Reporting from the WTM 2024 Tech Summit

It is the industry’s trillion-dollar question, whispered in boardrooms and shouted over networking drinks: Is the global travel infrastructure on the verge of a breakthrough, or is it collapsing under its own weight?

At the recent Tech Summit, this question wasn’t just rhetorical—it was the subject of a raucous, Oxford-style debate that pitted the optimists against the realists. The motion on the table was stark: “The road is long and getting longer, friction is built in and the prospects for improvements are not good. Travel technology and airline technology is the cause. We need to fix it.” ¹

For the next hour, a panel of industry veterans tore into the machinery of modern travel. What emerged was a portrait of an industry at a crossroads. On one side, a vision of a frictionless future where Artificial Intelligence (AI) wipes the slate clean, rendering decades of obsolete code irrelevant. On the other, a sobering look at the “tech debt” and “process debt” accumulated over 80 years—a tangled web of 1940s bureaucracy and 1980s mainframes that no amount of shiny new software can easily fix.

The debate ended with a bold, almost desperate prediction: In five years, the mess won’t matter, because the “AI machine” will simply take over. But is that a prophecy or a fantasy?

The “Fried Chicken” Gap: A Tale of Two Systems

To understand why the industry is betting so heavily on AI, one must first understand the depth of the current dysfunction. The debate opened with a jarring contrast between the East and the West, delivered by Roman, a panelist who had just returned from a summer in China.

“I don’t know if anyone spent any time in China recently travelling… but the user experience for technology there is just unbelievable,” Roman told the packed auditorium². He described a travel ecosystem so seamless it makes Western infrastructure look antique.

“You go to on the high speed rail, there are no tickets. You just put your ID in the machine and that’s it. You board the train,” he explained³. But the real clincher—the anecdote that drew gasps of envy from the audience—was about fried chicken.

“When you’re on the train, you can order KFC… to be delivered to you in your seat on the train,” Roman said⁴. The logistics are perfectly synchronized: the delivery driver meets the train at a specific station, hands the food to the conductor, and it arrives at the passenger’s seat mid-journey⁵. “It’s simply unbelievable,” he concluded⁶.

Compare this to the Western experience. Roman recounted the story of a friend traveling with Norwegian Air who was denied boarding at the airport simply because the Online Travel Agency (OTA) he booked with hadn’t paid the bill⁷. The passenger, holding a valid confirmation, was left stranded by a backend failure invisible to him but catastrophic to his journey⁸.

“These kind of experiences are incredibly frustrating for the consumer and they can’t understand the reasons behind that,” Roman argued⁹.

This “Fried Chicken Gap” illustrates the fundamental problem. Emerging markets like China built their travel tech on modern mobile-first platforms. The West is building on top of mainframes from the Cold War era.

The Defense: It’s Not the Tech, It’s the Product

However, blaming the “tech” is too easy, according to Massimo, who stood up to defend the engineers. He argued that travelers take for granted the sheer computational miracle that occurs every time they hit “Search.”

“When was the last time you… walked into a store… and they gave you a paper ticket?” Massimo asked, reminding the audience of how far the industry has come¹⁰.

He peeled back the layers of a modern flight search to show what the technology is actually doing in real-time. It isn’t just looking up a bus schedule. “You browse through hundreds of thousands of flights, schedule different airlines, it gets availability for each of them. Check the inventory. You check each and every single booking class to find the right match,” he rattled off¹¹.

Simultaneously, the system is calculating taxes, checking government eligibility rules, and ensuring minimum connecting times are met—all in milliseconds¹². “And you keep on telling us that technology is broken and you can make booking in literally less than a minute,” he countered¹³.

When challenged on why it takes “nine months to develop an app,” Massimo revealed the crushing complexity developers face. “Do you know how many features there are in that app? There are like almost 150 features,” he said^14141414. “It’s not like a web shop on Amazon that has a catalogue there. You pick up a product, you put it in the cart and you’re doing a checkout”¹⁵.

This distinction is crucial. Amazon sells static products. Airlines sell dynamic, highly regulated service contracts subject to international law, security protocols, and volatile pricing models. The code isn’t broken; the requirements are just insanely difficult.

The Ghost of Rio, 1947

If the coders are heroes, as panelist Ian suggested, then who is the villain? The answer appears to be history itself.

“The airline industry developed its underlying processes starting in 1947 at the first IATA traffic conference in Rio de Janeiro,” Ian explained¹⁶.

Think about that. The fundamental rules governing how airlines interact, how tickets are issued, and how money moves were decided in a hotel room in Brazil 78 years ago, when “computers” were people who did math and airplanes had propellers.

“We’re still using a lot of those standards and processes,” Ian noted¹⁷. This is the concept of “Process Debt.” The industry isn’t just suffering from old code (tech debt); it is suffering from old thinking. We have built shiny mobile apps on top of business logic designed for paper tickets and carbon copies.

“We really need a root and branch reform of process,” Ian argued¹⁸. “It’s not the technology’s fault, it’s the businesses”¹⁹.

The 10.5 Million Fare Nightmare

Nowhere is this “Process Debt” more visible than in airline pricing. Mark, a Global Distribution System (GDS) expert on the panel, provided a shocking glimpse into the pricing insanity that legacy systems are forced to support.

“I worked for one carrier that had 10 million fares,” Mark revealed²⁰.

The airline attempted a “fare simplification program.” Their data showed that out of those 10 million fares, only 500,000 accounted for 99% of sales²¹. Logic would dictate deleting the other 9.5 million unused price points to speed up the system.

But the sales teams revolted. “The sales guy said, ‘well, no, you can’t take away my Seaman with one leg booking on the first Tuesday after a full moon fare,'” Mark quipped²².

The result? “So we ended up with 10,500,000 file fares,” Mark laughed²³. They actually increased the complexity.

This anecdote perfectly encapsulates the “Legacy Trap.” Mark admitted that while new standards like NDC (New Distribution Capability) are touted as the solution, they often just add “a layer on top” of the existing mess rather than fixing the foundation²⁴. “We want the airlines to give us offers that are retail ready… but NDC is not a standard… it’s a schema,” he argued²⁵. Without true interoperability, the industry is just digitizing chaos.

The Invisible Handcuffs: Regulation and “Common Use”

Even when airlines want to innovate, they are often physically blocked from doing so by the very infrastructure they rely on: airports and governments.

Stephen, another panelist, argued that the industry is handcuffed by regulation. “We would love to do a whole heap of stuff above and beyond what we currently do today, but we’re stopped by governments around the world,” he said²⁶.

He pointed to the chaos of the pandemic as a prime example. “Every country had a different Covid regime… How can you do that? Every country having a completely different… regime,” he asked²⁷. This fragmentation forces tech teams to build custom solutions for every jurisdiction, draining resources that could be spent on innovation.

Furthermore, the physical hardware at airports—the kiosks and screens—is often off-limits. Stephen criticized the concept of “Common Use” (shared airport terminals) as a “straightjacket”²⁸.

“Because of common use, I can’t customize, I can’t treat my customers any different from Massimo’s,” Stephen complained²⁹. “We’re not permitted to put our own piping in the airport because it’s owned by SITA or some other provider”³⁰.

In this environment, an airline might have the best app and the best backend tech, but if they can’t install their own sensors or data hubs in the terminal, the seamless experience breaks down the moment the passenger enters the airport.

The Hail Mary: AI as the Great Compressor

So, the industry is stuck. It is burdened by 1947 rules, managing 10 million phantom fares, blocked by government red tape, and physically restricted by airport monopolies. The “To-Do” list to fix this seems insurmountable.

Enter Artificial Intelligence.

Toward the end of the debate, Christian, a panelist who compared the polarized discussion to the US election (“Are you for Donald Trump or against Donald Trump?”), offered a radical solution³¹. He suggested that we might never need to untangle the knots of the past because AI will simply cut through them.

“Five years from now, all this legacy technology will be overturned and be run by the AI machine,” Christian predicted³².

His theory relies on the concept of data compression. “This whole room, we have 37,000 something people talking to each other right now… passing data backwards and forwards. AI compresses all of this data,” he explained³³.

Christian’s argument is that the complexity of travel—which he noted is unique because “every single transaction is an international transaction” involving cross-border payments, cultures, and languages—is perfectly suited for AI^34343434. Humans and rigid SQL databases struggle to make sense of millions of variables. AI thrives on it.

If this vision holds true, the “10.5 million fare” problem solves itself. The AI doesn’t need a simplified list; it can ingest all 10 million options and find the “Seaman with one leg” fare in a microsecond, without a human ever needing to code a specific rule for it.

The “Retail” Dream: Searching by Desire

If AI fixes the backend, it could also revolutionize the front end, delivering the “Retail Dream” that Stephen alluded to.

Current travel search is incredibly rigid. You must tell the machine where you are going and when. But that isn’t how humans dream of travel.

“I want to be able to move away from origin, destination and a date to how much money have you got in your pocket and where do you want to go? What are your options?” Stephen said³⁵.

Imagine a search bar that simply asks: “I have $2,000 and I want to surf in February.” Today’s legacy systems choke on that query because they require specific airport codes and dates to check inventory. An AI layer, however, could interpret that intent, scan the entire global inventory, and package a solution. It bypasses the “rigid old search logic” entirely³⁶.

The Counterpoint: Automating the Mess

However, techno-optimism only goes so far. As the debate wound down, a sense of realism settled back over the room. The moderator, Linda, pointed out that despite the soaring rhetoric about AI, the “Against the Motion” team (those defending the current tech) had “failed miserably” in the eyes of the audience³⁷. The crowd voted overwhelmingly that the system is indeed broken.

The skepticism regarding an AI bailout is well-founded. AI models are trained on data. If the travel industry’s data is filled with “hanging chads” (a reference made by the moderator to the corruption of the voting process) and “boat rigging,” then the AI will simply learn to replicate those errors^38383838.

You cannot automate a mess. If an airline’s underlying process for boarding a plane or pricing a seat is fundamentally illogical (like the 1947 rules), putting an AI on top of it might just result in passengers getting bad information faster. As Mark pointed out with his fare example, the complexity isn’t just data—it’s political and human. An AI cannot convince a sales director to delete a legacy fare that he is emotionally attached to.

The Bill is Due

The Tech Summit debate didn’t end with a handshake, but with a warning. “I think we’ve got a long road ahead of us,” the moderator concluded³⁹.

The industry is currently in a race. On one track, the accumulation of “Tech Debt” and “Process Debt” is slowing innovation to a crawl, creating the friction that leaves passengers stranded and systems crashing. On the other track, AI is advancing at breakneck speed, offering a potential shortcut to modernization.

“There is tech debt out there that we’re all going to have to pay,” the moderator warned⁴⁰.

For now, the industry is betting the house that AI can pay that bill. But until that “AI Machine” arrives, travelers are left navigating a 2025 world with a 1947 map, hoping their app doesn’t crash before they can order their fried chicken.