Agentic AI and the Future of Travel: What executives need to know

During a McKinsey Live event, senior partners Jules Seeley and Kelly Ungerman discussed how the travel and hospitality industry is beginning to use agentic AI, what the future might look like, and how to seize the opportunities and mitigate the risks associated with the impending transformation.

According to recent research by Skift Research and McKinsey, 90 percent of the industry is experimenting with generative AI. But when it comes to agentic AI, adoption remains varied.

From the frontline to the back office, the leading travel and hospitality players have begun applying agentic AI throughout their organizations. Although the possibilities are exciting, so far, few organizations have seen the value reflected in their P&Ls.

Agentic AI has enormous potential to deliver richer and more tailored experiences to consumers. The question is, what happens to the industry when this agentic future arrives? We’ve identified four scenarios:

• Copilot commerce. AI agents are supporting tools embedded in the existing structure. The customer remains in control of the booking process, and OTAs and hotel websites continue to be the primary points of entry.
• AI-powered experience curation. Intermediaries help customers plan and organize the entire trip.
• Direct-booking renaissance. Customers engage directly with suppliers with proprietary AI agents that sift through data to create offers tailored to individual travelers.
• Agent takeover. AI agents for each customer and supplier negotiate and transact directly with each other, automating the entire process.

To realize the technology’s full impact, organizations will need to create new AI strategies, governance, and infrastructure—altering core business processes and ways of working. Companies will have to shift from scattered pilots to enterprise-scale transformations architected by cross-functional teams and championed by deeply engaged C-suite leaders.

Q&A from the session

1. How might agentic AI disrupt travel search and book, and what are the risks for suppliers and intermediaries?

AI is changing how travelers plan and book trips. What used to be a series of clicks along a funnel is becoming a continuous, conversation-driven cycle that spans search, booking, and trip management. On the upside, conversational search makes it easier to match traveler intent with the right supply through personalized recommendations. But that same precision brings new challenges for suppliers. As AI overviews and chat-based results capture intent earlier in the journey, they shrink the traveler’s consideration set, making visibility harder to maintain in a more curated world.

Intermediaries have been quick to partner with large language model (LLM) platforms, giving them efficient access to content and inventory. Still, they face their own risk of disintermediation if LLMs “own” the traveler relationship and push intermediaries into more behind-the-scenes infrastructure roles.

To stay ahead, travel organizations need to shore up both their technology and strategy. On the tech side, that means scalable, cloud-based systems that can support AI at enterprise scale. Strategically, it’s about earning traveler trust and becoming the authoritative source of data and execution for external AI agents. And inside the organization, success will require more than new tools; it will demand rethinking core processes, updating standard operating procedures, and investing in upskilling so teams can work confidently in an AI-enabled environment.

2. What will need to be true for agentic AI to gain traction and drive a majority of travel bookings?

At the end of the day, agentic AI has to earn travelers’ trust. Right now, fewer than 5% of people feel comfortable taking their hands off the wheel and letting an AI book on their behalf. For that number to climb toward 50%—and eventually mainstream adoption—agents will need to prove they can consistently deliver: finding the best options, booking through reliable channels, and getting it right every time. Transparency will be key. Travelers need to understand why an agent chose a certain flight or hotel, especially when prices, fare classes, or inclusions differ.

That kind of trust also depends on suppliers being ready behind the scenes. Airlines, hotels, and other providers will need to publish standardized, machine-readable offers covering dynamic pricing, content, cancellation rules, and ancillaries so that AI systems can weigh trade-offs and build complete itineraries. They’ll also need robust booking and servicing APIs, not just availability feeds, to enable smooth end-to-end transactions.

And as more bookings start through AI intermediaries instead of traditional websites or apps, the industry will need new commercial agreements: clear rules for who gets credit, who carries liability, and how revenue is shared.

* * *

For more on this topic, see the research reports Remapping travel with agentic AI, The state of AI in 2025: Agents, innovation, and transformation, Seizing the AI advantage, and The economic potential of agentic AI: The next productivity frontier; and the articles “How could hotel booking be disrupted by agentic AI?,” “Deploying agentic AI with safety and security: A playbook for technology leaders,” “The change agent: Goals, decisions, and implications for CEOs in the agentic age,” and “The agentic organization: The next paradigm for the AI era”; and a recording of the McKinsey Live event “Agentic AI: Moving beyond pilots to enterprise impact”—all on McKinsey.com.