The $100 million OTA war is over. Here's the new battleground.

TLDR

• The context: The hotel industry spent roughly $100 million over a decade trying to bypass OTAs through direct-booking campaigns—with minimal impact on independent hotel performance.
• The new threat: AI agents (ChatGPT, Perplexity, Google AI Overviews) are emerging as a new distribution layer sitting above both OTAs and hotel websites.
• Action items: Test your current AI visibility, audit your structured data for machine readability, and reallocate direct-booking spend toward API integrations that make you "AI-readable."

For a decade, the hotel industry poured $100 million into a single mission: cut out the OTAs. It was a massive, industry-wide crusade—and according to data from Skift, it barely moved the needle. For independent hoteliers, the question isn’t just about whether that money was wasted; it’s about why the fight they were chasing just became obsolete.1

The $100 million stalemate

Starting in 2015, major hotel chains launched aggressive "Book Direct" campaigns to reclaim guests. While brands like Hilton and Marriott saw modest 5–8 percentage point gains in direct-booking mix, the impact was negligible for independent hotels. Lacking massive loyalty budgets, most independents saw OTA dependency remain stagnant in the 60–80% range. For smaller properties, the decade-long push left them structurally unchanged and still trapped in high-commission cycles.

The shift to AI agents

The distribution landscape is shifting as AI agents—including ChatGPT, Google AI Overviews, and Perplexity—emerge as the new booking layer. When a guest asks an AI to "find a boutique hotel in Edinburgh," the agent parses inventory and completes the transaction via API. This bypasses the traditional interface entirely; the guest never visits an OTA, nor your hotel website. The AI agent now sits above the entire distribution stack.

If this model scales, the old rivalry between hotels and OTAs becomes a fight for second place. The agent owns the guest relationship, deciding which properties to surface. Whether the final booking flows through a direct engine or an OTA is merely a back-end detail invisible to the traveler.

Why independent hotels are currently invisible

The uncomfortable truth for independent hotels: OTAs are better positioned for the AI-agent future than most independent hotel sites. OTAs have structured inventory data, standardized rate/availability APIs, and rich content libraries that agents parse easily. Most independent hotel sites have none of these. A poorly structured hotel website with inconsistent descriptions and no API layer is functionally invisible to an AI agent.

If an AI agent can pull clean data from an OTA in milliseconds but hits a wall of "messy" code on your direct site, it doesn’t "try harder." It moves on. You don't just lose the booking; you effectively stop existing to the guest.

Three steps to stay relevant

1. Test your 'AI visibility': Ask AI tools for boutique hotels in your city. See if your property surfaces, and observe whether the agent pulls from an OTA, Google Hotel Ads, or your own site. If you don't appear, that is your baseline problem to solve.
2. Audit your structured data: Do you have schema.org markup for hotels? Is your room inventory machine-readable? Prioritize fixing this over spending on direct-booking ads.
3. Reallocate your spend: Chain-hotel math (spending millions to move the needle 5–8 points) doesn't work for a 28-room independent. Direct-booking budgets might be better spent on content richness and API integrations that make your property visible to agents.

The winners in the next decade won't be the ones who fought hardest for direct bookings—they'll be the ones whose data the agents can actually use.