Booking and Expedia at the crossroads of agentic AI

The evolution of Artificial Intelligence assistants is profoundly changing how users express needs, formulate queries, and make decisions. In this context, agentic AI is beginning to take hold—a technological layer capable of interpreting complex intentions and executing specific actions on behalf of the user. Its impact spans multiple sectors, although the hospitality industry is playing a central role due to its transactional, recurring, global, and high-value nature.

During 2025, the technological foundations that make this model possible were laid, especially with the establishment of integration frameworks that allow assistants to connect with the businesses where transactions take place. The year 2026 appears to be a turning point, marked by the launch of new products and, above all, by real adoption by both users and providers.

In this scenario, a key question arises for Booking and Expedia: to what extent is it viable to remain on the sidelines of an ecosystem where travel intentions could originate outside their traditional platforms? The expansion of agentic AI presents, on the one hand, the possibility of accessing incremental demand that currently doesn't reach OTAs, especially from users who express complex needs or who abandon traditional processes due to a lack of context or conversational support.

Another relevant factor is the progressive migration of users toward natural language-based interfaces. This trend is not solely due to the technological advancement of AI, but also to the conversational limitations of current intermediation models. The lack of seamless, multilingual experiences geared toward comprehensively addressing needs is accelerating this shift in the starting point of the user journey.

From an economic perspective, agentic AI also presents itself as a way to reduce the heavy reliance on performance marketing and dominant platforms for traffic acquisition. At the same time, failing to participate in this new channel opens the door for other players—global distributors, bed banks, modernized distribution systems, or even hotel chains—to occupy that space, taking advantage of the ease of technological integration and disrupting the current competitive balance.

However, adopting this model is not without risks. The potential loss of direct customer relationships, increased distribution costs, the threat of progressive commoditization of offerings, and the transfer of critical loyalty data represent structural challenges for large OTAs. Added to this are regulatory uncertainty, the loss of control over offering presentation, and the sensitivity of the financial market to high-impact strategic decisions.

In this context, developing proprietary conversational channels becomes crucial as a defensive and strategic element. Without internal expertise in natural language, context, and advanced query resolution, OTAs risk being relegated to a secondary role within the new decision funnel, entering the agentic ecosystem from a position of weakness.

The most likely scenario, according to the analysis, points to a hybrid model. Agentic AI would focus on capturing intent and initial inspiration, while OTAs would contribute inventory, contextualized content, and control of the final transaction. This integration would require minimal agreements on data sharing, ranking criteria, and consistency in results, preserving the strategic assets of each party.

Thus, the relationship between OTAs and large language models appears less as a confrontation and more as an inevitable interdependence, in which the ability to adapt early will be crucial in defining future relevance within the digital travel ecosystem.