Four futures of travel in the age of AI agents

This article was written by Hospitality Today. Click here to read the original article

From OTA evolution to full disintermediation – scenarios that will shape the next decade of digital travel

May 22, 2025

The rapid evolution of agentic AI is reshaping the online travel landscape at a pace few anticipated. Once thought to threaten the relevance of online travel agencies (OTAs), these autonomous systems – powered by advanced large language models – are now raising urgent questions for every stakeholder in the travel ecosystem. With Google, Mastercard, and OpenAI all deploying features that enable AI agents to research, book, and even pay for travel autonomously, the future of distribution is up for grabs. This strategic analysis outlines four possible scenarios that could define the next chapter of digital travel – each shaped by the level of consumer trust in AI agents and the degree of disintermediation from traditional OTAs.

Key takeaways

Agentic AI is accelerating: New developments—from Google’s “AI Mode” to Visa’s agentic payments – are transforming how travel is searched, booked, and transacted.
OTAs may evolve or be bypassed: AI agents could either enhance OTA interfaces or bypass them entirely by transacting directly with travel suppliers.
Four strategic scenarios emerge:
- Augmented Booking Experience: Low trust, low disintermediation – OTAs evolve with AI support, marketing shifts toward chat-based ads.
- AI Travel Concierge: High trust, low disintermediation – OTAs remain central, but serve AI assistants through algorithmic access fees.
- Direct AI Distribution: Low trust, high supplier control – brands invest in AI-driven retention within their own ecosystems.
- Neural Travel Marketplace: High trust, high disintermediation -AI agents dominate, OTAs fade, and marketing becomes machine-to-machine.
The future is uncertain but actionable: Travel companies must focus on adaptable AI infrastructures, flexible data systems, and agentic competencies rather than trying to bet on one winning scenario.
Strategic imperative: Success lies not in predicting which AI-driven model will prevail, but in building the agility to thrive in any of them.