May 2026
I recently read the BCG and NYU report on "AI-first hotels," and like most people in the industry, I found myself nodding along.
It's an ambitious vision, but not an unrealistic one. Hotels discovered through AI assistants instead of search. Pricing that adjusts in real time based on demand, sentiment, and context. Operations that become leaner, more responsive, and less dependent on manual coordination. Even the idea that hotels could be designed and optimized faster, shaped continuously by data rather than fixed assumptions.
If you read it, it's hard not to agree with the direction. AI is going to reshape hospitality. That part isn't controversial anymore.
What interested me more was not where the report says we're going, but where it quietly acknowledges we are today.
There's a line in the report that stood out to me, almost understated in how important it is:
"Successful AI relies on seamless communication among the many data sources and technologies that manage them… Without that foundation, hotels risk deploying sophisticated tools on disjointed systems."
That's the real story. And the industry already knows it's a problem. Fragmented systems, disconnected data, manual reconciliation. These are not new challenges. The report even points out that many hotel teams still spend significant time stitching together reports just to get a complete picture of their business.
So yes, AI-first hotels sound great. But that's not where most companies are operating today. And the gap between those two realities is where things get interesting.
The Harder Question
The report frames the challenge largely in terms of investment. Are you moving fast enough? Are you adopting the right tools? Are you preparing your organization for AI?
I don't think that's the hardest question.
The harder question is more fundamental: do your systems actually agree with each other today?
Because if they don't, AI doesn't fix that. It builds on it.
One of the things we've learned over the past seven years at Omniboost is that hospitality isn't difficult because it's complex. It's difficult because it's inconsistent.
On paper, everything looks structured. You have a PMS, a POS, an accounting system, and a framework like USALI that is supposed to tie it all together. In theory, that should create alignment.
In practice, it rarely does.
Two hotels can run the same systems and still produce completely different financial outputs. Not because the systems are wrong, but because of how they are configured, how they are used, and how decisions are made at the property level. Revenue is categorized differently. Tax logic is applied differently. Timing between systems is handled differently. Edge cases are treated in ways that make sense locally but create inconsistencies everywhere else.
That's not something you solve by adding a better tool. It's something you only start to understand after seeing it happen again and again, across different properties, markets, and operating models.
Over ten thousand integrations, those patterns become a standard. That's what we've built at Omniboost, and creating that standard is what we do by design.
What AI Actually Changes
What AI changes is not that underlying reality. It changes how we interact with it.
The report describes a future where decisions are faster, operations are leaner, and insights are generated instantly. That's all very real. But it also means we move further from the mechanics of the system. Instead of navigating workflows and reviewing outputs step by step, we start asking questions and trusting the answers we get back.
And that introduces a new dependency.
When you stop interacting with the system, you also stop validating what's happening inside it. You no longer see how the data flows, where it might be misclassified, or why something doesn't quite reconcile. You're relying on the system to already know what "correct" looks like.
But AI doesn't know what correct looks like. It learns from what it sees.
If the underlying data is consistent, that's powerful. It reinforces the right patterns and makes them easier to act on. If the data is inconsistent, which in hospitality it often is, AI doesn't correct that. It learns the inconsistency and scales it.
That's the part that's easy to underestimate.
Governance and Trust
This is where governance comes in. And I don't mean governance in the bureaucratic sense. Not policies, committees, or sign-off layers. I mean something more practical than that.
Who is accountable for the standard? Who decides what "correct" looks like? Who owns the data when it moves between systems?
Those questions matter more as AI takes over the interface, because trust in the output depends entirely on trust in the foundation. If you don't know who owns the standard, you don't know whether to trust the answer.
For hotels moving toward AI-powered decision making, that accountability has to be established before the tools go live. Not after.
The Invisible Work
The report touches on this when it talks about the investment dilemma. The foundational work required for AI, cleaning data, integrating systems, standardizing structures, establishing governance, is essential but largely invisible and often deprioritized because it doesn't deliver immediate, visible returns.
That resonates.
Because that invisible work is exactly what everything else depends on.
It doesn't show up in a demo. It doesn't improve the guest experience overnight. It doesn't generate headlines. But without it, even the most advanced AI capabilities struggle to deliver meaningful value.
We're Doing This Too
I want to be honest about something. At Omniboost, we're not just advising on this from the outside. We're in the middle of it ourselves.
Across our departments, we're running 90-day AI challenges. No prescribed direction. No defined playbook. Just a challenge, a timeframe, and the space to figure it out.
Some things work. Some don't. And that's entirely the point.
Because at this stage, there often isn't a right or wrong answer. The value isn't in arriving at a perfect solution. It's in learning what the question actually is.
It connects to one of our core values at Omniboost: Growth Mindset. The belief that you learn by doing, not by waiting until you have all the answers.
Which brings me to Yoda.
"Do or do not. There is no try."
Most people read that as a push toward confidence. But I think it's really about commitment. Stop theorizing. Start doing. Learn from what happens.
That's exactly where we are with AI right now. The companies that will come out ahead won't be the ones with the best strategy document. They'll be the ones that started, learned, adjusted, and kept going.
Getting the Basics Right
So while I agree with the direction outlined in the report, I think the real dividing line in the industry won't be between companies that adopt AI and those that don't.
It will be between companies that have done the foundational work and those that haven't. Companies that have built governance into the standard, not bolted it on afterward. Companies that trust their data because they know where it comes from and why it's right.
AI-first hotels are coming. But the ones that succeed won't necessarily be the ones moving the fastest.
They'll be the ones that took the time to get the basics right first.
And then started doing.
