How agentic AI threatens to upend OTAs’ dominance in search

No one knows how the agentic AI race will end, but its impact is already clear. 

Online travel agencies (OTAs) such as Expedia, Booking.com, and Trip.com are entering a new era of disruption – one that rivals the rise of Google search.

AI assistants are now planning, comparing, and completing bookings for users – reducing direct interaction with OTA sites.

This article examines how this change could impact their visibility, loyalty programs, and role in search.

Booking in the age of AI

I recently tested ChatGPT’s new Agent Mode to book a hotel in Paris. 

It managed the entire process within its own interface – searching Expedia, comparing results, recommending an option, and completing the booking – all without me having to visit Expedia’s website. 

The confirmation came from Expedia, but the experience was entirely generated by ChatGPT. That moment captures the future of travel booking.

The control point has shifted: instead of users navigating across sites, AI agents now act on their behalf. 

OTAs may still facilitate the transaction, but they lose visibility into the user’s journey, behavioral signals, and the ability to influence decisions through upsells or personalization.

The OTA dilemma

OTAs built their businesses on owning the customer experience. 

They encouraged users to:

• Compare hotels.
• Join loyalty programs.
• Read reviews.
• Add extras such as car rentals or experiences. 

Every interaction fed their recommendation engines and strengthened visibility in search results.

Agentic AI strips away much of that engagement. 

When ChatGPT or Gemini handles the process, users no longer browse multiple listings or interact with OTA features. 

They see a few options selected by the AI and complete the booking.

That drop in engagement isn’t just a UX issue. It also carries algorithmic consequences.

The ranking ripple effect

If AI-driven bookings lead to lower engagement metrics on OTA platforms, their search performance could take a hit. 

Google’s ranking systems – especially those informed by user signals – may interpret shorter dwell times, fewer clicks, and reduced repeat visits as signs of declining relevance.

The effect could extend to features like NavBoost, which rely heavily on interaction data to refine rankings. 

As fewer users click through to OTAs because their AI assistants complete the bookings, OTAs risk losing visibility over time. 

In short, the rise of agentic AI could erode the engagement signals that once helped them dominate search results.

Loyalty in an AI-first world

Loyalty has long been one of OTAs’ strongest defenses. 

Expedia Rewards, Booking Genius, and Trip.com’s loyalty tiers all aim to drive repeat bookings and direct traffic. 

But loyalty depends on habit, familiarity, and perceived value.

When agentic AI intermediates the journey, loyalty may shift from brand-based to assistant-based. 

Travelers may no longer feel loyal to Booking.com or Expedia, but to the AI system that understands their preferences and books efficiently on their behalf. 

That creates a deeper challenge: how do OTAs retain loyalty when the AI becomes the trusted travel companion?

The answer lies in data. OTAs that integrate their loyalty systems with AI ecosystems – allowing access to user preferences, points, and personalized offers – stand a better chance of staying relevant. 

Those that remain walled off risk being sidelined, with their loyalty programs rendered invisible to the AI layer managing the transaction.

This makes interoperability essential. 

Loyalty must evolve from a closed system of member-only rewards into one that connects with AI-led journeys. 

If an assistant can view and apply loyalty points directly during a booking, the OTA retains some influence. If not, loyalty will default to whichever AI makes the process easiest.

Dig deeper: How Google and Meta could disrupt travel discovery with AI

---

AI as the new search layer

The travel sector has always been driven by discovery and convenience. 

From guidebooks to search engines, each new layer has reshaped who controls visibility and conversion. Agentic AI now represents the next phase in that evolution.

Where search engines like Google and Bing once surfaced results, generative and agentic systems now interpret intent, predict needs, and complete actions. 

This marks a fundamental shift in how people search. Instead of typing “hotels in Paris” or “best time to visit Greece,” users express goals and preferences in natural language. 

For instance, “plan me a long weekend in Paris with a boutique hotel near the river.” The AI then manages everything from flights to restaurant reservations.

Here, agentic search becomes multi-modal. It combines text, voice, images, maps, and structured data into a single decision process that feels conversational and personalized. 

Users no longer compare blue links or scan multiple tabs. The AI synthesizes information across formats, makes trade-offs based on factors such as budget or travel time, and delivers an outcome rather than a list.

For OTAs, this shift changes the search journey entirely. Where they once competed for clicks on results pages, they now must compete for inclusion in AI agents’ recommendation sets. Visibility is no longer defined by position on a page but by presence in the conversation.

The new marketing battlefield

Expedia and Booking aren’t ignoring the trend.

• Expedia has already built AI integrations with ChatGPT and continues to expand its personalized tools.
• Booking’s Kayak.ai is testing conversational search and deeper personalization.

These moves aim to keep both brands in the conversation as users shift from screen-based browsing to dialogue-driven discovery.

But the competition is evolving quickly.

OpenAI’s Atlas browser, Perplexity’s Pro Search, and Google’s Gemini ecosystem are all developing travel-specific capabilities. 

They don’t need to become travel agencies to disrupt OTAs – only to own the interface where users make choices.

That control point determines what data is shown, which options are ranked, and which partners receive traffic.

Adapting AI and search strategies

For OTAs to adapt, they must treat agentic AI not as a threat but as a new distribution channel. That requires three key shifts in strategy.

AI-ready data

OTAs must structure and expose their inventory data so it can be easily read and interpreted by AI systems. 

Schema markup, robust APIs, and transparent metadata are essential. 

AI models rely on clean, consistent data to make accurate recommendations. Those who provide it will appear more often in agentic results.

Agentic tracking and attribution

Traditional analytics will no longer be enough. 

OTAs will need to track journeys that begin and end within AI interfaces, where standard cookies or UTM tags may not apply. 

New forms of agentic tracking will be needed to understand which interactions drive conversions and how AI-driven bookings contribute to overall performance.

Personalization through partnership

Loyalty and recommendation engines should not compete with AI systems but feed them. 

By partnering with major AI platforms, OTAs can ensure their user data and offers are factored into recommendations. 

A connected loyalty graph that works across ecosystems could become a key competitive advantage.

What happens next

The next two years will likely determine the extent of this transformation. 

If users adopt agentic interfaces at scale, OTAs may find themselves repositioned as infrastructure rather than destinations. 

Their brand visibility could fade, advertising revenue could shrink, and search performance could erode as engagement metrics decline.

Still, there’s opportunity ahead. OTAs hold vast datasets on traveler behavior, pricing trends, and conversion signals. 

If they can leverage that data to train their own AI models or power collaborative ecosystems, they can regain relevance in the evolving landscape.

The future of travel may not belong to those with the most hotels, but to those who best understand intent.

A new definition of ‘direct’

For two decades, “book direct” has meant bypassing intermediaries such as OTAs to reserve directly with the airline or hotel. 

In the age of agentic AI, that meaning is shifting once again. The new question is: direct to whom?

When an AI assistant books on your behalf, it’s not the OTA or the supplier that owns the relationship – it’s the system that understands you best.

OTAs must now compete not only on price and inventory but also on intelligence, data quality, and integration. 

Those who adapt quickly will stay visible in the AI era. The rest risk being reduced to fulfillment pipelines – quietly efficient, yet strategically invisible.

Dig deeper: Google’s AI-powered SERPs: Strategies for recipe, travel, and lifestyle bloggers