I worked with a GM once... sharp operator, mid-size convention hotel in a secondary market... who used to say staffing was like a dam. You can lose a little water over the top and nobody notices. You lose enough, and one day the whole thing just goes. Not slowly. All at once. Reviews tank. OT explodes. Your best remaining people start updating their resumes because they're tired of carrying everyone else's weight. He called it the "quiet collapse."

That's what I'm looking at this morning. Eighty thousand jobs gone from leisure and hospitality in June and July combined. And here's the part that should genuinely worry you... this didn't happen because people stopped traveling. Travel demand is still there. The summer season is running. Guests are showing up. They're just showing up to properties that have fewer and fewer people to take care of them.

This is not cyclical. This is structural. The labor force participation rate dropped to 61.4% in July, lowest since early 2021. People aren't between jobs. They're gone. And the ones who are still in the pool? Real wages for accommodation and food service workers actually declined over the past year... compensation up 2.4% against 3.3% inflation. We're asking people to work harder, for less, in an industry with 70-80% annual turnover. Then we act surprised when they leave. The AHLA says more than half its members are still understaffed. That's not a temporary condition anymore. That's the new baseline.

Here's where it gets really uncomfortable. Consumer confidence just dropped to 90.8. People are telling surveyors they think jobs are harder to find (which is true outside our industry... the broader economy lost 23,000 nonfarm jobs in July). So you've got a fall season potentially coming where leisure demand softens, business travel stays cautious, and your labor costs don't come down because the people simply aren't there. Labor CPOR was already up 1.8% year-over-year in Q1 to $46.79. That number is going higher. Not because you're paying people more (you're barely keeping up with inflation). Because you're paying overtime. Because you're paying agency premiums. Because you're eating the cost of service failures that show up in your review scores three weeks later and in your ADR ceiling three months after that. This is what I call the Invisible P&L... the costs that never appear on a line item but destroy more margin than the ones that do. The agency staffing premium, the OT, the comp'd night for the guest who waited 40 minutes for a check-in because you had two people at the desk... none of that shows up where your owner is looking. But it's real. And it's compounding.

The people who will get hurt worst are select-service operators with no labor redundancy. You're running a 120-key property with a skeleton crew already. One more person leaves and you don't have a staffing problem... you have a math problem. But full-service isn't safe either. You're charging $289 a night and the guest is wondering why the restaurant closed early and nobody answered the phone at the spa. The gap between what the rate promises and what the staffing delivers is where your brand equity goes to die.

I worked with a director of housekeeping once... 22 years at the same property, ran a team of 35, knew every room attendant's kid's name, their work authorization renewal dates, their bus schedules. She kept a spreadsheet she called "the real roster." Not the one in the labor management system. The one that showed how many people she could actually put on the floor on any given Tuesday. Her real roster was always shorter than the official one. She told me once, "The system says I have 35 people. I have 24. The rest are ghosts... open positions I'll never fill, people on their last 30 days of authorization, and two who stopped showing up but nobody's processed the termination yet." That was three years ago. Her real roster today is probably 18.

Let me be direct about what actually happened in Washington, because the headline floating around gets it wrong. The immigration bill didn't die. It passed. President Trump signed the Secure America Act on June 10th. Sixty-nine and a half billion dollars for ICE and CBP through September 2029. What died was any hope that the bill would include worker pathway protections... the visa expansions, the H-2B cap increases, the employment authorization stability that this industry has been begging for. Congress didn't fail to act. Congress acted. They just acted against us. That distinction matters because "gridlock" implies someone might still do something. Nobody is doing anything. Not this session. Probably not next session. The pipeline isn't frozen... it's being actively squeezed.

Here's the number that should keep you up tonight. Forty-nine percent of housekeepers in the United States are foreign-born. Not 15%. Not 25%. Forty-nine percent. In Miami, 65% of your hotel workforce... not just housekeeping, your entire hotel workforce... is foreign-born. Las Vegas, Houston, LA, New York... same story with different percentages but the same vulnerability. And June's jobs report didn't bring relief. Leisure and hospitality didn't add jobs in June. It lost 61,000 of them. That's not a soft patch. That's the workforce telling you it's leaving, and the government just told it to leave faster.

The operators who are going to survive this aren't the ones hoping for a legislative miracle. They're the ones who already stopped pretending this is a staffing problem with a staffing solution. Every-other-day cleaning isn't a sustainability initiative anymore... it's survival math. A 200-key full-service property running 85% occupancy needs roughly 28-30 room attendants if you're cleaning every occupied room daily at 30 minutes per credit. Drop to every-other-day stayover service and your daily requirement drops by 8-10 positions. That's not a nice-to-have efficiency play. That's the difference between making your room product work and not making it work. And if you're still staffed at 80-85% of what you need (which, based on the data, you probably are), it's the difference between burning out the people you have and keeping them.

The wage lever is already pulled. Seventy percent of owners reported offering higher wages as retention strategy in early 2026, up from 47% at the end of 2024. The money moved. The people didn't come. Because this was never just about money... it was about a workforce that is literally being removed from the labor pool by federal policy while the industry that depends on them writes polite letters to Congress. I've seen this pattern before. In the last major enforcement cycle, I watched three properties in the same market lose a combined 40% of their housekeeping staff in 90 days. Two of them closed floors. One went to three-day cleaning cycles before anyone was calling it a "sustainability program." They called it what it was: we don't have enough people to clean the rooms. The properties that got through it were the ones that had already restructured their cleaning model, cross-trained their teams, and stopped pretending the old staffing model was coming back. The ones that waited for the cavalry? Some of them are still waiting.

So here's something that doesn't happen very often in hotel technology: someone with a vendor-adjacent title saying "calm down." Max Spangler, VP of Technology at Charlestowne Hotels, told CoStar this week that hoteliers should stop chasing the AI hype and start integrating tools that actually help people do their daily work. He frames it as the 80/20 split... AI handles the first 80% (drafting, summarizing, comparing, organizing, analyzing), and your team owns the last 20% where judgment, tone, guest sensitivity, and owner expectations live. About 350 Charlestowne employees are using AI tools daily under this model.

Look, I agree with almost everything Spangler is saying here. And that's notable because I spend most of my time telling hotel operators that the technology being sold to them doesn't do what the sales deck promises. This is different. The 80/20 framing is honest because it acknowledges something most vendors won't admit: AI is a drafting tool, not a decision-making tool. The person who writes the first version of your RFP response, your guest recovery email, your maintenance summary... that work can absolutely be assisted by a language model. The person who decides whether that email actually gets sent, whether the tone matches the situation, whether the guest in room 412 needs a different approach than the guest in 308... that's a human job. That's always going to be a human job. Spangler gets this. Most of the industry does not.

But here's where I push back, and it's not on Spangler specifically... it's on the framing that "350 employees use AI daily" as a success metric. Usage isn't value. I consulted with a hotel group last year that rolled out an AI writing assistant to every department. Six months later, 80% of their staff had used it at least once. Sounds great, right? When we actually looked at what they were using it for, half of them were generating content they then rewrote entirely. The other half were using it for tasks that took three minutes manually and now took four minutes because they had to prompt, review, edit, and re-check. The tool wasn't bad. The implementation assumed that "available" equals "useful," and those are very different things. The question isn't how many people are using AI. It's how many people are doing better work because of it... and "better" means faster, more accurate, or higher quality. If you can't point to which one, you don't have adoption. You have activity.

The cybersecurity angle Spangler raises deserves more attention than it's getting. He's right that AI-generated phishing has eliminated the telltale signs (bad grammar, weird formatting, obvious spoofing) that we've trained front desk teams to spot for years. Your night auditor at 2 AM getting an email that looks exactly like it came from your brand's loyalty team, with perfect formatting, correct logos, and natural language... that's not hypothetical anymore. That's Tuesday. And most hotels are still running cybersecurity training built for the era when phishing emails had typos in the subject line. This is the part of the AI conversation that actually keeps me up at night, not whether your revenue manager is using ChatGPT to draft a comp set analysis.

Meanwhile, contrast Spangler's "give people useful tools" approach with what happened at Mews this week... roughly 170 people laid off because, according to their founder, AI is "fundamentally changing the economics of hospitality." That's a $2.5 billion company that raised $300 million this year cutting somewhere around 13% of its workforce and telling those people that AI changed the math on their roles. So which is it? Is AI a tool that handles the first 80% so humans can do the important 20%? Or is AI the reason 170 people don't have jobs? Both things can't be the industry narrative simultaneously. The answer, obviously, is that it depends entirely on who's making the decisions and what they're optimizing for. Spangler is optimizing for operational quality. Mews is optimizing for margins. Your technology vendor's AI strategy tells you everything about what they think your people are worth.