For nearly two decades, online travel agencies have faced predictions of their imminent decline. First, it was supplier-direct booking. Then mobile. Then metasearch. Then Google Flights. Today, the latest prediction is that conversational AI and AI agents will finally make OTAs obsolete.
As someone who has spent much of my career working across OTA growth, metasearch, travel distribution, Google Flights, Kayak, Skyscanner, Microsoft travel initiatives, API integrations and international customer acquisition, I believe the reality is more nuanced.
AI will transform travel discovery, trip planning and the way travelers interact with travel brands. But replacing OTAs is a very different challenge from helping travelers discover where they want to go.
The future of travel is likely to be evolution, not extinction.
Google Flights Already Changed Travel Discovery
Long before ChatGPT, Gemini and AI agents entered the conversation, Google had already reshaped travel search.
Google Flights changed how travelers compare airfare options. It reduced friction, increased transparency and moved flight shopping closer to the search experience itself. Over time, Google became not only the starting point for travel discovery but increasingly the environment where many travel decisions are influenced.
For OTAs, this was not a new challenge. The industry adapted.
Some OTAs became more sophisticated in pricing, merchandising, packaging, loyalty and customer experience. Others focused on routes, markets or inventory where they could provide differentiated value. This is exactly the type of operational perspective often missing from high-level discussions about AI replacing travel intermediaries.
The important lesson is that Google Flights did not eliminate OTAs. It changed how travelers discovered travel products.
AI Will Disrupt Discovery Faster Than Booking
The current wave of AI innovation is exceptionally well suited to one part of the travel journey: discovery.
Travelers increasingly want answers to questions such as:
- Where can I travel in Europe with a family of four for under $6,000?
- Which destinations have warm weather in February and direct flights from Toronto?
- Where can I combine beaches, culture and great food without changing hotels multiple times?
- What is the best destination for a short trip where schedule matters as much as price?
Traditional search interfaces require users to know what they are looking for. Conversational AI allows travelers to start with intent rather than a destination, date pair or filter set.
Google’s AI-powered travel direction shows this clearly. Travelers can increasingly describe the trip they want in natural language instead of manually testing destinations, dates and filters. This is a meaningful shift. But it is primarily a discovery shift. And discovery is only one part of the travel journey.
The Complexity Gap Between Inspiration and Fulfillment
One of the most common misconceptions in discussions about AI and travel is the assumption that finding travel products and booking travel products are essentially the same thing.
They are not.
Behind every flight booking sits a highly complex ecosystem that most travelers never see:
- Airline distribution systems
- NDC connections
- Global Distribution Systems
- Fare rules and ticket conditions
- Ancillary products
- Payment orchestration
- Fraud prevention
- Ticketing and fulfillment workflows
- Schedule changes
- Airline disruptions
- Refunds, exchanges and cancellations
- Customer servicing
While AI can help travelers navigate choices, these operational layers still require sophisticated infrastructure, supplier connectivity, settlement logic, compliance, customer support and deep travel expertise.
Travel is not simply a search problem. Travel is also an execution problem.
The Servicing Problem Nobody Talks About Enough
The area that receives surprisingly little attention in AI discussions is servicing.
Consider a common scenario. A family of four is traveling from Toronto to Athens through Frankfurt. A weather event disrupts the connection. One passenger has a wheelchair request. Another has a special meal requirement. The outbound ticket includes ancillary services purchased through multiple systems. The traveler wants to understand whether to accept an airline-proposed reaccommodation, request a different routing or cancel part of the trip.
Can an AI assistant help explain options? Certainly.
Can it independently resolve the entire situation end to end with the reliability, accountability and supplier authority travelers expect? Not consistently. Not today. And likely not in the immediate future.
This is where OTAs continue to create substantial value. The true test of a travel company is often not how effectively it books travel when everything goes right. It is how effectively it supports customers when things go wrong.
What Happens to Metasearch?
Some industry observers suggest conversational AI will eliminate metasearch. I see it differently.
In many ways, conversational AI may become the next evolution of metasearch.
The underlying objectives remain remarkably similar:
- Aggregate content
- Compare options
- Rank relevance
- Surface pricing
- Match traveler intent to inventory
- Help travelers make decisions
The difference is that travelers may increasingly interact through conversations instead of filters and search boxes. The interface changes. The underlying economics remain surprisingly familiar.
Travelers still need inventory. Travelers still need pricing. Travelers still need comparison. Travelers still need confidence. Whether the interaction occurs through a search bar, a metasearch interface or a conversational assistant, those needs do not disappear.
A More Likely Future for AI and OTAs
Rather than asking whether AI will replace OTAs, I believe the more useful question is: which parts of the travel journey are most likely to be transformed by AI?
- Discovery and inspiration: high near-term disruption through conversational search and AI recommendations.
- Trip planning: high disruption as AI improves itinerary building, comparison and personalization.
- Price comparison: moderate disruption as AI changes how results are summarized and ranked.
- Booking and fulfillment: gradual change because payments, ticketing, supplier rules and accountability remain complex.
- Post-booking servicing: limited near-term replacement due to disruptions, exchanges, refunds and supplier coordination.
- Complex itineraries: low near-term automation for multi-passenger, multi-supplier and exception-heavy trips.
What OTAs Should Do Now
The wrong response would be to dismiss AI as hype. The equally wrong response would be to assume AI agents will immediately own the full end-to-end travel journey.
The more practical path is to prepare for a world where discovery becomes more conversational while fulfillment remains operationally complex.
For OTAs, the strategic priority is not to fight AI. It is to become more visible, more useful and more trusted inside AI-driven discovery environments.
- Structured data and content quality: make inventory, policies, fees, ancillaries and servicing capabilities easier for search and AI systems to understand.
- Metasearch and Google Flights competitiveness: identify routes, fare types and markets where the OTA has a genuine advantage.
- Conversational commerce readiness: build experiences that support natural-language intent, flexible search and decision assistance.
- Operational differentiation: turn servicing, disruption handling, support and flexibility into visible value propositions.
- GEO as the next layer of SEO: optimize not only for search rankings but also for how generative engines interpret, summarize and recommend travel brands.
Final Thoughts
The travel industry has a long history of overestimating short-term disruption while underestimating long-term transformation.
AI will change travel. That much is clear.
Travel discovery is already becoming more conversational, more personalized and more intent-driven. Google Flights, Gemini and other AI-powered platforms are accelerating that shift.
But replacing OTAs is a far more complicated challenge than helping travelers choose a destination.
For the foreseeable future, someone still needs to issue the ticket, manage the payment, coordinate suppliers, handle disruptions, process exchanges, support travelers and navigate the countless complexities that occur between booking and returning home.
The future may be conversational. But travel remains operational. And that distinction matters.
About Adrian Ghisa
Adrian Ghisa is a travel technology, metasearch, OTA growth and distribution executive with experience across Google Flights, Kayak, Skyscanner, Microsoft travel initiatives, API ecosystems, strategic partnerships and international travel growth.
Discuss travel technology and OTA growth