The Great Revolution: AI vs OTAs in the 2026 Travel Market

For the last two decades, OTAs like Booking, Airbnb and Expedia have been the undisputed gatekeepers of travel distribution. However, the rapid adoption of agent-based AI and Large Language Models (LLMs) has fragmented the traditional customer journey.

The battle lines are now clearly drawn: AI is taking control of the discovery phase, while OTAs and direct booking channels fight to own the transaction.

The Collapse of the Traditional Search Funnel

Historically, travellers planned their trips through a predictable sequence: dreaming on social media, searching keywords on Google, comparing on OTAs and finally booking. Generative AI has compressed all of this into a single conversational environment.

The Shift to Conversational Discovery

According to 2026 data from Mindful Ecotourism, approximately 40% of global travellers now use AI-based tools for trip planning. Travellers no longer type “Hotel in Rome”; they give the AI complex parameters: “Find me a boutique hotel near Piazza di Spagna under €300, with flexible cancellation and fast Wi-Fi.”

A Huge Win for Direct Bookings:

A February 2026 study by Propellic, which analysed over 300 real travel planning scenarios on Google AI Mode, revealed a staggering shift. When users planned trips through Google’s AI, 56% of resulting booking flows went directly to the provider’s website, while less than 10% went through OTAs. In the AI ecosystem, users tend to click on direct links cited in text rather than scrolling through OTA comparison listings.

The OTA Counterattack: Integrate Rather Than Compete

Despite catastrophic headlines and recent stock market volatility for major travel groups, the OTA fortress is not crumbling; it is adapting. OTAs understand that while LLMs excel at planning, they fail as “merchant of record” (managing payments, cancellations and customer service).

Instead of fighting AI, OTAs are redirecting it:

• Infiltrating the Chat: Both Expedia and Booking were among the first to launch native apps within ChatGPT. They are positioning themselves as the underlying inventory engines for AI platforms.
• Expedia’s ‘Comet’ Browser: In collaboration with Perplexity, Expedia recently launched the Comet browser, merging natural language search with its vast travel database to create an end-to-end AI planning tool.
• The AI Traffic Redirect Strategy: OTAs are actively defending their ecosystems. When ChatGPT recommends a specific hotel and provides a Booking.com link, the latter silently redirects that user. Instead of landing on the specific hotel page, the UTM source is stripped and the user is placed into an “infinite scroll” feed of the entire destination.

Why OTAs Still Win by Default (the technical gap)

Wall Street analysts at Morgan Stanley recently noted that OTAs will likely remain fundamental to the travel landscape long-term. Because OTAs have spent decades structuring their data.

LLMs do not “rank” pages like traditional search engines; they synthesise sources they can reliably parse. OTAs win by default because they provide AI with clean, readable catalogues with rooms, rates and availability.

In contrast, most independent hotel websites hide their rates behind complex booking engines designed for human clicks, not machine reading. When an AI cannot easily verify a direct price, it defaults to the OTA.

The Opportunity to Increase Direct Bookings

To capitalise on the 56% direct booking opportunity presented by AI search environments, operators must rethink their digital infrastructure. Traditional SEO is no longer sufficient.

• Create an AI-Native Storefront: Properties must structure their digital data. Room types, bed configurations, real-time rates and cancellation policies must be encoded in structured formats (APIs) that AI agents can easily read and verify.
• Optimise Google Business Profiles: Since Google AI Mode heavily favours inline text links, ensuring your Google profile is hyper-accurate and detail-rich is critical for appearing in AI-generated itineraries.
• Aggressive Re-acquisition Campaigns: Over 68% of travellers still default to OTAs, but automated post-stay emails directed to OTA-masked addresses offer a reliable way to convert them to direct bookings for future stays.

What to Do to Avoid Falling Behind

Immediately: Audit your direct booking site for mobile speed (target: <2.5 sec load time) and schema markup implementation. Claim and optimise Google Business profiles for all properties. Test your properties on ChatGPT, Perplexity and Google SGE using natural language queries.

Short term: Invest in semantic content: neighbourhood guides, detailed amenity descriptions and local experience recommendations. Implement instant booking confirmation and a secure, reliable payment processing system on your direct site.

Medium term: Allocate 8-12% of marketing budget to AI channel optimisation (current industry average is 3-5%). Integrate your PMS with new AI distribution APIs as they launch.

Strategic: Diversify beyond Airbnb, Booking and Vrbo. Maintain presence on 4-6 booking channels to maximise AI data aggregation.

Summary

The narrative that AI will kill OTAs is fundamentally wrong. AI is simply the new distribution layer. OTAs are aggressively integrating into this layer, but the very nature of conversational AI is democratising discovery.

For properties willing to structure their data and build AI-readable storefronts, the opportunity to bypass OTA commissions has never been higher.