While undoubtedly convenient and beneficial for consumers, this new reality presents a challenge for TTH businesses.
For providers like hotels and airlines, it means consumers may have a harder time finding their products. All the metadata associated with product availability is held in siloed legacy booking systems built on outdated technology. LLMs have a very hard time scraping that data, which means they’re unlikely to share it with users who might otherwise be interested. That means missed opportunities.
For online travel agencies (OTAs) like Expedia that aggregate travel information across service lines, it means a risk of becoming obsolete. LLMs can very easily aggregate travel options just like OTAs, but in a conversational, accessible way that consumers increasingly prefer. As it becomes easier and easier to use generative and agentic AI for end-to-end travel booking, OTAs’ value proposition could diminish along with their customers’ loyalty.
Daunting though it may seem, these challenges don’t mean game over–at least, not if hotels and OTAs meet the moment and adapt. There’s plenty that companies and their leaders can do to evolve and stay relevant. Let’s dive into the opportunities at hand.
For hotels, the opportunity centers on rearchitecting.
If hotels can successfully structure their data for AI consumption, they can encourage more direct booking.
All product data, including availability, upgrade options, loyalty programs, and amenities, needs to be structured and indexable so LLMs can easily comb the site and retrieve the information they’ve been instructed to find. What’s more, each page needs a robust semantic layer to help AIs parse site logic and navigate options correctly. Other GEO-friendly features like FAQs can be helpful, too, providing ready responses for consumers who ask LLMs questions like, “Where’s the best place to stay near the Washington Monument in June?”
The idea is to entice LLMs away from OTAs and into direct booking by making data as accessible as possible. And then, once AI “hands off” the lead, users who land on the hotel site need to be met with a superior, personalized experience. Hotels will be better positioned to own financial transactions if they can leverage loyalty data to provide more attractive options than a generic OTA listing.
It may take years to modernize legacy technology or migrate to a new, general standard, but it will be well worth the lift once LLMs can find what they’re looking for and funnel users directly to hotel websites.
For OTAs, it’s about carving out a new niche–and monetizing it.
While hotels are working behind the scenes to update their architecture, OTAs can capitalize on the opportunity to give LLMs what they need: aggregated, cleaned, and structured data.
OTAs are already good at coalescing thousands of disparate travel providers into one seamless stream; it’s their competitive moat. They should lean into this core competency, rather than trying to fight a losing battle over UI on their own websites or apps.
The savvy way forward is for OTAs to let AI capture the front-end user experience and instead become the indispensable, high-utility backend for the entire agentic travel economy. Tools like ChatGPT and Gemini need a way to retrieve inventory across multiple providers, and OTAs could create the ultimate shortcut by delivering a model context protocol (MCP) server that collects and aggregates available inventory.
This means that, rather than AI reaching out to try to scrape data from individual hotels or airlines, they would simply tap into two or three MCP endpoints belonging to, say, Kayak to build a comprehensive overview. This service–of providing a verified, clean inventory stream–can be monetized to replace the revenue lost from on-site retail media.
OTAs could also go one step further by becoming an essential memory layer for travel. Currently, OTAs don’t excel at sourcing deep user preferences or maintaining long-term travel context. But if organizations can build specialized travel assistant agents that maintain users’ entire travel history and preferences (including specific properties/companies to prioritize or ignore), and then make those agents accessible to AI tools, they can bolster their value proposition and become must-have resources.
Between driving continued relevancy and opening up new revenue streams, there’s a great deal of upside potential for OTAs who ride the AI wave and embrace their new position in the market.
The bottom line is this: AI is fundamentally changing TTH for the better, and the providers and OTAs who stubbornly resist this change will get left behind.
As industry consolidation continues, there’s a lot up for grabs for the businesses who can meet consumer needs and evolve with the times. What it will take is both an emphasis on data management and an acknowledgment of strengths.
For providers like hotels, their strengths lie in travel loyalty and proprietary data. For OTAs, it’s about data delivery. Continuing to lean on those strengths while treating AI as a new distribution layer–not a competitor–is what will make the difference and drive lasting success.
