Only 6% of hotels appear in AI search results as travel industry grapples with legacy systems

The Skift Data + AI Summit 2026 brought together 15+ industry leaders to assess where travel AI stands after years of investment and hype. The conversation was sobering: while AI adoption is accelerating, the industry is discovering that technology deployment alone does not generate returns.

The most striking data point is that 95% of GenAI projects report zero ROI. This is not a technology problem—Marriott's Colin Coleman made clear that "the technology is not the hard part anymore." Instead, the bottleneck is organizational: companies lack the workforce expertise to identify and prioritize problems worth solving, and most pilot projects end without clear ownership and accountability. As Priceline's CTO noted, "the planning-to-building ratio has inverted itself," signaling that preparation now exceeds execution.

The second critical insight is distribution: only 6% of hotels surface in AI search results, a finding from Curacity's Cornell research that maps how different traveler segments (budget, aspirational, luxury) interact with AI at different funnel stages. This is a revenue problem because AI search is rapidly becoming the primary entry point for travel booking. Nick Slavin, CEO of Curacity, framed it plainly: "AI search is the new front door for travel."

Hilton and Air Canada's leaders stressed that modernizing core infrastructure is now a prerequisite, not a nice-to-have, because legacy systems cannot scale AI. Conversely, speakers from Sierra and Amadeus warned that waiting to deploy agentic AI is no longer viable—companies already running AI agents in production have moved past the pilot phase and established competitive moats. Richard Valtr, founder of Mews, captured the highest aspiration: "technology should help you deliver better welcomes and better goodbyes," a reminder that the purpose of AI in travel is human experience, not cost-cutting.

On personalization, Priceline's CTO highlighted a shift from behavioral data alone to preference data—what customers explicitly prefer—and stressed that letting customers see and edit their own preference profiles is now a trust requirement. Broader themes include unclear ownership (who is accountable once a pilot ends?), investor focus on revenue generation rather than cost reduction, and a recognition that the traditional use-case approach fragments organizations into point solutions. The summit's consensus was clear: the travel industry has entered a new phase where execution, accountability, and infrastructure capability matter more than technology innovation itself.

The Skift summit reveals a maturation crisis in travel AI adoption. While the industry has moved past asking whether AI matters—AI search is now the primary funnel entry point—it faces a harder question: why do most AI projects fail to deliver returns? The data suggests a systemic problem: companies have invested in GenAI without simultaneously modernizing the infrastructure, organizational structures, and talent pipelines required to operationalize AI at scale.

Hilton and Air Canada's emphasis on architecture modernization as a prerequisite, not a precursor, to AI signals that legacy hotel systems and distribution chains cannot support AI without foundational overhaul. Equally revealing is the gap between pilot and production: Sierra and Amadeus speakers warned that companies still treating agentic AI as a future capability are already behind those running agents in production. The implication is that the competitive window is closing—delay is now a strategic disadvantage.

The invisibility problem—only 6% of hotels in AI search—is both a data and a distribution issue. Curacity's research shows that travelers use AI differently by budget segment and decision stage, yet most hotels lack the connected data or distribution partnerships to reach them. Priceline's emphasis on letting customers see and edit their own preference data hints at a deeper shift: trust and transparency are becoming competitive differentiators, not nice-to-haves.

No comments yet. Be the first to share your thoughts!