There's a clear trend in travel right now: global travel and corporate travel giants are quietly building MCP. KAYAK built an agent-to-agent connector. Amex GBT integrated Claude directly. Even Didi-style ride-hailing is moving toward MCP.
Most people haven't processed what this means. I talk to friends in corporate travel, and the first reaction is always "what's MCP? What does it have to do with my travel booking?" A lot. This could mean future corporate travel booking isn't employee opens an app, searches, books, and files expenses. It means AI handles the whole thing automatically.
This post covers what major travel players have built on MCP and what it means for corporate travel.
KAYAK: Agent-to-Agent Connector, AI Compares Prices Directly
KAYAK is one of the world's largest travel comparison platforms. They recently launched an agent-to-agent connector. Short version: AI agents call KAYAK's search and comparison capability directly.
What does this mean? Before, users booked flights and hotels by opening KAYAK's site, entering city and dates, scrolling results, comparing prices. Now the user tells AI "book me a flight from Shanghai to Beijing next Wednesday." AI calls KAYAK's MCP interface, runs search, comparison, and filtering automatically, and recommends the best option.
No app to open. No dozens of result pages to scroll. AI gives you the optimal answer.
That's the MCP effect. KAYAK shifted from "a website people use" to "a tool AI calls."
Amex GBT: Claude Integration, AI Books Your Business Travel
Amex GBT (American Express Global Business Travel) went more aggressive. They integrated directly into Claude's MCP ecosystem.
Amex GBT is one of the world's largest corporate travel management companies, serving tens of thousands of enterprises. Before, employees booked travel like this: open the travel platform app → search flights and hotels → pick options that meet travel policy → submit for approval → pay → after the trip, file expense reports with receipts.
After integrating Claude, the flow becomes: employee tells Claude "book me a business trip to Shanghai next Friday." Claude calls Amex GBT's MCP interface—automatically searches flights and hotels that meet company travel policy, submits for approval, pays, syncs the itinerary to calendar, and generates the expense report after the trip.
The employee says one sentence. AI handles everything else.
That's the future of corporate travel: from "employee manually operates" to "AI autonomously books."
Even Ride-Hailing Is Going MCP
Not just international giants. Domestic players are following. Hello Chuxing recently did MCP-style integration.
Before, users needed the Hello app: enter start and end points, pick a time, publish the ride, wait for a driver. Now the user tells an AI assistant "I need to get from home to Hongqiao Station tomorrow at 8 AM." AI calls Hello's MCP interface, publishes the ride, matches a driver, and confirms the order.
Even a relatively niche ride service is going MCP. You get the trend.
The Domestic Scene: Hotel and Flight MCP Ecosystem Is Taking Shape
Domestic development is moving fast. RollingGo's hotel MCP provides an all-in-one booking skill for ClawHub, Coze, and Qclaw agent platforms. It supports 40+ major LLM clients directly including Cursor, Claude Code, Codex, and Windsurf. Behind it: 2M+ global hotel properties and 110K+ direct-contracted hotels. Direct inventory with real-time price confirmation. What you see is what you can book. Completely free, no call-volume limits. It aggregates 500+ global suppliers, anchored upstream of OTA supply. Strong pricing on international hotels plus Shanghai, Hong Kong, Japan, Korea, and other popular destinations. Developer partners set country-based markup rates and earn commission.
Want to try a mature domestic hotel MCP? Grab an API key and get started.
Setup in Qoder:
{
"mcpServers": {
"RollingGo-Hotel": {
"type": "sse",
"url": "https://mcp.rollinggo.ai/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}
Restart Qoder and you're calling hotel search, room detail, and price confirmation directly. Corporate travel booking, from manual to AI automation, is this simple.
Behind it: 110K+ direct-contracted hotel properties, anchored upstream of OTA supply. Strong pricing on international hotels plus Shanghai, Hong Kong, Japan, Korea. Before, these AI apps integrated hotel APIs themselves. Now they connect MCP directly, saving massive integration costs.
From international to domestic, from OTAs to TMCs to vertical mobility services, the whole travel industry is going MCP. Not isolated. Industry-wide trend.
The Future of Travel Booking: From Manual to AI-Autonomous
Why are giants building MCP? Because the entire travel booking model is fundamentally changing.
Before:
- Employee opens app, searches flights and hotels
- Compares prices, times, locations
- Judges whether it meets company travel policy
- Submits for approval
- Pays
- Files expense reports after the trip
Employee spends 30 to 60 minutes. Admin teams spend hours reviewing receipts and checking policy compliance.
Future:
- Employee says one sentence: "I'm going to Shanghai next Wednesday, back Friday"
- AI searches flights and hotels meeting company travel policy
- AI compares prices and filters optimal options
- AI submits for approval (policy-compliant options auto-approved)
- AI pays (company account)
- Itinerary syncs to calendar automatically
- Expense report generates after the trip. No receipts.
Employee says one sentence. Admin review workload drops dramatically.
This isn't sci-fi. Amex GBT is doing it with Claude. KAYAK opened MCP interfaces. Domestic hotel and flight MCP is maturing fast.
What This Means for Corporate Travel Leaders
If you're a corporate travel or admin leader, start paying attention now. Not because you need to rip out your existing system tomorrow. Because this change is coming fast.
What you can do now:
1. Check whether your current travel platform has open APIs or MCP capability.
Future AI travel assistants need to call your travel system's data. If your platform is closed, AI can't connect when the time comes. UX suffers.
2. Think about whether your company travel policy can be structured.
For AI to auto-judge "does this flight meet policy?" the policy has to be structured. Like "economy class, under 2000 each way, book seven days in advance." If your policy is vague and buried in a PDF, AI can't process it.
3. Follow MCP ecosystem development. No rush to launch, but stay informed.
Early stage right now. In a year or two, AI travel assistants will be everywhere. Understand it now and you move first.
KAYAK, Amex GBT, Hello Chuxing all building MCP isn't a coincidence. The whole travel industry is moving in the same direction: from "people operate platforms" to "AI calls tools."
Future travel booking isn't "people find services." It's "AI handles everything." Employee says one sentence. Rest is automated.
The trend is clear. It's just a question of speed.
Is your company still doing manual travel approval? Drop your current travel flow in the comments. I'm curious how painful it is.
Top comments (0)