RevPAR and AI: What We Learned Running Marriott & IHG Properties

We operate properties under Marriott and IHG flags through 360 Hospitality Group. Revenue management has been using algorithms for years — your PMS already has some version of AI adjusting rates. That's not the story here.

The story is the 60% of hotel operations that still runs on spreadsheets, WhatsApp messages, printed schedules, and tribal knowledge in the manager's head. That's where the real opportunity is.

The revenue management myth

When people talk about AI and hotels, they usually mean dynamic pricing. Rates that adjust based on demand, comp set, booking pace. The flag systems do some of this. Third-party tools do more of it.

But most independent and mid-market hotel operators aren't losing RevPAR because their pricing isn't dynamic enough. They're losing it because they don't have clean visibility into what's happening across their channels, and they don't have time to act on the signals that are already there.

The AI opportunity in revenue is less about smarter pricing and more about: what's happening right now, across all your channels, and what should you do about it? That's a data aggregation and decision-support problem, not a pricing algorithm problem.

Where the real operational leverage is

Labor scheduling.This is the biggest cost driver after property expenses, and it's almost always managed manually. A GM with good intuition can schedule pretty well. A scheduling system with AI assistance that incorporates occupancy forecast, historical patterns, and known events can schedule better — and can do it for every shift across every department, not just the ones the GM has mental bandwidth for.

Guest communication at scale. Every hotel has the same problem: too many guests, not enough staff hours to communicate with them individually. Pre-arrival messages, upsell sequences, review request timing — these are all automatable. Not with generic blast emails, but with triggered, personalized communications based on booking data and stay behavior.

Procurement and F&B cost control.The variance between budgeted and actual food and beverage cost is one of the most painful numbers in a hotel P&L. A lot of it comes from informal procurement decisions — a sous chef ordering extra inventory, a manager approving a purchase that doesn't go through the system. AI that provides visibility into where the variance is happening — by department, by vendor, by item — doesn't solve the behavior problem, but it surfaces it fast enough to address it.

The multi-property visibility problem

If you operate more than one property, your biggest operational problem is probably not any one thing at any one property — it's not having a clean view across all of them simultaneously.

Property management systems don't make this easy. The data is in there, but extracting a consistent view across multiple PMSs, flag systems, and accounting platforms requires someone who knows all the systems well enough to pull it together manually.

AI-driven dashboards that pull from multiple sources and surface the key metrics in one view — occupancy, ADR, RevPAR, labor percentage, F&B cost, guest scores — are genuinely transformative for multi-property operators. Not because the metrics didn't exist, but because having them in one place, without a 2-hour manual pull, changes how decisions get made.

What we changed at our properties

The three things that moved the needle most in our own operation:

1. Automated daily performance summary sent to every GM before their morning standup — occupancy, revenue, labor hours, and any outlier alerts from the night before.
2. AI-assisted labor scheduling that incorporates occupancy forecast and generates a recommended schedule that the manager reviews and approves, rather than builds from scratch.
3. Guest review request timing optimized by stay type — the send time and channel (email vs. SMS) vary based on booking source and length of stay. Review score improved measurably within 90 days.

None of these are revolutionary. They're the basic blocking and tackling of operational AI — applied to an industry that still runs most of its operations on gut feel and tribal knowledge.