For thirty years, online travel has run on the same loop: a search box, a list of results, a checkout page. The web let travelers compare flights and hotels without calling an agent. Mobile put that comparison in their pocket. Now a third shift is underway, one where an AI assistant does the searching, comparing, and increasingly the booking itself. AI-native platforms like trip1.com are built for it, and the industry's largest players are chasing it too, with Expedia's leadership calling artificial intelligence travel's "third chapter."
trip1 is a hotel-booking platform with access to more than 3 million properties across 200+ countries. It has operated since 2025 and, unlike the incumbents, has opened up its entire search, rate-comparison, and booking system so that AI assistants can use it directly inside a conversation. For investors watching the online travel agency (OTA) space, that kind of shift matters because it touches the two things that drive these businesses: customer acquisition cost and conversion.
From search to agent
The near-term change is a move from tools that suggest to tools that act. A conventional site returns options and leaves the work to the traveler. An agentic system aims to assemble a full itinerary, rank flights, filter hotels, and rebook when plans change, all with minimal input.
Consumer appetite looks real. In Booking.com's global research, a large majority of travelers said they want AI involved in future trip planning. But there is a catch the industry calls a "trust gap": travelers happily let AI help them discover a trip, yet they still want a recognized brand handling the transaction when real money is on the line. How much control the traveler keeps versus hands to an agent is now the central design problem in travel technology.
The incumbents are moving first
The established platforms are spending where they see value. Expedia has rolled out natural-language planning that turns a plain-English request into a bookable itinerary, plus search by "vibe" rather than filters. It also leans into Answer Engine Optimization, structuring content to get cited inside AI assistants, which it calls one of its fastest-growing acquisition channels. Kayak has launched a chat-based agentic booking testbed. Airbnb reports its AI assistant now resolves a meaningful share of support inquiries without a human. ChatGPT itself has begun hosting travel apps, with Expedia and Booking.com among the first partners.
The logic is simple: personalization lifts conversion, automated servicing lifts margin, and being visible inside AI-driven discovery protects the top of the funnel as traffic fragments across chatbots, social feeds, and search.
How you book a hotel with ChatGPT
The connective tissue behind this shift is the Model Context Protocol (MCP), an open standard that lets AI assistants plug directly into outside services and take real actions rather than just offer advice. trip1 exposes its hotel inventory and booking system as MCP tools, so an assistant like ChatGPT can search and reserve without the user ever opening a browser tab. In practice the flow is short:
Connect the trip1 MCP server to your AI client using the details at trip1.com/agents. It works with ChatGPT, Claude, Gemini, and other MCP-compatible tools.
Describe the trip in plain language: destination, dates, budget, and anything that matters, such as "near the train station" or "well-reviewed and under $150."
The assistant queries trip1's live inventory, filters by price, rating, and distance, and returns a shortlist with real rates, cancellation terms, and availability.
You confirm a choice and the assistant creates the reservation, returning a secure checkout link to complete payment.
The advantage over asking a chatbot for hotel tips is accuracy. Without a live connection, an AI answers from stale training data and may list hotels that are unavailable or mispriced. With MCP, the numbers are real-time.
Where the challengers fit
The incumbents' edge is trust and inventory depth. Their weakness is decades of legacy infrastructure that makes reinvention slow. That gap is what AI-native platforms are trying to exploit. trip1 frames its approach as "sentiment travelling," using AI to build itineraries around the emotional arc of a journey rather than pure logistics. Founder Tomas Achmedovas argues the booking part of travel has long felt like a chore, with forms designed for an earlier internet, and that it no longer has to. The platform starts with hotels and plans to connect flights, restaurants, and experiences into one continuous itinerary.
Whether a young platform delivers on that vision is an open question, since early products routinely lag incumbents on completeness and polish. But the bet illustrates the thesis: if AI becomes the interface for travel, the winners may not be the ones with the biggest footprint, but the ones built for an agent-first world from the start.
What to watch
Three things separate durable value from hype. First, execution: a polished itinerary is worthless if pricing or availability break down. Second, monetization: bookings from AI channels remain small even at the largest players, and the economics of agent-mediated discovery are unsettled. Third, the trust gap: the platforms that convert AI-assisted browsers into buyers will be those that pair automation with credibility.
Travel is an emotional category disguised as a transaction. The emerging consensus is that AI will amplify the human side rather than replace it, handling the busywork so travelers, and the brands they trust, can focus on the rest. For a sector this large, even a modest shift in how trips get booked is worth watching.
