AI Will Know Your Travelers. And the World’s Travelers. Will You?

Conversion Gap Is Closing

Travel companies face a new distribution question: will AI simply drive visitors to brand sites, or will it absorb the entire discovery and booking process? At a Skift Global Forum executive roundtable, presented by Merkle, Jonah Brown, VP of Industry Strategy at Merkle, and Julie Hoffmann, Global Director of Industry Strategy for Travel and Loyalty at Adobe, analyzed this tension with top travel brand executives.

AI-driven traffic is closing the conversion gap. No one at the table could say with certainty where transactions will actually occur as travelers increasingly turn to generative AI and the LLMs behind it. „Where will the transaction actually happen? In the agent? Or in the LLM generally?” one participant asked. Adobe data presented at the table showed AI referral traffic grew 119% in July 2026 and 3,500% between May 2025 and May 2026. Initially, AI-referred visitors engaged heavily but converted less often. That gap has narrowed. AI-referred traffic now converts at a rate only about 1% below non-AI traffic. This means brands need content that meets the expectations of two distinct audiences. Travelers respond to emotion and human connection. At the same time, AI agents demand structured, trustworthy, and precise information.

Travel Service Remains Hybrid

Participants compared AI agents to human travel advisors who, in the past, curated destinations, hotels, and experiences into deeply meaningful itineraries. AI can eliminate the repetitive work of routine, repeatable travel — especially in business travel, where policies, preferred suppliers, and known routes set clear boundaries. However, leisure travel, often complex with high expectations, may still require human judgment, especially when a booking fails or a disruption occurs. Travel companies must decide which tasks an AI agent can handle and when a traveler needs a real person.

Personal Discovery Must Stay Individual

The group wondered if AI recommendations would steer all travelers to the same hotels, restaurants, and attractions. Recommendation systems tend to favor what is most mentioned, reviewed, and chosen, so discovery risks drifting toward an average „sea of sameness.” A useful personal agent should reflect who the traveler is and what they want, not recycle the most popular itinerary. Brands can help by publishing richer, contextual content, maintaining preferences that show what regeneration, comfort, or exploration means for each individual. The goal is guided choice that leaves room for the joy many travelers find in the „work” of planning.

Relational Memory Is the Strategic Advantage

Generative AI engines can model discovery, but brands hold a relational layer built on years of real interactions. This advantage only counts if the data is usable. Hotels and travel platforms run by financial services firms have a „full person” context. Companies continue to value direct bookings because direct relationships provide identity, purchase history, credibility, loyalty status, and travel preference context across multiple providers. „Whether it comes from Booking.com, an OTA, or direct, it’s still a booking,” said one participant from a hotel group. „Obviously, we prefer direct bookings because we get more information about the guest.” That participant’s acute priority is ensuring the hotel remains discoverable regardless of the booking channel. Participants want connected travelers who anticipate the next step, for example, pairing a confirmed hotel with ground transport or recognizing a repeated corporate itinerary.

Loyalty can also protect rates, benefits, and differentiated services that public AI platforms cannot yet match. For another participant, the question is: „How do you ensure you stay part of the conversation, even when discovery happens in a space you can’t control?” Scaling this process requires clear ownership, coordinated budgets, redesigned workflows, and a data model that lets AI detect patterns and act in real time, without burdening teams with yet another disconnected tool. This content was created collaboratively by Merkle, Adobe, and Skift Studio.