The Hidden Constraint Holding Back AI in Travel, According to Mews Founder Richard Valtr

Richard Valtr has built one of the most modern hospitality platforms in the market, challenging legacy property management systems at their core. As founder of Mews, he has pushed a guest-centric data model long before AI made it urgent.

At Skift Data + AI Summit 2026, Valtr brings a perspective grounded in real deployments, where performance shows up in revenue, workload, guest experience, and system efficiency.

1. What has surprised you most about applying AI at scale in your organization?

“How much of the work is cultural rather than technical. Building with AI, especially using agents, follows the same adoption curve we’re seeing across the market: the early adopters move fast, but the real work is helping skeptics and late adopters see the value for themselves.

What’s also striking is how much delegating to AI changes the nature of the work. It opens up new ways to rest, refine and rethink strategies, ideas and products far more quickly than before.”

2. What is the single most important factor for successfully scaling AI in travel operations?

“Architecture. Most hospitality systems were built around the room rather than the guest; that made sense for the accounting principles of the time, but it means guest data is fragmented by stay rather than forming a continuous profile.

If you don’t know your guest, you can’t market to them, and if you can’t speak to them effectively, you can’t maximize the revenue from them.

If your AI is making decisions on this incomplete structure, it’s always working from an incomplete picture of revenue.

The unglamorous but essential prerequisite for AI at scale is getting the underlying architecture right first. Most organizations skip that step and then wonder why the results stay marginal.”

3. Are there common misconceptions about AI in travel operations you would like to clarify?

“The most common one is that AI in hospitality means fewer people, but the more interesting near-term reality is almost the opposite.

A 100-room hotel today might handle a few personalized guest requests a day. Once 30 or 40 percent of travelers have AI agents handling their preferences on their behalf, those agents have, as I put it, zero shame. They’ll ask for everything, on behalf of every guest, every stay.

The volume of incoming requests is going to increase significantly. The hotels that will handle that well are the ones whose systems are already designed to automate the back-of-house, so staff have the capacity to respond to what’s coming at the front.

The second misconception is that good AI should be visible. The most valuable implementations are the ones you don’t notice: pricing that adjusts overnight, anomaly detection that surfaces a booking pattern before it becomes a problem, suggested actions that arrive before anyone thinks to ask.

Every one of those unseen actions is time that doesn’t land on a staff member’s screen. We’ve built around a design principle we call user disengagement – the goal is for staff to spend less time looking at screens, not more. That’s what lets them look at guests instead.”

4. Where do you see the biggest opportunities for AI to impact business outcomes in travel?

“The shift from property management to profit management. Most hotels today run revenue, operations and guest experience as separate functions.

When they share a single operating system, you can apply intelligence across the whole business simultaneously: pricing decisions that account for staffing constraints, upsell offers that reflect availability, housekeeping schedules that respond to guest behavior.

The near-term impact is in revenue management through better demand forecasting and automated pricing optimization that translates directly into RevPAR gains.

Longer term, the more significant shift is in what guests will expect. There’s a Louis CK bit about Wi-Fi on planes where he talks about the idea that something unthinkable five years ago becomes the baseline, and its absence makes you angry.

Personalization delivered by AI agents is on that curve. Once guests experience it somewhere, they’ll expect it everywhere. The hotels that have invested in the underlying infrastructure – unified guest data, connected systems, automated operations – will be positioned to meet that. The ones still running fragmented stacks will be playing catch-up on an expectation they didn’t see coming.”

Hear More from Richard Valtr at Skift Data + AI Summit 2026

Join data, revenue, and technology leaders working through the same questions in real time.

This is where operators compare what’s actually delivering results, where assumptions get challenged, where new approaches take shape, and where the next wave of AI in travel becomes clearer.

Join us on June 3 in New York City for Skift Data + AI Summit 2026.