Why Legacy Airline Tech is Making Carriers Invisible to AI

Eric Marketts

Here's a scenario for all of you. . A traveler opens their AI assistant — Gemini, ChatGPT, whatever they're using — and types: book me a flight to Tokyo in October. The agent searches. It compares schedules, prices, ancillary options. And if the airline has structured, machine-readable offer and order data, it can complete that booking. No browser. No OTA. No GDS in the loop. Direct to confirm.

Steph Nell

And for most airlines, it stops before the booking.

Eric Marketts

Not because the agent failed. Because the content was never built for a machine to read. The offer isn't structured. The ancillaries aren't queryable. The order system can't respond to an automated confirmation request. The airline is invisible — not to the traveler, but to the agent acting on their behalf.

Steph Nell

And that's not a hypothetical. That is the state of the industry right now .

Eric Marketts

Which is what this episode is about. Welcome back to The V1 Airline Retailing Report. I'm Eric Marketts.

Steph Nell

And I'm Steph Nell. For anyone joining us for the first time — quick orientation. Every Monday, we break down two or three stories that actually matter in airline and travel distribution. Not the press release version. The strategic version — what it means, who benefits, what the risks are, and what the industry keeps getting wrong. For today's sources and references, see the episode details.

Eric Marketts

Three stories this week. And they form an argument, not just a list. First: new research from Accelya and Atmosphere Research quantifies the Offer and Order execution gap in a way that should make every airline commercial leader uncomfortable. Seventy-two percent of airlines say O&O transformation is a priority. Only twenty-seven percent have started. We'll break down what's driving that number and why it's not just slow execution.

Steph Nell

Second: Sabre's CEO publicly accused Amadeus of using its Altéa Passenger Service System to block competing Offer and Order solutions — and said Sabre is pursuing regulatory and legal options. The industry has known this dynamic exists for years. This is the first time someone at that level has named it out loud.

Eric Marketts

And third: OpenAI quietly pulled its "Buy Now" button from ChatGPT earlier this year. A lot of people read that as a retreat. We'll explain why it's actually a deadline — and why the O&O execution gap and the PSS lock-in from the first two stories are the exact reasons the deadline matters.

Steph Nell

The through-line across all three: the infrastructure debt isn't just a retailing problem. It determines whether your airline shows up when an agent goes looking.

Eric Marketts

Let's get into it.

Eric Marketts

Story one. In April of last year, Accelya and Atmosphere Research Group published a report on the state of Offer and Order transformation in the industry. Seventy-eight airline executives surveyed, twenty-eight phone interviews — American Airlines, British Airways, Emirates, others. The data landed with some attention when it came out. But Sabre's CEO said something from an earnings call this month that makes this report worth pulling out again — because what he named publicly is what this data had been quietly documenting for over a year . Sixty-six percent of airlines are now using NDC. Seventy-two percent identify Offer and Order transformation as important. But only twenty-seven percent — fewer than one in three — have taken substantive steps to actually begin the work .

Steph Nell

Say that again, because I want people to actually sit with it. Seventy-two percent say it's a priority. Twenty-seven percent have started.

Eric Marketts

That is a forty-five point gap between aspiration and action.

Steph Nell

And I don't think the instinct to explain that as corporate inertia or slow decision-making is the right one. Because the organizations being surveyed know what the stakes are. Ninety-one percent cite revenue generation as a key advantage. Eighty-one percent say it helps them quantify customer value. Seventy-nine percent connect it to attracting new customers. These aren't people who haven't thought through the business case.

Eric Marketts

So what is stopping them?

Steph Nell

The report projects NDC bookings tripling — from seven percent to twenty-one percent of total bookings in three years. EDIFACT GDS distribution declining by more than fifty-seven percent over the same period. The math is moving. And the window the report identifies for most carriers to actually complete the transition is 2028 to 2029.

Eric Marketts

Which is the bull case, in a way. If the adoption curve follows the pattern of every prior major transition — e-ticketing, GDS adoption, NDC itself — the execution phase compresses once a critical mass of early movers can demonstrate ROI. The twenty-seven percent who have started are building the proof cases. Once those numbers are public and consistent, the remaining seventy-three percent will move faster than the NDC adoption timeline did. The underlying revenue opportunity is real and it's been quantified. McKinsey puts ancillary uplift from personalization at ten to thirty percent without adding a single new product. That's not rounding error .

Steph Nell

That's the optimistic read. Here's the harder one. The carriers that haven't started O&O transformation by now aren't just behind on a timeline. They're behind on institutional knowledge, vendor relationships, pricing science, and data infrastructure — all of which take years to develop. By 2028, the retailing capability gap between airlines that started in 2024 and airlines that start in 2026 will be the widest it has ever been. And it compounds. The carriers moving fastest are also the ones attracting the most experienced talent and the most capable technology partners. That flywheel does not slow down .

Eric Marketts

So the real risk for the seventy-three percent isn't missing a project deadline. It's arriving at the starting line of the agentic era with infrastructure that can't support the game.

Steph Nell

That's the critical take. Because here's what the report doesn't fully surface, but the data implies: a lot of the seventy-three percent are not slow. They're stuck. There's a specific structural reason why Offer and Order transformation is hard to start — particularly for airlines on certain PSS platforms — and that reason has a name. We cover it in story two .

Eric Marketts

Sabre had its strongest financial quarter in more than two years. Seven hundred and sixty million dollars in revenue in Q1 2026. Up eight percent year over year. Adjusted earnings before interest, taxes, depreciation, and amortization is up twenty-one percent. By any measure, a good quarter to report.

Steph Nell

And Sabre CEO Kurt Ekert used the earnings call to say something that nobody at that level has said publicly before.

Eric Marketts

He accused Amadeus of monopolistic behavior in airline IT. The specific claim: Amadeus is using its control over the Altéa Passenger Service System to make it prohibitively difficult for airlines to implement competing Offer and Order solutions. The word he used was "trapped." Altéa customers who want to work with Sabre's OSD platform — Offer, Order, Settlement, and Delivery — or with any other O&O provider, cannot do so without significant friction imposed by Amadeus. Ekert said Sabre is, quote, "working on approaches to that from a regulatory and a legal standpoint ."

Steph Nell

To understand why this matters, you have to understand what Altéa is. Altéa is not just a booking system. It runs airline operations. Departure control. Check-in. Inventory. Ticketing. It is the operational nervous system for dozens of major carriers globally. Changing your PSS is not a software upgrade. It is a multi-year, nine-figure program. And the dependency Sabre is describing — where the PSS vendor controls the conditions under which you can engage with competing O&O platforms — is exactly the structural barrier that explains a meaningful portion of the twenty-seven percent execution rate from story one.

Eric Marketts

So the bull case here is real. If Sabre's regulatory or legal approach gains traction — if the EU or any competition authority takes a serious look at PSS interoperability requirements — it could force open the Offer and Order technology market in a way that the market hasn't been able to achieve on its own. Sabre's OSD platform, Accelya's FLX, and new entrants would all get a fair shot at airlines that today have no practical path to a different solution. That's a genuinely better market outcome. Airlines get choice. Competition drives better products and lower prices. The seventy-three percent who haven't started might find the path significantly less obstructed .

Steph Nell

The bear case is that this is a competitive dispute dressed in regulatory language — and regulators know it. Sabre's OSD commercial success depends directly on winning Altéa customers. Ekert's argument is not neutral. The EU examined GDS distribution practices for years and moved slowly. A formal antitrust investigation into PSS interoperability would take years to resolve. During that entire period, the status quo holds. And airlines waiting for a regulatory ruling before beginning their O&O transformation will still be waiting in 2030 .

Eric Marketts

And Amadeus's response tells you everything about how they view the situation.

Steph Nell

They didn't rebut it.

Eric Marketts

Not directly. Instead, Amadeus announced an expansion of its platform into biometric identity, artificial intelligence integrations, hospitality, and payments. The company is broadening its footprint across the entire airline technology stack — not just distribution, not just PSS, but every layer that sits between the airline and the passenger .

Steph Nell

When a market-dominant player declines to answer an antitrust accusation and instead announces expansion into adjacent markets, it is not deflecting. It is saying the original market is already won. The next one is what matters.

Eric Marketts

And the critical take here is about what layer they're moving to. Biometrics. Identity. AI. Payments. These are the same layers that Google, Apple, and Anthropic are approaching from the consumer side. Amadeus is building from the airline side toward the same junction. Whoever owns the interface between the airline and the AI agent — whether that's the identity verification, the payment rail, or the AI integration layer — owns the next generation of airline distribution. Sabre's antitrust play is a tactical fight over the PSS layer. The strategic fight is already moving somewhere else. And that is where story three is planted.

Eric Marketts

In March 2026, OpenAI quietly removed the "Buy Now" button from ChatGPT's travel interface. No announcement. No post-mortem. It was there, and then it wasn't. The stated reason: travel is too complex, and users weren't willing to complete transactions inside the chat interface .

Steph Nell

And a lot of people in the industry read that as vindication. See — agentic booking is still years away. The consumer isn't ready. The tech isn't ready. We have time.

Eric Marketts

Which is the wrong read entirely.

Steph Nell

Walk through the data.

Eric Marketts

IDC published this in January — from their FutureScape research on hospitality and travel. By 2030, thirty percent of travel bookings will be executed by AI agents. That's a formal IDC forecast. Four years from now. That is not a distant horizon . And alongside that, a January 2026 consumer survey of three hundred U.S. leisure travelers found that ninety percent are aware AI can help book travel — but only two percent are willing to let it book on their behalf. So awareness is nearly universal. Trust is almost nonexistent.

Steph Nell

Those numbers are not contradictory. They're sequential. The awareness is there. The willingness follows trust. Trust follows successful transactions. Successful transactions require infrastructure. The consumer is not the bottleneck. The infrastructure is.

Eric Marketts

The bull case for the OpenAI pullback is actually counterintuitive. OpenAI retreating from transactional AI is good news for airlines — it buys time. It forces the industry to hopefully build the trust infrastructure before the transaction volume arrives. The airlines and vendors that invest now in machine-readable offer and order data, reliable agent APIs, and clean post-booking servicing will be the ones whose inventory gets booked when the two percent becomes twenty percent .

Steph Nell

The bear case is that the industry is misreading the pullback as a reprieve rather than a warning . The "Buy Now" button didn't fail because agentic booking is conceptually flawed. It failed because the content underneath it wasn't ready. If a traveler asks an AI to book a flight and the agent finds three carriers — one with machine-readable, dynamically priced, ancillary-complete offer data, one with a partial NDC feed, and one with nothing structured at all — the agent books the first one. Not because the traveler chose it. Because the agent could complete the transaction. That is what airline invisibility looks like in practice.

Eric Marketts

And the critical take is precise. The same two problems from the first two stories are what's blocking agentic airline distribution right now. The twenty-seven percent O&O execution rate means most airline content is not machine-consumable at the level an AI agent needs to quote, book, and service a ticket. The PSS lock-in Sabre named publicly is a primary reason more airlines haven't fixed that yet. You cannot separate those three stories. They are one story about one problem with three different faces.

Steph Nell

And here's the math that makes this urgent. IDC's thirty percent by 2030 projection implies forty million or more bookings annually running through AI agent interfaces — if that number is even close to right. That transition starts now, not in 2028. Airlines that wait for the consumer trust curve to close before they build machine-readable infrastructure will be building it while the volume is already flowing around them to carriers that built it first .

Eric Marketts

The window to shape how AI agents access airline inventory — what data they see, what offers they can quote, what they're authorized to confirm — is open right now. Airlines that don't engage with that question in 2026 will be negotiating from weakness in 2028. Or they won't be negotiating at all, because the agent won't have found them.

Steph Nell

The agents are ready. The content isn't. That's not a technology statement. It's a strategy statement .

Eric Marketts

Three stories. One argument. Let me close it out and then Steph will add the last word.

Steph Nell

Go.

Eric Marketts

The Accelya data establishes the scale of the problem. Seventy-two percent of airlines say Offer and Order transformation is a priority. Twenty-seven percent have started. That gap is not corporate inertia. For a significant share of the seventy-three percent, it is structural — they are building on infrastructure that was not designed for the destination they need to reach, and the vendor that controls that infrastructure has not made the path easy.

Steph Nell

The Sabre-Amadeus dispute names the structure. The PSS lock-in that Sabre's CEO called out publicly on an earnings call is the same constraint that shows up in the Accelya data as unexplained execution failure. This isn't the first time the industry has known about it. It is the first time someone at that level has said it in public with legal action on the table. That changes the conversation, even if the regulatory outcome takes years.

Eric Marketts

And the agentic AI timeline provides the deadline. OpenAI's "Buy Now" pullback was a correction, not a cancellation. The consumer trust gap is real. The IDC 2030 projection is real. And the infrastructure that determines whether an airline's content is bookable by an agent in 2030 is the same infrastructure that needs to be rebuilt right now. The O&O execution gap isn't just a retailing lag. It's a visibility problem. Airlines that haven't started are not running behind. They are building toward a future in which they don't appear.

Steph Nell

If this episode was useful — share it with someone making decisions in distribution, retailing, or travel technology. And if you have a perspective on the Sabre-Amadeus dispute, the O&O execution timeline, or how your organization is thinking about agentic readiness — reach out. These conversations are better when more informed people are in 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 .