So here's something that should make every hotel operator using a revenue management system pause for about five seconds. On July 29, a federal appeals court in Philadelphia reversed the dismissal of a class-action lawsuit accusing Caesars, MGM's Borgata, and Hard Rock Atlantic City of using a shared AI pricing platform... Rainmaker, owned by Cendyn... to effectively fix room rates. The plaintiffs allege that these competing hotels fed non-public pricing and occupancy data into the same algorithm, accepted its recommendations roughly 90% of the time, and the result was artificially inflated rates. The 3rd Circuit said that's enough to let the case proceed.

Let me be clear about what's actually being alleged here, because the technology matters. This isn't about hotels using revenue management software. Every hotel uses revenue management software. This is about competing hotels feeding sensitive, proprietary data... real-time occupancy, rate positioning, demand signals... into a shared platform that then tells all of them what to charge. And they all listen. Think about that architecture for a second. You and your three biggest competitors are all sending your private data to the same system, and that system is telling all of you to raise rates at the same time. You don't need a phone call between GMs to coordinate pricing. The algorithm does it for you. That's the allegation, and honestly... the architecture makes it plausible. I've evaluated dozens of RMS platforms. Most of them ingest competitive data. The question the court is now asking is: at what point does "competitive intelligence" become "coordinated pricing"? And that's a question nobody in hotel tech has had to answer under oath before.

What makes this especially interesting (and by interesting I mean legally chaotic) is that the 9th Circuit in San Francisco dismissed a nearly identical lawsuit against Las Vegas casino-hotels using the same Rainmaker software last year. So now you've got two federal appeals courts looking at the same technology, the same vendor, the same defendants in some cases... and reaching opposite conclusions. That's a circuit split. For the non-lawyers in the room, a circuit split is basically the legal system saying "we have no idea, someone bigger needs to decide this." That someone bigger is usually the Supreme Court. The DOJ and FTC have already filed statements in similar cases arguing that using an algorithm doesn't shield you from antitrust liability. They're watching this.

Here's where this hits closer to home than Atlantic City casino floors. The RealPage lawsuits in multifamily housing... same theory, same structure, landlords feeding data into shared pricing software... have already resulted in DOJ intervention and settlements. Hotels are next in that crosshair. If you're an independent or a branded property using an RMS that also serves your comp set (and most of them do... that's literally how they build their datasets), you need to understand what data you're sharing and how the vendor uses it. I talked to a revenue manager last month who had no idea that her RMS was ingesting rate data from three of her direct competitors and using it to generate her pricing recommendations. She thought it was just pulling from public OTA rates. It wasn't. The vendor's data-sharing architecture was buried in page 47 of a terms-of-service document nobody reads.

Look, I'm not saying your RMS is a price-fixing tool. Most of them probably aren't. But the legal landscape just shifted, and "probably" isn't a defense. The question this case is forcing is whether the hub-and-spoke model... one vendor, multiple competitors, shared data, coordinated recommendations... constitutes an implicit agreement to fix prices. If courts start saying yes, that changes the entire economics of revenue management technology. Vendors will have to redesign how they handle competitive data. Hotels will need to understand exactly what they're feeding into these systems and what comes out the other side. The days of treating your RMS as a black box that just spits out a rate recommendation are over. Or they should be.

So here's the pitch: AI watches demand signals in real time, adjusts your room rate hundreds or thousands of times a day, and auto-generates personalized bundles... spa credit plus late checkout plus a room upgrade, packaged and priced dynamically for each guest based on their booking behavior. Airlines have been doing this for years. Hotels are next. One budget chain is reportedly changing prices up to 15 million times a day. The reported upside? RevPAR gains of 10-20%. Ancillary revenue bumps of $15-$40 per stay. A 20-35% lift in direct booking conversion from AI chatbots. The numbers are real enough to get your owner's attention. They got mine.

But let's talk about what this actually does at property level. Because I consulted with a hotel group last year that bought into one of these AI pricing platforms... mid-tier vendor, decent reputation, solid demo. Implementation took four months instead of the quoted six weeks. Their PMS integration broke twice during peak season. The revenue manager spent more time troubleshooting rate discrepancies than actually managing revenue. And the "dynamic bundles" the system generated? Half of them offered amenities the property didn't have. The AI didn't know there was no spa. It just knew spa bundles convert well. Nobody on the vendor side had bothered to map the system's offer library against the property's actual amenity set. That's a demo feature, not a production feature. There's a difference.

Look, I'm not anti-AI pricing. I'm an engineer. I've built rate-push systems. The underlying technology is legitimate... real-time demand forecasting, price elasticity modeling, automated channel optimization. When it works, it works. Hilton just launched an AI trip planner in beta. Major chains are embedding this into their tech stacks at the corporate level, where they have dedicated teams, clean data pipelines, and the engineering resources to handle edge cases. For a 3,000-property portfolio with centralized revenue management, this makes sense. The math scales. But the airline comparison keeps getting thrown around like it's a simple analogy, and it's not. Airlines have standardized inventory (a seat is a seat is a seat, mostly). Hotels have 50 different room types, inconsistent PMS data, local comp set dynamics, and a night auditor who needs to understand why the rate on a walk-in just changed three times since they clocked in.

The Dale Test question here is brutal. When this system misfires at 1 AM... and it will, because every system eventually fails... what's the recovery path for the person at the desk? Can they override the AI rate? Do they even know how? What happens when a guest pulls up a rate on their phone that's $30 lower than what the front desk is showing because the AI adjusted between the time the guest searched and the time they walked in? That's not a hypothetical. That's a Tuesday. And if your answer is "the system handles it automatically," you've never watched a guest argue about a rate with a 22-year-old front desk agent who has no idea what algorithm priced the room. The real cost isn't in the subscription fee. It's in the training gap, the integration maintenance, the staff confusion, and the guest friction that doesn't show up on the vendor's ROI slide.

Here's what I'd actually do if I were evaluating this for an independent or a small portfolio. First, ignore the 15-million-rate-changes-a-day headline. That's a volume metric, not a performance metric. Ask the vendor for properties in your comp set running their system and get actual RevPAR index movement, not projections. Second, demand a full integration audit before you sign anything... what PMS version are you running, what's the data handshake, what breaks during night audit. Third, if you're running anything older than a 2020-era PMS, the integration cost alone might kill your ROI. That $15,000 infrastructure upgrade your property needs? It just became a prerequisite, not an option. And fourth... the bundles. Make sure any dynamic bundling system maps to YOUR amenity set, YOUR staffing levels, YOUR actual property. If the AI is offering guests things you can't deliver, you haven't upgraded your revenue strategy. You've automated disappointment.