Sam Shank: AI Agents Threaten the Ad Model Behind Booking and Expedia

Sam Shank, co-founder and former CEO of HotelTonight, says AI-driven booking could challenge the sponsored search model that generates much of the revenue at Booking.com and Expedia. Speaking with Good Morning Hospitality host Wil Slickers at Skift Global Forum, Shank laid out where he thinks distribution leverage moves next — and it's not toward the incumbents' ad machines.

This is a practitioner's read, not a forecast backed by disclosed data. Shank didn't put a number on how much sponsored-search revenue is at risk, and neither Booking nor Expedia has shown signs of that model cracking in filings. But the argument matters because it comes from someone who built and sold a booking business inside the current distribution stack.

What does Shank think breaks?

The core claim: sponsored search — paid listings that OTAs auction to hotels — loses its grip when travelers stop browsing results pages and start asking AI agents to book for them. An agent that compares and transacts doesn't need a ranked grid of paid placements.

"AI-driven booking could challenge the sponsored search model that has fueled major online travel agencies like Booking.com and Expedia," Shank said. That's a direct hit at the two largest players in Western online accommodation distribution, whose marketing-and-ads businesses sit on exactly that mechanic.

The caveat writes itself: today, AI agents mostly assist discovery, not transactions. Nobody has shown a large volume of bookings flowing through agentic channels. The threat is structural, not yet measured.

Where does he see the openings?

Shank sees two winners if the shift happens:

• Specialized travel agents — vertical AI products with domain depth rather than generalist assistants.
• Booking infrastructure — the pipes that execute AI-driven transactions, whoever fronts the customer relationship.

If he were running HotelTonight today, Shank said, he would reposition it as a specialist agent for last-minute accommodations — same niche, new interface. The takeaway: brands with a sharp focus may translate better into an agent world than broad catalogs do.

How do hotels and vacation rentals stay visible?

Shank discussed how hotels and short-term rental operators can maintain visibility as discovery shifts from search pages to AI-powered tools. He didn't prescribe a playbook in the interview summary, and that gap is telling — right now, nobody controls ranking inside an AI answer. Suppliers who spent years optimizing for Google and OTA search are facing a channel where the ranking rules are opaque and changing.

For sellers of travel, the practical question is whether AI discovery rewards direct relationships and structured content, or whether a new intermediary simply replaces the old one.

What could slow adoption?

Shank named trust, privacy, and security as the factors that will shape how quickly travelers hand over booking to AI agents. Handing an agent payment credentials and travel preferences raises stakes that sponsored-search browsing never did.

He also sketched a future where general-purpose AI assistants and specialized travel agents coexist — the assistant as front door, the specialist as executor. That division, if it holds, determines who owns the customer and who owns the margin.

The bottom line: Shank is describing a plausible redistribution of booking economics, not a measured one. Watch whether Booking and Expedia start disclosing AI-channel revenue or agentic partnerships before pricing this as disruption rather than discussion.