The Pipe, the Pivot, and the Protocol - Westjet, Travelport and does MCP replace NDC

Eric Marketts

Welcome back to The V1 Airline Retailing Report. I'm Eric Marketts.

Steph Nell

And I'm Steph Nell. Week two. We made it .

Eric Marketts

We made it. And Steph, I have to say — the response to episode one was genuinely encouraging . A lot of people in distribution, airline commercial teams, TMC leadership. The audience we wanted is actually showing up.

Steph Nell

Which means we have to keep earning it. And this week I think we've got the stories to do that.

Eric Marketts

We really do. For anyone who's new — quick orientation. The V1 Airline Retailing Report is a weekly show. Every Monday, Steph and I break down two or three stories that matter in airline and travel distribution. Not the press release version. The real version — what it means, who benefits, what the risks are, and what the industry keeps missing. For today's news stories, please see the episode details for links and references.

Steph Nell

This week we have three stories, and I want to say upfront — they connect. They're not three random headlines. They're three different speeds of the same transformation playing out simultaneously.

Eric Marketts

Set it up.

Steph Nell

WestJet is going live with NDC this month. They've described their own rollout as "boring." We'll talk about why that word is deliberate and what it actually means for the industry. Travelport has a new CEO, fifty million dollars in fresh capital, and a landmark partnership with United Airlines that doesn't look like any GDS deal we've seen before. And then we have the conversation that's been running hot at every industry conference — whether MCP, a protocol that's barely eighteen months old, could make NDC obsolete before NDC reaches full adoption.

Eric Marketts

That last one is the one I'm most fired up about . Because I think most people are asking the wrong question.

Steph Nell

We'll get there. Let's start at the beginning. WestJet.

Eric Marketts

So WestJet has been building toward an NDC launch for a while, and this month it happens. They're using Accelya's FLX platform as the offer engine. Content is going through all three major GDSs — Sabre, Travelport, and Amadeus — plus a direct connect option. And the airline's own leadership team has used the word "boring" to describe how they're approaching the rollout. Steph, what do you make of that?

Steph Nell

I love that word choice. Because boring, in this context, is not an accident. It's a deliberate signal to the travel agency community. If you've been in this industry for any length of time, you know what the not-boring NDC rollouts looked like.

Eric Marketts

American Airlines. Lufthansa. The carriers that turned NDC into a channel war .

Steph Nell

Exactly. Content pulled from GDSs, agencies facing surcharges or losing access to fares, TMCs spending eighteen months trying to manage client expectations through someone else's distribution fight. Every one of those situations created real damage — agency relationships that took years to repair, corporate travel pushback, regulatory heat.

Eric Marketts

WestJet watched all of that happen and decided — that's not our playbook.

Steph Nell

The explicit strategic decision is: we don't launch until all three GDS connections are certified and ready. Agents get to choose their access channel on day one. No coercion, no forced migration, no channel surcharge. Optionality from the start.

Eric Marketts

Which sounds straightforward. But it's actually hard to execute.

Steph Nell

Because it means your launch date is hostage to the slowest GDS integration. NDC inside a GDS is not a simple connection. The GDS has to map the airline's offer attributes — bundles, ancillaries, dynamic pricing logic — into its own data model. Every GDS does that differently. And the "fully ready" bar is a moving target. Waiting for all three is a real constraint.

Eric Marketts

So the risk is that you spend eighteen months getting the pipe right, you finally go live, and agents connect — and the content looks thin. Basic fares. No real differentiation from what they were getting through EDIFACT.

Steph Nell

That's the bear case. And it's the outcome that has killed enthusiasm for NDC at other carriers. If the first wave of agents connects and finds the offers look the same, the goodwill burns fast .

Eric Marketts

So what makes the bull case work? What has to be true for WestJet to actually deliver on the promise?

Steph Nell

WestJet has to have invested in the offer logic, not just the pipe. They're using Accelya's FLX platform, which is purpose-built for offer management — it can generate hundreds of thousands of personalized offers per second. But the platform is only as good as what WestJet puts into it. Pricing science, fare family differentiation, ancillary bundling strategy. If that work has been done, what comes through those GDS connections on day one is genuinely different from EDIFACT. If it hasn't, the pipe is there but the product isn't.

Eric Marketts

The pipe is not the product. The offer is the product .

Steph Nell

That's exactly right. And here's the broader point: the fact that WestJet can run a tri-GDS NDC launch with a modern offer platform at all is meaningful. This was not possible for a carrier of WestJet's scale three years ago. NDC has gone from experiment to operational playbook for mid-market carriers. That's a genuine sign of infrastructure maturity.

Eric Marketts

So the story isn't just about WestJet. It's about where the industry is in the adoption curve.

Steph Nell

The question for every carrier watching this is no longer "can we build this." The question is "do we have the commercial and product team to make it matter." Those are very different questions. And a lot of carriers have answered the first one and are still struggling with the second .

Eric Marketts

Okay, story two. Travelport. And I want to start with the CEO choice, because I think it tells you everything about what the board believes the company needs to become. John Mangelaars took the CEO role on April 1st. Before Travelport, he was CEO of Skyscanner.

Steph Nell

Which is a very different background from a traditional GDS executive.

Eric Marketts

Completely different . Skyscanner is a meta-search platform. It aggregates multi-source content — flights, hotels, car hire — and presents it in a format that's consumable by search engines, by comparison tools, and increasingly by AI systems. Mangelaars built his career understanding how content needs to be structured for machine consumption. That is not a skill set you develop running a traditional GDS.

Steph Nell

And it's the exact skill set you need if your pivot is toward being an AI distribution infrastructure layer.

Eric Marketts

Which is exactly what Travelport is saying it is. The company posted twelve percent EBITDA growth last year, reaching three hundred and thirty-nine million. Shareholders committed fifty million in new capital to accelerate the next phase. And the positioning has shifted — not a GDS, but an AI-enabled distribution infrastructure platform. The pitch is: we normalize multi-source content — NDC, low-cost carrier direct, EDIFACT — into a structured feed that AI-native travel builders can consume programmatically.

Steph Nell

That pitch has real strategic logic to it. Because the GDS model, in its legacy form, was always at risk as direct booking channels grew. The question was what Travelport becomes when airlines move to seventy percent direct. The answer, apparently, is the layer that aggregates and normalizes what's left.

Eric Marketts

And then there's the United deal. Which I think is the more interesting story.

Steph Nell

Talk about it.

Eric Marketts

United and Travelport announced a co-development partnership. Not a content deal. A co-development relationship. Travelport gets early access to United's NDC technology roadmap. They're building tooling together. Shared engineering resources. United will also bring its OBT extras into Travelport's Deem platform. This is a structural relationship, not a commercial one.

Steph Nell

Which means Travelport gets embedded into United's architecture in a way that is much harder to unwind than a standard distribution agreement.

Eric Marketts

That's the bull case. If you're a tech partner with co-development access to an airline's NDC roadmap, you're not just distributing content — you're helping shape it. Your integration is tighter. Your switching costs are higher. That's a defensible position that survives the AI disruption cycle.

Steph Nell

So where's the skepticism?

Eric Marketts

Travelport has been transforming for a long time . This is not the first reinvention announcement. EBITDA growth is real, but fifty million dollars is not a war chest at a company that still carries significant debt from its PE era. That's a funding tranche. And the "AI distribution infrastructure" positioning is what every GDS and aggregator is claiming right now. It's a narrative, not a moat.

Steph Nell

And United is already seventy percent direct . The contested territory is the remaining thirty percent. Co-development access to United's roadmap doesn't automatically translate to capturing more of that share.

Eric Marketts

Right. But here's the critical take, and I want to be precise about this. The question of whether Travelport succeeds is almost secondary to the question of whether the model it's attempting is the right model. And I think it is.

Steph Nell

Say more.

Eric Marketts

Traditional airline-GDS relationships are passive. Airlines deliver supply. GDS moves inventory. There's no real co-development, no shared architecture, no structural interdependency. What United and Travelport are describing is active engineering partnership. And in a world where AI agents are moving toward becoming the primary booking interface, the passive conduit has no story. The engineering partner has a story — because they're embedded in the airline's architecture and much harder to replace.

Steph Nell

Distributors who remain pipes get disintermediated. Distributors who become partners get embedded .

Eric Marketts

That's the template. Whether Travelport can execute it with more than one airline, at scale, before a newer platform beats them to it — that's the open question. And it connects directly to the third story.

Steph Nell

At the Airline Distribution 2026 conference, the conversation that kept coming back — the one that nobody seemed to want to be the last person still in the room for — was whether MCP could make NDC obsolete before NDC even reaches full adoption. Eric, give people the grounding on what MCP actually is.

Eric Marketts

So MCP stands for Model Context Protocol. It's a standard developed by Anthropic and released in late 2024. The core idea is that it gives AI agents a standardized way to call external tools and data sources. Instead of an AI system requiring deep, hardcoded API integration to interact with, say, a flight search engine, you publish an MCP server that describes your tools and how to call them. The AI agent figures out the rest. It's natively designed for how large language models actually work.

Steph Nell

And the travel industry has started publishing these servers fast.

Eric Marketts

Very fast. Booking.com, Expedia, Turkish Airlines, Amadeus, Sabre — all moving to publish or develop MCP servers. The pitch is that if an AI agent can call your MCP server, it can search, price, and eventually book travel without the brittle XML schema complexity that has slowed NDC adoption for years. And from that, the "MCP could replace NDC" narrative took off.

Steph Nell

Which sounds logical. And I think you want to push back on it.

Eric Marketts

I do, because the framing is wrong . NDC is a content standard. It defines what an airline offer contains — fare components, ancillaries, conditions, pricing logic. It's the language an airline uses to describe what it's selling. MCP is a communication protocol. It defines how an AI agent calls a tool. These are different layers of the stack. They are not competing with each other.

Steph Nell

An AI agent using MCP could be calling an NDC-powered endpoint on the back end.

Eric Marketts

Exactly. They can coexist. They almost certainly will coexist. An MCP server is the interface. NDC — or Offer and Order — is the content model inside the system that MCP server is calling. Saying MCP will replace NDC is like saying HTTP will replace SQL . Different jobs. Different layers.

Steph Nell

So if that's the wrong question, what's the right question?

Eric Marketts

The right question is: if AI agents become the primary interface through which travelers shop and book, who controls the agent?

Steph Nell

And the answer to that is not airlines .

Eric Marketts

The answer to that is Google, Apple, Anthropic, OpenAI, Perplexity. The AI platforms. And those platforms are not in the travel industry. They don't have IATA memberships. They don't have settlement relationships with airlines or BSP agreements. They care about user engagement and margin on the transaction. Their incentives are not aligned with airline commercial strategy.

Steph Nell

Which means an airline publishing an MCP server is not taking control of its distribution. It's handing a tool to an agent whose owner may have very different interests.

Eric Marketts

This is the same structural mistake the industry almost made in 2012. When NDC was first positioned as the path to GDS independence, the logic was: if we control the offer, we control the relationship. But the GDS wasn't just a pipe — it was the interface. The agency relationship sat on top of the interface. So even with NDC, the intermediary layer didn't go away. It evolved. It adapted.

Steph Nell

And now the intermediary is evolving again. Into an AI agent.

Eric Marketts

And the AI agent's owner is not in the travel industry. That's the sentence nobody at Airline Distribution 2026 wanted to close on . Because if you follow it to its conclusion, all the NDC versus MCP debate is almost beside the point. The airline industry spent fifteen years arguing about who controls the offer. The next fight is about who controls the agent that presents the offer. And that fight is already underway — and the airlines are not currently winning it.

Steph Nell

So what do airlines and distribution players actually do with this?

Eric Marketts

Three things. First — separate MCP adoption from distribution strategy. Publishing an MCP server is infrastructure. It's not a moat and it is not a competitive position. Second — watch what the AI platforms do with their travel partnerships. Expedia's deal with Perplexity, Tripadvisor's early MCP relationships — these are early signals of where the agent layer is forming. The partners who get embedded early at that layer have an advantage. Third — and this connects back to the Travelport story — the distributors and technology platforms that are actively co-developing with airlines, building structural relationships, are the ones who stay relevant. Passive pipes get replaced. Engineering partners get embedded.

Steph Nell

The pipe is always the first thing that gets disintermediated .

Eric Marketts

Every single time.

Steph Nell

Let's close it out. Three stories this week, and here's the thread that runs through all of them. WestJet goes live with NDC this month and calls it boring — which is a sign that NDC infrastructure has matured to the point where mid-market carriers can execute it as a routine operational program, not a transformation bet. The question shifts from "can you build the pipe" to "do you have the product to put through it."

Eric Marketts

Travelport raised capital, installed a new CEO with an AI-native background, and signed a co-development deal with United that looks more like an engineering partnership than a content agreement. The model Travelport is attempting — being an active architecture partner to airlines rather than a passive distributor — is the right model. Execution and scale are still to be proven.

Steph Nell

And MCP is eighteen months old and already changing the conversation. Not because it replaces NDC, but because it forces the real question: in a world where AI agents become the primary booking interface, who owns the agent? That's the conversation the industry needs to be having honestly — and mostly isn't .

Eric Marketts

Nobody in this industry gets to work on just one of those three problems at a time. That's what makes this moment genuinely interesting to cover. And genuinely difficult to navigate if you're inside one of these organizations.

Steph Nell

If you found value in this episode, share it with someone making decisions in distribution, retailing, or travel technology. These conversations get better when more informed people are having them.

Eric Marketts

We'll be back next Monday. Until then — I'm Eric Marketts.

Steph Nell

I'm Steph Nell. Thanks for listening.

Eric Marketts

Stay sharp out there .