Will AI Search benefit the direct online channel or the OTAs?

Helping independent hotels and small groups use tech to build a healthier P&L, my bias, wishful thinking and true belief is that AI travel search is a treasure trove of untapped profitability for unique properties doing incredible things.

The key is thinking in terms of channel drift versus the more dramatic 'shift'-- and what you will do with the delta in net revenues to propel further innovations for bigger, better EBITDA down the road.

Imagine if you could drift 2% of your OTA guests over to direct year-over-year by engineering your website and BE for AI search. That's a win for cost per acquisition and ancillary spend, but it's also significantly better data with more clarity from discovery and website behavior through to guest sentiment analysis.

If you know how to use it, that 2% compounds into more tech budget for fun AI or automation tools, more paid search budget using lookalike audiences (LALs) for the next 2% drift, and more time / opex to actually fix the issues driving down CSAT.

After a few cycles of compounding, that's when the hotels embracing this start to really pull away in both guest experience and value to ownership.

We're at a crossroads and this can go in either direction.

And the answer depends on whether we, as an industry, do something to push it in the right direction.

That's one of the main reasons why AI Hospitality Alliance exists.

As to who will benefit from AI search, it depends entirely on what hotels are willing to do. And by AI search, I mean mostly agentic search, not chatbot search. For the first time in forever, AI search via OpenAI or Anthropic will give hotels the opportunity to exploit a structural advantage (leveraging customer knowledge) that OTAs won't be able to match. Will they put in the work to do so? That remains to be seen. If they do, they'll be able to pull in direct bookings, and if they don't, the OTAs will grab that demand. This isn't all-or-nothing; there are levels of participation, but they have to get in the game effectively to draw in direct bookings.

Re the financial question, Google will operate differently from OpenAI and Anthropic because they want to monetize like they do with GHA--it's still going to be pay-to-play for hotels in the Gemini world. But both OpenAI and Anthropic have business models that are heavily incentivized to optimize recommendations for the guest's wants, not for ad revenue. Both have pledged not to let ads impact recommendations, and I believe that will be true for a long time. Their competitive strategies depend on it.

Ultimately, I am of the view that everyone will benefit—provided they understand their role in the new ecosystem.

As a technology, this AI layer will eventually sit at the apex of the stack. It will be the primary consumer interface to underlying architecture, information, and capability. When a technology becomes that fundamental, it eventually serves the entire industry.

However, we must look at this through the lens of strategic outcome. It is clear that advertising and sponsored influence will quickly surround AI-based search. No one is walking away from those revenue streams, and businesses haven't lost their willingness to influence the outcome in their favor.

My position remains grounded in differentiation and authenticity. If you have a genuine reason to be found, the customer will find you. Success won't come from what you have in common on the digital shelf; it will come from your unique value proposition. Without those differentiators, navigating this new landscape will be difficult.

We operate in a high-value transaction industry where people care about the end product and service experience. That experience is our trump card, but you must be bold enough to play it.

Right now, AI Search feels much more “dangerous” for the top funnel than for the booking funnel itself.

Discovery, inspiration, comparison, trip planning, all of that is being radically compressed by conversational AI. The classic “50 tabs open” behavior is disappearing very quickly. In that sense, AI Search could certainly erode part of the traditional aggregator's advantage.

But it is also true that OTAs themselves are often the source layer feeding these AI systems. So paradoxically, AI could end up reducing traffic to OTAs while still reinforcing their infrastructural dominance underneath the interface layer.

And then there is another uncomfortable truth: perhaps hospitality is far more in love with the idea of direct booking than travelers actually are.

Consumers optimize for convenience, trust, simplicity, loyalty aggregation, flexibility, and perceived security. Not for distribution ideology.

The actual transaction layer remains extremely complex: payments, refunds, cancellations, upgrades, loyalty programs, fraud prevention, customer support, localization, dynamic pricing, edge cases. OTAs already solved most of that operational mess at scale.

If I had to answer Max’s question directly, my answer would currently be: OTAs, absolutely, even if mostly in the background.

OTAs win, at least for the next few years, and the gap will be hard to close after that.

1) Infrastructure: OTAs have clean APIs, structured inventory across hundreds of thousands of properties, and teams shipping integrations as fast as protocols ship. Most hotels run on legacy booking engines whose vendors have not moved on MCP, ACP, UCP, or WebMCP. The vendor chokepoint that blocked Book on Google a decade ago is the same one blocking agent commerce today.

2) Monetization: OpenAI turned on ads, walked back in-chat checkout after Walmart reported conversion three times lower inside ChatGPT than on their own site, and is shifting to embedded merchant flows. Affiliate commissions from OTAs solve this in one step: zero integration cost, instant access to 750,000 hotels, attribution that already works.

3) Brand context: chat strips the imagery, reviews, and reassurance signals that move someone through a booking. When trust drops at checkout, users default to what they know, which is the OTA.

The exit is first-party, warehouse-first data architecture exposed directly to agents. Hotels building it now will skip the intermediary. Most will not, and will pay commissions to a new layer on top of the old ones.

To benefit from AI search and compete with the OTAs for mentions, referrals and ultimately - direct bookings from the AI platforms, hoteliers must:

• invest in AEO (Answer Engine Optimization) to get your property’s expertise and unique value proposition extracted in the answer boxes and AI overviews ( ex. on Google’s Gemini).
• invest in GEO (Generative Engine Optimization) to have your property cited by name and featured in the AI trip planning recommendations (ex. on ChatGPT).
• ask the property’s CRS, Channel Managers and cloud PMS vendors to “project” your rates into Google’s UCP and AI platforms like ChatGPT via MCP or A2A. Ex. Cendyn CRS already enables hotel rates to be visible in AI search results in partnership with DirectBooker.

In the same time, do not forget to invest in old-fashioned SEO to get your website appear high in the rankings of relevant search results on Google and other search engines. Let’s not forget: Search marketing on Google and all of its formats: Google Ads (GA), Google Hotel Ads (GHA), organic listings (SEO) consistently contributes to over 50% of hotel website bookings. Google is not going anywhere! In Q1 2026, Google’s search revenue grew by 19% YoY.

AI Search will probably help OTAs more than hotels in the short term.

Platforms like Booking.com and Expedia Group already have huge amounts of hotel inventory, booking technology, payments, and customer support, making them easy partners for AI platforms like OpenAI and Google.

However, AI Search could also help strong hotel brands because users may get direct recommendations without browsing many websites.

Hotels with strong brands, good reviews, loyalty programs, and clear online content may win more direct bookings.

Smaller independent hotels may become even more dependent on OTAs.

Most likely:

• AI platforms will handle travel discovery and recommendations
• OTAs will still handle many bookings
• Hotels will compete to become the AI’s recommended option.

ChatGPT, Claude, Gemini, Grok and others, each have a different commercial architecture — which produces structurally different booking behaviours.

ChatGPT is already in monetisation mode. Its integrations with Booking.com and Expedia are not neutral. Unless a user explicitly asks for the direct rate, the OTA wins by default.

Claude has no affiliate dependencies baked in. It follows the logic of the query, which more often leads to the brand website — particularly when price transparency matters.

Gemini pulls from Google's broader index. Whoever has the better-optimised web presence wins. That's a technical race, not a loyalty one.

So hotels are not facing one AI channel. They are facing three distinct distribution logics running in parallel: affiliate-driven, query-driven and SEO-driven.

The OTAs won search because hotels ceded it. The same mistake is available to make again.

Be visible across all three. Keep brand.com technically clean. Maintain rate parity. And understand that as AI literacy among guests grows, the path of least resistance will shift — toward whoever earned the trust first.

A few days ago, Google published an extensive SEO guidebook in the era of AI. Long story short: GEO (generative engine optimization) builds upon the foundation of SEO, not the other way around.

In other words, AI Search is certainly a different animal in terms of how it shows results to the user. But how it gets its information relies on proven SEO techniques: URL accessibility, fan-out ranking, query-answer match, topic cluster ranking and so on.

Thus AI search is not all that different than traditional search on Google or via social media (where many folks now flock, from Instagram to YouTube and TikTok). Hotels that invest and have a solid online presence across key digital platforms are more likely to outperform those who simply rely on OTA distribution, or think that GEO tactics alone will save them. It won't.

AI search will likely help both the direct channel and the OTAs, but probably not equally.

OTAs start with a major advantage. They already control massive amounts of pricing data, reviews, search traffic, and booking behavior. AI systems value structured content, reputation signals, and conversion history, all areas where OTAs are extremely strong.

At the same time, AI search may create one of the best opportunities direct booking has had in years. Traditional search is heavily driven by SEO rankings and paid search budgets, making it next to impossible for a single hotel to compete with a marketplace that consolidates thousands of hotels. AI search will rely more on authority, consistency, positioning, and relevance. That creates an opening for hotels with a strong story and clearly differentiated experience.

This is where alignment becomes critical. Revenue management, marketing, distribution, and operations all need to support the same positioning. AI will not just compare rates. It will evaluate reputation, experience, value, consistency, and relevance to traveler intent.

In the end, AI search will likely reward the hotels and brands that combine strong digital infrastructure with clear strategic positioning, while OTAs will continue to benefit where hotels fail to tell their own story effectively.

Ultimately, AI Search will benefit both the direct online channel and OTAs, but not equally. It will depend on who controls guest relationships, data, and the booking experience.  

For hotels, AI Search is a major opportunity to strengthen the direct channel.  Instead of browsing dozens of websites, guests increasingly expect instant recommendations, personalised itineraries, and conversational booking experiences.  AI may reduce reliance on traditional search engines and allow brands to connect with consumers much earlier in the decision journey.  

However, OTAs are unlikely to disappear.  AI platforms require large, reliable, real-time inventories of hotels, flights, rentals, reviews, pricing, and availability.  And OTAs already possess this ecosystem, along with mature affiliate and monetisation models. OTAs make highly valuable partners for AI-powered travel planning. The partnership with ChatGPT and Booking.com is an ideal example, allowing travellers to plan and book trips conversationally and seamlessly. 

There’s a high chance that AI assistants become the new front-end of travel discovery, while OTAs remain the backend. OTAs could retain significant influence, even if travellers interact less directly with OTA websites.  And AI Search will reward the businesses (and thereby, direct bookings) that provide the best data, personalisation, and booking convenience.   

Thank you, Max, for initiating the conversation with relevant OTA data. Theoretically, AI assistants will pull data from both hotel and OTA websites to generate personalized recommendations. In practice, OTAs currently possess major advantages in AI search over hotel websites: massive inventories, structured data, established affiliate programs, and mature booking infrastructures that AI platforms can easily integrate with.

AI platforms need scalable and reliable travel data, and OTAs already aggregate millions of listings, reviews, pricing options, and ancillary services in machine-readable formats. This creates a natural bridge between AI platforms and OTA websites, where most hotel websites fail to compete effectively at this point.

In the long term, hotel direct channels could regain competitiveness by improving their direct-booking technologies, strengthening loyalty ecosystems, and optimizing their content for AI discoverability. Trust signals, structured data, recommendation quality, and seamless booking experiences will matter more than visibility alone in an AI-driven environment. The real question now is how quickly hotels can prepare to compete for recommendability in an ecosystem increasingly shaped by algorithms rather than human browsing.

This debate feels too binary. It does not have to be one or the other. Both arguments are partially right, but miss the bigger shift taking place.

The real issue is not whether AI Search benefits direct channels or OTAs. It is whether AI fundamentally changes who owns customer intent. Historically, search engines owned discovery, OTAs owned comparison, and suppliers owned fulfillment. AI Search compresses all three into a single conversational experience, changing the economics of travel distribution.

OTAs still have a major advantage. Their inventory is standardized, globally aggregated, API-ready, and enriched with structured content that AI systems can easily process. In contrast, much of the hotel ecosystem remains fragmented, with inconsistent technology and limited AI readiness.

But AI may also reduce the value of aggregation itself. If an assistant can narrow millions of options down to three highly personalized recommendations, the OTA advantage of “infinite choice” becomes less important. With time, consumers may prioritize trust and confidence over browsing endless inventory.

That creates an opening for suppliers with strong brands, loyalty ecosystems, differentiated experiences, and the ability to integrate seamlessly into AI-driven booking workflows.

AI Search may not eliminate intermediaries, but it will redefine them.

The OTAs will win the sprint. The question is who wins the marathon.

Right now, OTAs have the structural advantage in AI search and hotels shouldn't kid themselves about that. AI platforms currently surface OTA listings by default, leaving hotels invisible at the exact moment a traveler is making a decision. It makes sense: OTAs aggregate millions of travel products, have years of experience optimizing for digital discovery, and hand AI platforms a ready-made monetization model. And it's getting more serious: OTAs are partnering with these platforms to build agentic booking experiences, which means when a traveler books through an AI interface, it's increasingly the OTA earning that commission. In the short term, that's a hard combination to compete with.

But here's where the narrative shifts. AI search doesn't rank hotels the way traditional search engines did, it recommends the sources it trusts most. Think less algorithm, more knowledgeable friend. And that's actually where OTAs have a weakness: they don't know enough. They've got inventory and pricing, but they can't tell a traveler why your rooftop bar is worth the splurge or match a guest's preferences to a room category based on years of first-party data. The problem is, if a hotel's own rates and story aren't accessible, the AI will find what it can, and right now, that means surfacing OTA listings. The booking goes to the OTA. So does the commission. Through technologies like Model Context Protocol (MCP), hotels can now connect their rates, availability, and brand story directly into platforms like ChatGPT, Gemini, and Claude, reclaiming visibility without ceding the guest relationship. The hotels that move now will own that conversation. The ones that wait will keep paying OTAs to show up in a channel that could have been theirs.