Data Isolation Is AI's Biggest Obstacle in Hospitality

The hotel industry is not short of ambition when it comes to artificial intelligence. Personalization at scale, pattern matching, real-time decision-making, guest intelligence—these ideas have been making the rounds on the conference circuit for the past few years. We're on the right track, but our data hasn't kept up with our goals.

The foundation for everything that AI promises is data. Right now, much of that data sits in silos. It's fragmented and lost across hundreds of platforms and screens. No model or algorithm can act on data it doesn't have. 

At full power, AI can analyze behavior across a global portfolio, distill it into actionable insights, recommend your next move, and execute it in near real time. The gap between that vision and where hotels are today comes down to one thing: data isolation. It's the foundational barrier our business has been working around, and the problem we must solve before we can realize any of the exciting promises of AI. 

Before any hotel can think about cross-property intelligence or portfolio-wide pattern matching, they have to solve a more immediate problem. Their own data is a mess. 

The average hotel collects guest information across the front desk, the spa, the restaurant host stand, the pro shop, or the loyalty database. None of these were designed to talk to each other. A guest who checks in, grabs a drink, books a tee time, and orders late-night room service leaves data across a handful of separate systems. The hotel rarely sees the complete picture. 

You cannot send the right offer to the right guest at the right time with that incomplete profile. Clean and unified guest profiles are table stakes. The hotels that have not yet solved these problems are not ready for AI. When you do get it right, the benefit is tangible. You're seeing increased open rates from targeted campaigns, more RevPAR from those campaigns, and guests who feel recognized. But these outcomes, relative to what comes next, are a ceiling. 

The GM, the loyalty card, and the ceiling they share

There's a version of guest intelligence that the world's best hotels have had since the days of paper ledgers and physical room keys. It's the GM that knows the regulars and their preferences—the corner room, which whiskey they drink, and their kids' names. That's the kind of recognition that has always built loyalty. And it works on the scale of a single person's memory at a single property. 

Loyalty programs were the industry's attempt to scale that primitive form of guest intelligence. They were built to carry guest recognition across properties within a single brand, making sure that a loyal traveler felt known at every property in the chain, even ones they’ve yet to visit. What those programs offered was a more advanced version of what the great GM was already doing. It’s faster, more consistent, and larger in scale, but fundamentally the same thing. And they only worked within a single chain. It didn’t work for management groups operating mixed portfolios. Most of what the industry calls AI today is a faster version of the same thing: better recall, same roadblocks. 

Take Caroline Brittel, a guest archetype hiding in plain sight across most hotel portfolios. She takes a different spring break trip every year – always within the same hotel group, different property, different country, same pattern. She books in October for the following March. Always requests a pool or ocean view. Big spender at the spa. She has done this for the past decade.

Every hotel Caroline has ever stayed at has had one entry in its system. One stay. One record. One guest who may or may not come back. None of them knows she is a spring break traveler or how to reach her in October. None of them knows about the water view or the spa. The data that would tell them all of this exists — distributed across ten properties that have never shared a signal.

Caroline is not invisible because the data does not exist. The frustrating part is that everything needed to know about Caroline exists within one organization. She is not lost to a competitor's database or hidden behind an OTA. She is a guest of the same hotel group — checked in, paid, reviewed, and gone — across ten properties that have never compared notes. The data is there. The architecture to use it is not.

The next frontier: cross-property behavioral intelligence

The GM's memory and the loyalty card were both recognition systems. What becomes possible with cross-property behavioral intelligence is something different: pattern detection across data that no individual hotel has ever had access to. The behavioral signals distributed across thousands of properties, analyzed simultaneously, reveal patterns that are invisible at the property level. Caroline Brittel's October-March travel rhythm. The "bleisure" traveler who adds room nights and upgrades every time. The family that plans two years out. These patterns exist in the aggregate, and always have. The industry has never been able to see them.

Getting there requires solving a data architecture problem that the industry has largely avoided confronting. Connecting real guest data across properties at scale poses a serious risk of exposure under GDPR and similar privacy regulations. You can't merge millions of guest records and mine them for patterns without running into data privacy requirements that exist for good reason.

The architecture that makes cross-property intelligence possible while respecting privacy is to anonymize data and work at the behavioral layer. Strip out the identifiable information — names, emails, any details that identify an individual — and analyze what remains: the behaviors. When someone books, how far in advance they book, what they spend, and what they request. These signals, analyzed at scale, reveal the patterns. Once identified, the AI can apply the patterns to real guest data to drive real outcomes — without the analysis ever touching personal information. Privacy stays private.

Hotel groups are sitting on something no individual property has: a portfolio view. The guest who stays at the budget property in Chicago and the flagship in Dubai is a stranger at both. Somewhere in the enterprise, that relationship exists in its entirety. Most groups have never been able to see it.

From detection to execution

Pattern matching is only the start. The more important part is doing something with them before the opportunity is lost. 

Even at the property level, hotel sales and marketing teams are short on time, not data. An AI system that only surfaces insights when someone thinks to query it is an improvement over what exists now, but a marginal one at best.

The real shift comes from reactive to proactive systems: AI that watches the data continuously, flags the signal when it appears, and acts — without waiting for a person to start the process.

Caroline Brittel is due to book this October. A proactive system already knows that, because it's seen the pattern. It doesn’t wait for her to start searching; it reaches her with the right offer at the right time on her preferred channel. The hotelier sets the guardrails like brand standards, offers, details, and channels, and AI tools operate within them using a wealth of data we haven't seen before. All at a speed and a scale that no human team can match on their own.

This is the version of AI in hospitality that the industry has been dreaming of. It requires data discipline to get there. Without unified property-level data, cross-property intelligence is impossible. Without cross-property intelligence, the patterns remain hidden in the ether. Without the patterns, the AI has nothing real to act on beyond what we have right now.

What hotel groups should prioritize now

To get from where we are today to where we need to go is a process, not a jump. 

The first move is to unify data at the property level, connecting systems that currently operate in isolation into a single guest record that reflects the full picture of every stay. This is foundational, and nothing else works without it.

The second move is to ask harder questions about your enterprise data. Specifically, whether those systems across the portfolio can share signals in a way that is both technically and privacy-sound. That means a data layer that anonymizes guest records before any cross-property analysis touches them. It preserves the behavioral signal while stripping the personal detail. Most groups do not have this yet. Those who start building toward it now will not be scrambling to catch up when the tools that require it become standard.

The third move is the most important one. Ask what the AI system tells you when you are not looking. A system that only answers when asked is a dashboard with better branding. The bar worth setting is proactive intelligence — pattern-based, operating within your guardrails, acting before you think to look.

The data discipline required to get there is unglamorous work. It is also the only thing that separates AI that performs from AI that promises.

Caroline Brittel books a different trip every October. The hotel groups that know that will earn her loyalty. The ones that don't will be meeting her for the first time every time.