AI Agents Won't Kill the OTAs. But They'll Make Your Commission Problem Worse.

So here's what's actually happening. Bernstein puts out this analysis saying AI agents are creating "terminal value risk" for Booking and Expedia... margin compression, eroding supply moats, the whole existential threat narrative. Wall Street reacts. OpenAI scales back its direct booking ambitions, and Booking's stock jumps 8%, Expedia jumps 12%. Everyone exhales. Crisis averted.

Except nobody's asking the question that matters to the person running a hotel: what happens to YOUR cost of distribution while these trillion-dollar companies figure out their AI strategy?

Look, I've been watching the OTA "disruption" narrative cycle for years now. Google was supposed to kill them. Metasearch was supposed to kill them. Now AI is supposed to kill them. And every single time, Booking and Expedia don't die... they adapt, they spend more, and they pass those costs downstream. Booking is targeting $450 million in cost savings by 2027, with a chunk reinvested into AI automation. Expedia cut 3% of its headcount (down to 16,000) and is plowing those savings into machine learning. Both companies are partnering with OpenAI and Google Gemini. They're not sitting still. They're spending aggressively to make sure that when you search for a hotel on whatever AI platform emerges, their inventory shows up first. And who funds that arms race? You do. Through commissions, through rate parity restrictions, through the loyalty program assessments that keep climbing.

Here's the part that actually matters at property level. Bernstein's own numbers tell the story: a 1% improvement in conversion rates could boost OTA EBITDA by 30%. Think about what that means. These platforms are optimizing conversion with AI... getting better at turning a browsing guest into a booked guest... and capturing more of that value. Meanwhile, 40% of travelers say they'd book directly through an AI chat interface if pricing and payment were integrated. That's your direct booking channel getting squeezed from both sides. The OTAs get smarter at converting, AND new AI platforms start funneling demand through their own pipes. Online hotel booking penetration could push from 66% to 80%, which sounds like growth until you realize the intermediary's cut grows with it. More bookings going through more middlemen, each one taking a piece.

I talked to an independent owner last month who told me he tracks his "true cost per booking" across every channel... OTA commission, loyalty assessment, brand marketing contribution, rate parity discount, all of it. His OTA bookings were costing him north of 22% when you stacked everything up. His direct bookings were at 8%. And his OTA mix was climbing, not falling, because the platforms keep getting better at capturing demand before the guest ever sees his website. That's not a technology problem. That's a distribution economics problem. And AI isn't solving it for the hotel... it's accelerating it for the platform.

The real shift here isn't whether AI kills Booking and Expedia. It won't (not anytime soon). The real shift is that AI makes every intermediary more efficient at extracting margin from the transaction... while making it harder for individual properties to compete for attention in an AI-mediated search environment. Your website, your SEO, your metasearch strategy... all of that was built for a world where a human types a query into a browser. When an AI agent queries multiple sources, compares prices, and presents options in a conversational interface, the rules change. And nobody's rewriting those rules in favor of the 150-key select-service in a secondary market. They're rewriting them in favor of whoever has the deepest API integration and the biggest data set. Which is... Booking and Expedia.