties to ensure their online travel agency (OTA) listings are as complete as possible, because that can drive results from AI search.
“They tend to think OTAs don't necessarily play as important a role. We find it's the opposite,” said Hosting Holidays’ Murphy.
“If you are on a lot of different online channels, that content helps build up your profile. So having a detailed profile … that contributes massively. So don't only audit your own website but also look at the channels.”
Gulland, meanwhile, said one advantage is using AI to prepare data analysis for Wigwam Holidays’ franchisees, covering areas like distribution and revenue management.
“Putting their numbers together used to take hours,” he said. “Analysis now takes a couple of minutes. So, before our staff have meetings, it's just all there on the dashboard.”
Wigwam Holidays’s developer is now using AI exclusively, Gulland added, with work that took 18 months now reduced to three months.
What’s not working
So far, so good. But the jury is out on other features, with questions raised around trust and data foundations.
When it comes to revenue management, Hosting Holidays' Murphy said he was “nervous” about ceding too much control.
“AI might not know about an event that’s going on, it won't always have the memory,” he said.
He also recalls testing AI answering phone systems for its contactless hotels.
“There were benefits. But if people feel like they're speaking to a robot or AI, they'll become disengaged, and we did find that in the reviews as a result,” he said. “So it's just getting a balance because, don't get me wrong, you're going to deal with an awful lot of calls that say ‘Can I check in early? Can we check out on this date?’ Anybody that runs a hotel or short-term rental will know that.”
Gulland said Wigwam Holidays misses 250 phone calls a month, and so an AI voice system would be looked at carefully over time.
For Hunt, a concern is AI’s optimism. He said his team constantly has “crazy ideas,” one of which was whether to buy a church. The accountant stepped in to point out it wasn’t a good idea.
“He calmly talks us through it, whereas AI is so confident, and you ask a question like that to AI … it will somehow find a way for this crazy idea to work,” he said.
Hunt also said he considered an AI-powered cleaning roster but didn’t go through with it.
“I like having my hand in what's going on in the business, what's happening where,” he said.
Finding the middle ground
One area the panel agreed on was the need to build the right foundations. Gulland noted the biggest difference AI made was when Wigwam Holidays connected its database with an model context protocol connection, which is the layer that sits between the large language models (LLMs) that power AI agents and data sources and tools elsewhere on the internet.
“That means you can talk to your business, all the knowledge of your business is brought together … that's where it saved us the most time,” he said.
Hosting Holidays’ Murphy said AI requires a lot of time to set up. “We tried different software solutions that automatically respond to questions. We had a few issues … if your point of truth is not accurate, it might go off-piste. If it does that, and guests pick this up, it could be a problem,” he said.
According to a survey in Actabl’s State of Hotel AI 2026 report, just 42% of respondents said they were confident their data was ready to support their AI goals.
‘We are all learning’
Independent consultant Julian Tee, founder of Compass Hotels Management, believes the panel’s comments reflect wider industry trends.
“We are all learning,” he said. “I see so many different types of businesses rely on different IT platforms, and we can add AI to that. That could be structured or unstructured.”
He agreed AI was shining through for back-office functions, and as a lever for growth. He said one of his clients was expanding from dining into offering accommodation, and testing AI voice systems.
“They are adamant that it is coming,” he said. “They're building their future plans around that. The type of property they own could lead to voice bookings by people who are interested in the area, who might want to know more about the area before they make a booking.”
Yet he remains unconvinced anyone offers a perfect solution, for now. As a result, the client outsources unanswered phone calls to a third party.
“They could replace that overflow service with AI as soon as there is a platform that is capable of answering accurately and not frustrating an individual,” he said. “The moment someone nails this, there will be an enormous opportunity in the market.”
Tee also cautioned against building too many tools in-house. By doing so, small businesses do not benefit from the expertise and bigger data sets that specialist providers bring to the table.
“If I use an example of feedback review platforms. So, all the usual ones, Booking.com, TripAdvisor, Google and so on, it’s possible to use an AI platform and instruct it to pull in information,” Tee said.
“You can give instructions as to what to do with that, in terms of presentation or an output, and that will provide results. But put that up against off-the-shelf products and look at the amount of work that's been done to produce high quality outputs, with an ability to look at comp sets and other advantages.”
For independent businesses, the message seems to be: Don’t wait for the perfect AI solution. Technology providers, meanwhile, need to ensure products keep pace with the curiosity and experimentation already happening among their customers.
“The availability of access to various different AI platforms will generate that curiosity, and as more people talk about how they’ve done this, or done that, or I've tried this, or I’ve been streamlining ... curiosity will breed trial and error, inevitably,” Tee concluded.