HotelTonight Co-Founder: AI Agents Threaten OTAs' Sponsored Search Model

The sponsored search model that generates much of the revenue at Booking.com and Expedia could come under pressure from AI-driven booking, HotelTonight co-founder and former CEO Sam Shank said at Skift Global Forum.

Shank laid out his position in a Good Morning Hospitality bonus episode with host Wil Slickers. His core claim: AI agents reshape how travelers discover and book online, and that shift strikes directly at the paid-placement economics that powered the major online travel agencies.

The conversation is a practitioner's read on where distribution leverage could move, not a forecast backed by disclosed data. Shank built and ran HotelTonight, the last-minute accommodations platform, before exiting the CEO role, and his argument draws on that operator experience rather than on published figures.

What replaces sponsored search?

Shank sees two openings if AI-driven booking erodes the sponsored search model. The first is specialized travel agents — AI agents built for a narrow task rather than general assistance. The second is the booking infrastructure that executes AI-driven transactions underneath those agents.

That infrastructure layer matters because an AI agent still needs something to complete the reservation. Whoever supplies that plumbing sits in the transaction path, a position the OTAs have occupied for two decades.

Would he reposition HotelTonight?

Yes. Shank said he would reposition HotelTonight as a specialist agent for last-minute accommodations. The brand's original premise — same-day rooms — maps onto the specialist-agent model he predicts, where a focused agent outperforms a general-purpose assistant on a defined use case.

He also offered a view on how the two categories may coexist: general-purpose AI assistants and specialized travel agents working in relationship with each other rather than one eliminating the other.

What does this change for hotels and operators?

For hotels and vacation rental operators, Shank addressed the visibility question directly: how do properties stay discoverable as travelers shift from search-result pages to AI-powered discovery? His answer, in practical terms, is positioning — operators need to think about how their inventory appears to agents, not just how it ranks in a paid listing.

Adoption, he cautioned, will not run on capability alone. Shank named trust, privacy, and security as the factors that will determine how fast travelers let an AI agent complete a booking on their behalf. A traveler who will not hand payment details and preferences to an agent will keep booking the old way, whatever the technology can do.

For an industry that has watched OTAs take a large share of bookings through paid search, the open question is whether AI agents redirect that flow — and who collects the toll when they do.