Amadeus Moves Upstream: The Booking Was Never the Point

Eric Marketts

Last week we ended on three words. Own it or rent it. The model got cheap, we said, so your edge moves to the harness and the data only you own. Build the offer engine, own the order, own the moment the traveler decides. And we left you with a question. Will airlines build it before someone else owns the layer for them .

Steph Nell

Right. And this week the industry answered a harder version of that question. Because there is an assumption buried inside "own the harness." The assumption is that the thing the harness produces, the booking, is still where the money is. Three moves this month say it is not.

Eric Marketts

That is the episode. For fifteen years the entire NDC fight was about one thing. Who controls the offer. The airline or the GDS. That was the war. This month, three separate players made moves that tell you the industry already moved past that war while a lot of carriers were still fighting it .

Steph Nell

The new fight is on both sides of the booking. Who controls the demand before the offer. And who controls the agent after it. The booking in the middle, the segment, the thing every fee in this industry was built around, is becoming the least valuable part of the chain.

Eric Marketts

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

Steph Nell

And I'm Steph Nell. Quick orientation if you're new. Every Monday we take the two or three things that actually moved in airline and travel distribution and tell you what they mean. Not the press release. The strategic version. Links and sources are in the episode details.

Eric Marketts

Three stories today, and they fit together into one argument. Story one. Amadeus launched an advertising platform and basically admitted the GDS pipe was never the asset. Story two. Google published a commerce protocol that turns the AI agent into the booking channel, and the first travel players are already live on it. Story three. Lufthansa quietly put a single order ID into production, and it is the unglamorous thing that makes the other two stories actually possible.

Steph Nell

Demand on one end. The agent on the other. And the order structure underneath that decides whether you can play in either. Let's get into it .

Eric Marketts

Story one. Amadeus held an Advertising Summit in Nice and introduced something called the Amadeus Travel Advertising Platform. It's AI-powered, it's built on an agentic campaign framework they developed with Accenture, and the pitch is simple to say and enormous in what it implies. They want to act on a traveler's demand signals before that traveler ever opens a booking engine. Before search becomes a transaction .

Steph Nell

And the tell is who Amadeus named as the competition. Not Sabre. Not Travelport. They pointed at Google, Meta, and The Trade Desk. An advertising-technology competitive set. A company that built a fifty-year empire on the booking segment just stood up and said our rivals are the ad platforms.

Eric Marketts

Let's slow this down, because I think a lot of people hear "advertising platform" and file it under marketing and move on. Steph, walk through what Amadeus is actually selling here, because it is not banner ads.

Steph Nell

It is not banner ads. So picture the travel funnel. At the very top, somebody is dreaming. They are scrolling, they are searching a destination, they are reading about Tokyo. In the middle, they go to a booking site and they shop. Flights, dates, fares. At the bottom, they buy. A seat, a segment, a booking. For decades, Amadeus made its money at the bottom of that funnel. Every time a booking completed through its system, a segment fee. That was the business .

Eric Marketts

The classic GDS model. Get paid per segment.

Steph Nell

Per segment. Now here is the thing about the top of the funnel, the dreaming and searching part. All of that activity throws off signals. What destinations are heating up, which routes are filling, how price-sensitive a given market is this week, when demand is about to spike. And those signals flow through Amadeus's pipes constantly, whether or not anyone ever completes a booking. For fifty years Amadeus treated those signals as exhaust. Byproduct. The booking was the product. What they just announced is the decision to package the exhaust and sell it. To monetize the intent that flows through the system instead of waiting for the booking at the end of it .

Eric Marketts

So the product is the demand signal itself. Sold to whom?

Steph Nell

Sold to anyone who wants to reach a traveler earlier and more precisely than they can today. Airlines, hotels, tourism boards, anyone whose whole problem is capturing a traveler before a competitor does. Amadeus is saying, we can see demand forming before you can, so let us help you act on it. That's the platform.

Eric Marketts

Okay. Give me the bull case, because there is a real one.

Steph Nell

There's a strong one. This is Amadeus finally being honest about where its durable advantage actually lives. The moat was never the EDIFACT plumbing. Anyone can move a message. The moat was decades of search patterns, booking trajectories, and demand forecasts that no pure-play ad platform can match for travel-specific precision. Google knows a lot about you. Google does not know that this specific city pair is about to see a demand spike in the next ten days the way Amadeus does. So if Amadeus opens a higher-margin revenue line selling demand intelligence, that is a line that grows even as the per-segment booking economics get competed down to nothing. They're diversifying off the segment fee before the segment fee collapses. Strategically, that's the right read of their own business .

Eric Marketts

And the bear case.

Steph Nell

The bear case is that Amadeus is walking onto Google and Meta's home field with a fraction of their consumer reach and a fraction of their ad-tech maturity. Travel advertising dollars already concentrate on those platforms. Marketers already know how to buy Google and Meta. Amadeus is asking them to learn a new platform with a smaller audience. And there's a second, sharper problem. A GDS selling demand intelligence to airlines, at the exact moment those airlines are fighting to own the direct relationship with their own customers, can look like the intermediary trying to stay relevant by selling the airline data the airline thinks it already owns. That's a tough room .

Eric Marketts

So what's the take an airline commercial team should walk out with?

Steph Nell

The critical take is that this is the clearest signal yet that the GDS incumbents expect the booking transaction to be commoditized. Read it that way. When the company that built its empire on the segment fee starts selling everything that happens before the segment fee, it is telling you exactly where it thinks the margin is going. And the margin is moving away from the booking. The race is no longer to own distribution. Distribution is becoming a utility. The race is to own intent, the demand, the moment of desire upstream of the transaction. Airlines should not read this as Amadeus launching an ad product. They should read it as Amadeus hedging against its own core business, and ask themselves the obvious question. If the company that runs the pipe is betting the booking gets commoditized, what does that mean for an airline that still treats the booking as the prize .

Eric Marketts

Hold that, because story two is what happens on the other side of the booking. If Amadeus is moving upstream of the transaction, Google just moved to own the transaction itself.

Steph Nell

Demand on one end. The agent on the other. Same squeeze on the booking in the middle .

Eric Marketts

Story two. Google published the Universal Commerce Protocol. UCP. And before anyone glazes over at the word protocol, this one matters, and we need to correct something while we're here. Steph, last week on this show we referred to UCP and tied it to Amadeus. Set the record straight, because the ownership of this thing is the whole story .

Steph Nell

We did, and I want to be precise about it because precision here changes the strategic picture. UCP is Google's protocol. Google built it, co-developed with Shopify, and it is backed by an enormous coalition. Amazon, Microsoft, Meta, and on the payments side Stripe, Visa, and Mastercard. Amadeus is in the picture, but as a foundational partner helping adapt UCP for lodging. Amadeus is a participant. Google is the author. And when you're talking about who writes the standard the whole industry runs on, author versus participant is the entire game. So last week's framing pointed at the partner. The thing to watch is the author .

Eric Marketts

So what is UCP, in plain terms?

Steph Nell

It is an open standard for agentic commerce. Let me unpack both words because they both carry weight. Agentic means an AI agent acting on your behalf, not just answering you. And commerce means it doesn't stop at recommendation, it completes the purchase. So UCP is a common language that lets an AI assistant run the entire buying journey end to end. Catalog search, building a cart, linking your identity, checkout, payment, and post-purchase order management. The whole thing, in one conversation, no forms, no booking-engine page .

Eric Marketts

Now we've talked on this show about MCP, the Model Context Protocol. How is UCP different? Because I think people are going to blur them together.

Steph Nell

Good, and they are different in a way that matters. Think of MCP as the protocol that lets an agent read and understand. It connects the model to live information and tools so the agent can see your inventory, your fares, your rules. MCP is about context. It's about the agent knowing things. UCP is the layer on top that lets the agent act. Transact. Move money, create the order, own the checkout. MCP lets the agent read the menu. UCP lets the agent place the order and pay the check. And Google didn't stop at theory. In March they shipped a major update that added cart support and product catalog access, so the protocol can now hold a real shopping cart, not just a single item. Direct UCP checkout is live in the US, with global rollout planned later this year .

Eric Marketts

And travel specifically?

Steph Nell

Google has confirmed flights and hotels are next, after restaurants and event tickets. And it's not waiting on the airlines to be ready. In parallel, a startup called Mindtrip launched a conversation-led flight booking experience built on Sabre's AI-native air APIs, with PayPal handling the agentic payment. So you can already see the shape of it. A traveler talks to an agent, the agent shops live air content through Sabre, and PayPal settles the transaction inside the conversation. The booking moved from a human staring at a booking engine to a machine-to-machine flow the traveler never sees .

Eric Marketts

Okay, the bull case. Because an open protocol sounds like exactly what this fragmented industry has been begging for.

Steph Nell

And there's a real one. A common protocol collapses what I'd call the integration tax. Right now, if you're an airline and you want to be bookable inside ten different agents, you potentially build ten different bespoke connections. That's the N-by-M problem, every supplier wiring to every channel by hand. It's expensive, it's slow, it's why distribution has been a mess for thirty years. If instead you expose your catalog, your cart, and your order management once, through UCP, then any compliant agent, Google's, an OTA's, a startup like Mindtrip, can transact with you without a custom build. You reach demand wherever it forms. The traveler gets genuinely frictionless booking. In the best light, this is what NDC promised at the data layer, finally extended to the entire purchase .

Eric Marketts

And the bear case. Which I suspect is large.

Steph Nell

It's large. Google writing the commerce protocol for travel is the exact disintermediation risk airlines have feared for fifteen years, just with much better packaging. Walk the logic. If the agent owns discovery, owns the comparison, and owns the checkout, then the airline becomes a UCP-compliant supplier sitting behind Google's interface. Your brand gets stripped down to a row in an agent's result set. Your upsell, your bundles, your merchandising, all of it gets controlled by someone else's ranking logic. You spent a decade fighting the GDS to control the offer, and now the offer renders inside an agent you don't own, ranked by rules you don't set. Ceding the protocol means ceding the merchandising surface, which is the thing modern retailing was entirely about .

Eric Marketts

So somebody inside an airline hears this and says, fine, we'll just sit it out until it matures. What do you tell them?

Steph Nell

The critical take is this. NDC was a fight over who controls the offer. UCP is a fight over who controls the agent. Those are different fights, and a lot of carriers are about to lose the second one because they're still congratulating themselves on the first. Airlines spent ten years and a fortune building offer-and-order capability specifically to escape the GDS's grip on merchandising. And they are now on the verge of handing the new merchandising surface, the agent, to whoever writes the agentic standard. So the strategic question is not, should we support agents. That's settled, the travelers will decide that for you. The real question is, do we shape the protocol or do we implement someone else's. And here's the part people miss. Sitting out is not neutral. Sitting out is a vote for Google's version. The protocol gets written whether you participate or not. The only choice you control is whether your interests are represented in it .

Eric Marketts

And the place where airlines could have a voice in shaping it, IATA, the carriers themselves, is exactly where the energy isn't right now.

Steph Nell

That's the worry. The standard-setting body for this could be the airline industry. Instead the standard is coming out of Mountain View, and the airline industry is debating it after the fact .

Eric Marketts

Story three. And on paper this is the least exciting headline of the three, which is exactly why we saved it for last. Lufthansa Group and Amadeus rolled out a single Order ID. It unifies a booking and all its ancillaries into one record, aligned to the IATA ONE Order standard, and it is now running in production across multiple carriers and millions of passengers. The press framing is operational simplification. Steph, the press framing is underselling it badly .

Steph Nell

Badly. Because what sounds like an IT housekeeping project is actually the gate that the first two stories have to pass through. Let me explain why, and to do that I have to explain the thing it replaces. The PNR .

Eric Marketts

The Passenger Name Record. The thing the whole industry has run on forever.

Steph Nell

Since the 1960s. The PNR was designed for a world of travel agents and reservations. And here's its core problem in a modern retailing context. A trip doesn't liv in one clean place. You've got the PNR, then you've got a separate e-ticket, then ancillaries like a bag or a seat get recorded in yet another structure, sometimes as a separate document entirely. So a single traveler's single trip is scattered across multiple records that were never designed to function as one object. For a human agent in 1995, fine. For an automated system in 2026, that fragmentation is poison .

Eric Marketts

Why poison specifically? Make it concrete.

Steph Nell

Okay. Say an AI agent books you a flight, a premium seat, and a bag, and then your plans change and you need to rebook. For the agent to do that cleanly, it has to read the entire state of your trip, understand what you bought, modify it, re-price it, and re-issue it, all without a human untangling four different records and reconciling them by hand. If your trip is scattered across a PNR, an e-ticket, and separate ancillary documents, the agent hits a wall. It can't reliably see the whole picture, so it can't safely act on it. ONE Order fixes that at the root. It collapses the airline's order — the flight and every air ancillary, the bag, the seat, the lounge pass — into one record with one ID. Now, the standard's ambition goes further. On paper that same order can hold a hotel, a transfer, a tour. But that is the vision, not the deployment. What Lufthansa actually put into production is air and air ancillaries. The full-trip order, hotel and ground and activities living inside one airline order, is the next frontier. We are not there yet, and we should be honest about that..

Eric Marketts

And that's the connection to stories one and two.

Steph Nell

That's the entire connection, and it's why this is the load-bearing story. Go back to Amadeus selling demand intelligence in story one. And Google running agentic checkout in story two. Neither of those works against an airline whose orders aren't machine-readable. An agent cannot complete a clean UCP transaction against a trip that's smeared across four legacy records. Amadeus cannot help you act on demand in real time if your own order structure can't be read and modified programmatically. A clean, unified order is the precondition for all of it. It's the admission ticket .

Eric Marketts

So give me the bull case on ONE Order.

Steph Nell

The bull case is that Lufthansa just proved order-based retailing works at industrial scale. Not in a lab, not in a slide, not in a single-route pilot. In production, across multiple carriers, across millions of passengers. One record for the trip and every ancillary means cleaner servicing, far fewer reconciliation failures, and crucially a structure an AI agent can actually read, modify, and rebook against. So a carrier with this foundation can walk straight into a UCP-style agent flow and offer real-time dynamic upsell, while everyone else is still trying to untangle PNRs. Lufthansa didn't just modernize plumbing. They built the on-ramp to every other game on this episode .

Eric Marketts

And the bear case.

Steph Nell

The bear case is that this migration is brutally expensive and brutally slow, and most of the industry is years behind Lufthansa. So you get a two-tier world. A small handful of order-ready megacarriers who can play in advertising and agentic commerce and dynamic offers, and a long tail of airlines still stitched to legacy PSS, locked out. And here's the cruel timing problem for that long tail. The migration cost may land on them at the exact moment the agentic shift commoditizes the booking they're spending all that money to modernize. They pay full price to rebuild the infrastructure right as the margin in the thing it produces gets competed away. That's a genuinely bad spot, and it's where a lot of mid-size carriers are going to find themselves .

Eric Marketts

So what's the take that ties the whole episode together?

Steph Nell

The critical take is that headlines one and two are impossible without headline three, and almost nobody is framing it that way. Amadeus cannot monetize demand and Google cannot run agentic checkout against an airline whose orders are not machine-readable. Order-based retailing has been sold to airline boards as an internal efficiency project. Clean up the back office, reduce reconciliation cost. And that framing dramatically undersells what it is. It is not plumbing. It is the gate to every downstream value pool there is, advertising, agentic commerce, dynamic offers, all of it. So the airlines treating ONE Order as an IT cleanup are misreading the single most strategic investment on their roadmap. The ones who get it are treating order modernization as the thing that decides whether they get to compete at all in the next five years .

Eric Marketts

Let me put the three together, because individually they're interesting and together they're an argument. Value is leaving the booking transaction from both ends. Amadeus is moving upstream to own the demand before the offer. Google and the agents are moving downstream to own the checkout after it. And order-based infrastructure, the Lufthansa story, is the only thing that lets an airline play in either game. Read together, they describe a distribution stack where the segment, the booking, the thing this entire industry priced itself around, is becoming the commodity. And everything adjacent to it is the prize .

Steph Nell

And the uncomfortable part is what it means for where you put your attention. For fifteen years the strategic conversation in this industry was who controls the offer. That conversation is basically over, and a lot of carriers are still having it. The booking was never the point. The point was always the demand on one side and the agent on the other, and the structured order underneath that lets you reach both. The airlines that win the next phase are the ones whose order infrastructure is machine-readable enough to plug into both ends, and who understand that the transaction in the middle is the least valuable thing they own .

Eric Marketts

Two things to watch this quarter. First, watch which airlines start talking about ONE Order as a commercial strategy instead of an IT project. That language shift is the tell for who actually understands the stakes. And second, watch whether IATA or the airlines themselves try to shape the agentic commerce protocol, or whether they let Google define it by default. The next couple of quarters decide whether carriers shape the agent layer or just implement it . And if this sharpened your thinking, send it to someone making distribution, retailing, or AI decisions in the next twelve months. These are 2026 calls, not 2028 calls.

Steph Nell

And if you take one thing into your Monday, take this. Audit your order. Not your NDC roadmap, not your loyalty program. Your order structure. Because every other move on this episode, monetizing demand, plugging into agents, dynamic upsell, every one of them assumes a machine can read what your customer bought. If it can't, none of the rest is available to you, no matter how good your strategy deck looks .

Eric Marketts

And one quick ask before you go. The V1 Airline Retailing Report is available on every major platform, Apple Podcasts, YouTube, and Spotify. So wherever you're listening or watching, hit like, follow, and subscribe. It genuinely helps. Every follow and every rating pushes the show in front of more people building airline retailing, and it's what lets us keep producing this every single week. So follow the show, leave us a quick rating, and you'll never miss a Monday .

Steph Nell

And if you're inside a carrier doing the hard version of this, getting your order onto a single ID, deciding whether to engage with UCP, we want to hear how it's actually going. That's the conversation the industry needs to be having out loud.

Eric Marketts

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

Steph Nell

I'm Steph Nell. Thanks for listening.

Eric Marketts

The booking was never the point. Stay sharp out there .