Most of the talk about AI in hotels centers on speeding up back-of-house operations. Richard Valtr wants owners to think bigger.
The Mews founder believes AI tools will ultimately prove most useful not to operators chasing efficiency, but to the people who own the asset — because AI will eventually link net operating income, and even the cost of capital, down to individual operational decisions.
"Fundamentally, the owners are the ones who care the most about the impact of every single decision, whether it's a purchasing decision or operations or trying to [upsell]," Valtr said. "Then, say we expect to have this many upsells every single day. We expect the bottom line to be impacted in this particular way. We expect costs to be here."
Valtr and Greg Naidoo, Mews' chief development officer, laid out that argument to a packed house during the "AI's Growing and Strategic Role in Hospitality" roundtable at this month's NYU International Hospitality Investment Forum in New York.
The promise, as Valtr framed it, is granularity: AI will let owners benchmark cost decisions on an hourly, daily, or weekly basis and decide in real time how aggressive they want to be.
"AI centralizes everything," he said. "One of the things… that's been the story of maybe the last 10 years, is that owners are now much more understanding of what is happening on their properties than ever before and AI just supercharges that. We see that owners are now much more understanding of every single cost item and how that actually impacts the guest experiences in a much broader and more thorough way than before, and that's also thanks to the increased digitization of the operations of a hotel."
Where the profit actually shows up
For Naidoo, who spends much of his time talking to owners, the efficiency story is only half of it. The bigger prize is revenue.
"Profitability is definitely an absolute must in terms of creating AI efficiency across the entire business, but I think there's an element of trying to also create products that guests will end up buying more and even spending more, because that's where most of the revenues are coming from," Naidoo said.
He pointed to a concrete shift: Owners are now using data to monetize spaces they once left on the table.
"Before this was never a thing," he said. "You wouldn't monetize every single space that you have and now hotels are trying to do it."
It's a fast-moving target. AI is in an incredibly interesting place, Valtr said, because "it feels like every single thing that we knew six months ago is now no longer true six months later."
He opened the discussion by polling the room: who uses AI daily, and who counts as an advanced user. The founder was pleasantly surprised by the number of daily and advanced users.
So far, Valtr said, he hasn't seen many owners using AI to interact with the hotel's property management system, but he expects that to change.
"If you open up LinkedIn, you will see everyone and his dog talking about how AI is going to transform this industry, but I don't think we have seen those use cases yet," he said. "Strangely enough, we haven't actually gotten too many of those requests from proprietors yet, but in our minds that's really only a question of time."
That gap is part of what Mews is building toward. In May, the company introduced "Mews OS," which it bills as hospitality's first true unified operating system for hotels.
"We're trying to unify a lot of those different data sources into one kind of engine, which is both the analytical engine of the hotel, but also another one which speaks more to the action-oriented one," he said. "We're thinking about basically the brains of the operation, plus the hands in those ways and making sure that those are then actually translated to the devices that every single team member uses."
The ROI question
If the upside is clear, the math is not. One roundtable question cut to the chase: How do hoteliers measure return on an investment that isn't cheap now and will likely only get pricier?
"This is probably what's still holding [the use of AI] back," Valtr said, noting he isn't seeing many groups leaning into their Claude accounts or accessing data through their Google workspaces yet.
Part of the resistance, audience members said, is plain distrust. Some users intentionally avoid AI because they don't want the "AI to get smarter." But the room agreed that hesitation should fade as the technology becomes more familiar.
For Valtr, the stakes are bigger than any single tool. Large language models, he argued, are the first technology to become genuinely transformative for hospitality.
"The cloud was never really that. We started in the cloud and built for the cloud, but there's a reason we've been doing it for 14 years: it was never that transformationally useful," Valtr said. "Fundamentally, you know, the real gains were marginal, and it was more about kind of the future execution. Now, (with AI) a lot of those things are actually really coming to the fore."