Five Tensions in the Room

That ratio belongs to Carnival Cruise Line and it is not unusual. MIT found that 95% of organizations report zero ROI on generative AI projects.

The skills that get you to a pilot (curiosity, speed, low governance) are often not the ones needed to get to production. The result? Most travel organizations haven’t meaningfully scaled AI yet. Skift Research found that fewer than one-quarter of companies in the industry say they’ve scaled generative AI and are using it widely, and only 2% say the same about agentic AI. The organizations that crack this first will have an advantage that compounds over time and leaves others in their wake.

AI in pricing, customer service, and personalization runs on traveler trust. But even millennials and Gen Z, who are more likely to embrace AI than older travelers, aren’t giving that trust to companies yet. Part of this is because the pressure to move fast creates visible governance gaps that erode credibility.

This is already playing out in practice. Teams are pulling back, adding human oversight, and rethinking what “ready” actually means when speed and trust pull in opposite directions.

The travel industry is reorganizing work around AI faster than it is preparing people for it. BCG and NYU found that only 2.9% of travel workers have AI skills, the widest gap of any major industry, while 36% of hotel and airline employees report receiving no formal AI training at all. AI upskilling has taken hold at a handful of travel companies and several have noted that where it’s working with teams, it required deliberate change management investments most organizations haven’t made. We are finding that AI adoption looks less like a technology rollout and more like an organizational transformation.

At the same time, CFOs are cutting headcount and citing AI on earnings calls, seeking proof in the margins. As roles are redeployed and workflows redesigned, the human element of change management threatens to undermine any meaningful acceleration of AI.

The build vs. buy decision looks different depending on where you sit. For a platform operating at global scale, owning more of the stack can mean better margins, clearer differentiation, and greater flexibility. For others, it can mean distraction, debt, and delay.

One question that most companies have not yet answered is which capabilities are genuinely differentiating, and which are effectively commodity.

Where a company sits on the AI maturity curve often determines the right answer, but the curve keeps moving. The leaders in this room have made different bets and today is a chance to stress-test them against each other.

Travel brands have spent years optimizing the direct booking experience, but AI is changing the calculus entirely. AI doesn’t send travelers to your website to browse and compare. It returns a shortlist based on what it can find, parse, and trust about you. If your data aren’t structured to speak to AI systems, you may not make that list at all. The new shelf space isn’t search rankings or OTA placement, it’s whether AI can read you clearly enough to recommend you.

That makes this perhaps the sharpest tension of the five: expose your data and restructure your operations to capture AI-driven traffic, or protect the first-party booking funnel you’ve built and risk becoming invisible. The industry is moving from search/scroll/compare to ask/shortlist/decide.