← Back to Feed

Booking.com Is Spending Billions on AI. Your Front Desk Is the Collateral.

Booking Holdings is pouring resources into AI that plans trips, handles complaints, and cuts customer service costs by double digits. If you're an independent hotelier, the question isn't whether this technology is impressive... it's how much of your guest relationship you're about to lose.

Booking.com Is Spending Billions on AI. Your Front Desk Is the Collateral.
Available Analysis

So let's talk about what Booking Holdings actually announced at VivaTech a couple weeks ago, because the stock ticker story is noise. The real thing happening here is an OTA spending aggressively to put an AI layer between your guest and your hotel... and doing it well enough that the guest might never need to talk to you at all.

Here's what they've built. Priceline has an AI travel agent called Penny that's showing increased engagement in early testing. Booking.com has rolled out natural language search, smart filters, and... this is the one that should make you sit up... agentic service flows that handle complaints and cancellations without a human. Agoda has already hit double-digit year-over-year reductions in customer service cost per booking through AI automation. This isn't a pitch deck. This is production code running at scale across multiple platforms. I've evaluated enough hotel tech to know the difference between a demo and a deployment, and this is deployment. The architecture is real. The results are measurable. And the strategic intent is crystal clear: own the guest from inspiration to post-stay, and make the hotel the fulfillment layer.

Look, I get why Booking is doing this. Their stock is down roughly 19% year-to-date because investors are worried that general AI models (think Google AI Mode, ChatGPT) could disintermediate OTAs entirely. So Booking's play is to become the AI layer itself... build the "Connected Trip" that manages everything so the guest never needs to leave the ecosystem. It's a defensive moat disguised as innovation. And from an engineering perspective, it's smart. They're partnering with OpenAI, Google, Anthropic, Amazon. They're not building foundation models, they're building the travel-specific application layer on top of the best available models. That's the right architectural decision. But here's the thing nobody in the hotel industry is talking about: every efficiency gain Booking makes in customer service is a touchpoint they're pulling away from your property. Every complaint their AI resolves is a complaint your front desk never hears about... which means you never get the chance to fix the underlying problem, and you never get the recovery moment that turns a frustrated guest into a loyal one. I consulted with a hotel group last year where 30% of their repeat guests cited a problem-resolution experience as the reason they came back. You don't get that if the OTA's AI handled the complaint before your team even knew it existed.

The financial picture makes the strategic picture worse. Booking is running a 23.3% adjusted EBITDA margin and growing revenue 16% year-over-year. They bought back $3.6 billion in shares in Q1 alone. They have the capital to keep building this for years. Meanwhile, 63% of bookings at many independents already flow through OTAs, and that number isn't shrinking. When an OTA with this kind of financial firepower starts using AI to own the pre-arrival, in-stay service, and post-stay feedback loops... you're not a hotel anymore. You're a supplier. And suppliers don't set terms. They accept them.

The question I keep coming back to is my standard one: what happens at 2 AM when nobody's here? Except now I'm asking it about the guest relationship, not the technology stack. What happens to your ability to know your guest when Booking's AI is handling their complaints, adjusting their itineraries, and personalizing their next trip... all without your property touching any of it? The answer is you become invisible. And invisible suppliers get commoditized. If you're running an independent or a small portfolio, you have maybe 18-24 months before this AI service layer is polished enough that guests genuinely prefer it to calling your front desk. Start building your direct booking infrastructure now. Not a loyalty program that mimics the big brands. Something real... a guest relationship that the OTA's AI can't replicate because it requires a human who actually works in your building and knows the guest by name. That's your moat. The only question is whether you'll build it before Booking finishes building theirs.

Operator's Take

Here's what I'd bring to your next owner meeting if you're running an independent or a soft-branded property with significant OTA dependency. Pull your channel mix report from last quarter. If more than 40% of your revenue comes through Booking or Expedia, you've got a guest relationship problem that's about to get worse... fast. The actionable move this week: audit every guest touchpoint where the OTA currently sits between you and the guest. Pre-arrival communication, complaint resolution, review solicitation. For each one, build a direct alternative. Even something as simple as a personal text from your front desk manager 24 hours before arrival changes the dynamic. This is what I call the Vendor ROI Sentence applied to your distribution partners... if you can't articulate what value the OTA is providing beyond heads in beds, you're paying 15-22% commission for a relationship someone else owns. Get your direct channel strategy on paper before Q4. Not a wish list. A plan with a number attached to it.

— Mike Storm, Founder & Editor
Source: Google News: Booking Holdings
🏢 Anthropic 🏢 Google 📊 Penny 🏢 Agoda 🏢 Booking Holdings 🏢 Booking.com 📊 Customer service automation 📊 Guest relationship management 📊 OTA disintermediation risk 📌 Priceline
The views, analysis, and opinions expressed in this article are those of the author and do not necessarily reflect the official position of InnBrief. InnBrief provides hospitality industry intelligence and commentary for informational purposes only. Readers should conduct their own due diligence before making business decisions based on any content published here.