The lucky dip of stumbling on a cheap seat on a popular flight is starting to disappear. Airlines are handing their pricing to artificial intelligence, and on busy routes that mostly means one thing: higher fares. The days when a savvy traveller could swoop in at the last minute and grab an unexpected deal are fading as sophisticated algorithms take over the cockpit of revenue management. What was once a game of guesswork and rule-of-thumb pricing is now a high-stakes data-driven operation that leaves little room for bargains.
Carriers have long priced seats with analysts and rules of thumb, such as bumping fares by 20 per cent once a flight is a quarter full. This traditional approach relied on historical data, simple triggers, and a team of specialists who could only react to changes after they happened. AI replaces those spreadsheets with models that weigh dozens of variables in real time and adjust continuously to demand. The result is a far more dynamic and precise pricing engine that constantly recalibrates fares based on everything from booking patterns and competitor moves to weather forecasts and even social media sentiment.
How AI pricing works
At the heart of this transformation is machine learning, a form of artificial intelligence that allows computers to learn from data without being explicitly programmed. In the context of airline pricing, these models ingest massive amounts of information—historical ticket sales, current search volume, seat availability, route popularity, economic indicators, fuel costs, and even the browsing behaviour of anonymous potential customers. The algorithms identify patterns and correlations that human analysts would never spot, allowing airlines to predict with remarkable accuracy how much a given passenger is willing to pay at any moment.
Unlike the old rules of thumb, AI-driven pricing is not static. A fare can change multiple times a day, sometimes within minutes, as the system reacts to new bookings, cancellations, and market shifts. For example, if a competitor drops its price on the same route, the AI might immediately adjust to remain competitive—or it might hold firm if it detects that demand is strong enough to justify a higher fare. This continuous recalibration is a key reason why the traditional bargains that used to appear on busy routes are becoming increasingly rare.
The shrinking price gap
The effect of this technology is to shrink the pricing gaps that once let travellers find bargains. In the past, airlines would often leave certain fare buckets open for late-booker discounts or special promotions, hoping to fill seats that might otherwise go empty. But AI models are far more adept at estimating exactly when a seat will be sold and at what price. They sell fewer seats below what they think you will pay, and pack planes closer to capacity. On popular routes, where demand is consistently high, the algorithms quickly identify that almost every seat will be filled regardless of price, so they have little incentive to offer deep discounts.
This trend is not hypothetical. Airlines from Delta to Virgin Atlantic are adopting the tools to squeeze more from every flight. Delta, one of the largest carriers in the world, has publicly acknowledged that it is investing heavily in AI-based revenue management. Virgin Atlantic, the UK-based carrier, has also deployed similar systems, and both report that the technology is helping them improve load factors and yield. The result is that passengers on premium business routes are seeing fewer flash sales and last-minute deals, while the gap between the cheapest and most expensive economy seats narrows.
Cheaper too, sometimes
It is not all bad for flyers. The same models can cut fares on quieter routes to fill empty seats, so off-peak and low-demand flights could get cheaper. If a route is struggling to attract passengers, the AI may slash prices aggressively to stimulate demand, sometimes offering fares well below what a human analyst would have authorised. This is because the marginal cost of carrying an extra passenger is low, and the algorithm understands that a half-empty plane is a lost opportunity. For flexible travellers willing to fly at odd times or to less popular destinations, there are still deals to be found—and sometimes even better ones than before.
According to Bryan Terry, an aviation consultant at Alton Aviation Consultancy, consumers should expect that airlines will be smarter about their pricing. They will “exploit that capability to raise fares where possible and cut prices where they have room to stimulate demand.” In other words, the AI is designed to maximise revenue, not to be fair to passengers. It will squeeze the maximum from those who are willing to pay more, while offering just enough low fares to keep the planes full. Whether that ultimately helps or hurts travellers depends on the route, the timing, and the individual's ability to be flexible.
The secret sauce: AI pricing platforms
Israeli startup Fetcherr is one of the firms driving the shift. Its platform is used by nearly a dozen carriers, including Canada’s WestJet and Brazil’s Azul. The company's AI models are designed to simulate thousands of pricing scenarios in real time, allowing airlines to respond instantly to changes in demand, fuel costs, and competitive actions. During recent Middle East disruptions, Fetcherr's system repriced flights worldwide in response to oil swings, cancellations and shifting demand, demonstrating the power of AI to handle complex, fast-moving situations that would overwhelm human analysts.
“Our models analyse dozens if not hundreds of classes of variables to come up with fares. You can only now do that because of AI,” said co-founder Uri Yerushalmi. The company says it lifts revenue mainly by filling more seats, not by raising ticket prices. This is an important distinction. By pricing more accurately, the AI can attract price-sensitive customers who would have been scared off by a standard fare, while still capturing higher revenue from business travellers who book late. In theory, this is a win-win: the airline fills more seats, and some passengers get lower fares than they would have under the old system.
AI is also extracting value even after you buy your ticket. Volantio, a software company used by Japan Airlines, spots passengers who might swap a busy flight for a voucher. The system identifies passengers who are likely to be flexible—perhaps those who have low-cost fares or non-refundable tickets—and offers them an incentive to move to a less crowded flight. This frees up a seat on the busy flight, which the airline can then resell to a last-minute business traveller for $1,000 or more. The original passenger gets a voucher and a new flight, and the airline captures revenue it would otherwise have missed. This practice, known as voluntary denied boarding, is becoming more sophisticated with AI, and similar tools are moving through hotels and booking platforms.
The worry: surveillance pricing
The bigger fear is where this goes next: fares aimed at each traveller’s highest “willingness to pay,” what analysts bluntly call the “pain point.” Consumer advocates and US lawmakers have warned that airlines could use generative AI for “surveillance pricing”, charging different people different fares for the same seat based on data such as browsing history or income. Imagine boarding a flight where the person sitting next to you paid half as much for their ticket simply because they used an incognito browser or live in a different postal code. This is not just a hypothetical worry—it is the logical extension of a technology that is already being deployed.
The concern is that AI could combine real-time market data with personal information gleaned from cookies, loyalty programs, and credit card transactions to build a detailed profile of each passenger. This profile would allow the airline to estimate the maximum fare each individual would be willing to pay, and set the price accordingly. A business traveller who always books at the last minute could be charged a premium, while a budget-conscious tourist might get a discount—but the tourist would never know that they are paying more than they need to, because the price is hidden and personalised.
That is no longer purely hypothetical. The US Federal Trade Commission has opened a civil investigation into whether airlines use individualised data profiles to push up prices, as reported by Simple Flying. The investigation aims to determine if airlines are engaging in deceptive or unfair practices by using personal data to set fares without transparency. Some states have moved first, with Maryland passing a Protection from Predatory Pricing Act, which prohibits businesses from charging different prices based on personal data like income level or location. Regulators elsewhere are circling too, after cases like China’s Trip.com fine for hidden pricing practices.
For now, the companies say they are not there yet. Delta, which declined to comment, has publicly denied setting fares using personal information. Fetcherr says its models use aggregated market data, and Volantio says its offers are not personalised. The distinction is important: aggregated data might include broad trends like search volumes and booking patterns, but it does not identify individual passengers. However, the technology to move from aggregated to personalised is already available, and it is only a matter of time before some airline tests the boundaries.
The broader implications
The shift to AI-driven pricing is not confined to airlines. Hotels, car rental companies, and even online travel agencies are adopting similar tools. The same machine learning algorithms that optimise airline fares are being used to set hotel room rates, dynamic pricing for ride-sharing, and even surge pricing for concert tickets. This is part of a broader trend toward algorithmic pricing across the economy, where businesses use data to maximise revenue in real time. While this can lead to greater efficiency and lower prices in some cases, it also raises concerns about fairness, transparency, and consumer trust.
For the travel industry, the immediate impact is clear: the old trick of hunting for a hidden bargain fare is quietly getting harder. As AI becomes more sophisticated, the price you see is increasingly the price the algorithm has determined you are willing to pay. The games that travellers used to play—clearing cookies, searching on different devices, or booking at odd hours—are becoming less effective. The AI is learning to see through these tactics, and it is getting better at extracting the maximum from each transaction.
There is also a human cost. Revenue management analysts who once built spreadsheets and manually adjusted fares are seeing their roles evolve. Some are being retrained to oversee and tune the AI models, while others are being replaced altogether. The skills required to work in airline pricing are shifting from statistics and intuition to data science and machine learning. This is a reflection of the broader transformation that AI is bringing to many industries, where the role of the human expert changes from doing the work to supervising the algorithms that do the work.
Despite the concerns, the airlines argue that AI pricing is not inherently bad for consumers. By making pricing more granular and responsive, they can offer lower fares to price-sensitive customers who would otherwise stay home. A family saving up for a vacation might secure a bargain on a midweek flight that the old pricing system would have left too expensive. A student travelling with a flexible schedule might find a deeply discounted fare to fill an otherwise empty plane. The challenge is ensuring that these benefits are shared broadly, rather than being captured entirely by the airlines.
Regulators are beginning to take notice. The FTC investigation in the United States, the Maryland law, and similar actions in other countries signal that the issue is on the policy agenda. The question is whether current consumer protection laws are sufficient to address the risks of AI-driven pricing, or whether new regulations are needed. Some lawmakers have called for transparency rules that would require airlines to disclose when a price has been personalised, as well as safeguards against discrimination. Others argue that the market will self-correct, as competition between airlines and consumer choice will keep pricing in check.
In the meantime, travellers should be aware of what is happening behind the scenes. The price you see for a flight is no longer the result of a simple formula; it is the output of a complex algorithm that is constantly learning and adapting. The best strategies for finding a good deal are to be flexible, book early or very late, and to compare prices across different channels. But even these strategies may not be enough to beat the AI. After all, the airlines are investing billions of dollars in these systems because they work, and they work because they are able to capture more revenue from every passenger.
The future of airline pricing is likely to be more sophisticated, more personalised, and more opaque. Some experts predict that we will move toward a model where every passenger pays a unique fare, based on a combination of their willingness to pay, their travel patterns, and the competitive dynamics of the route. This may sound like science fiction, but the building blocks are already in place. The AI systems that are being deployed today are the first step toward that future, and they are already changing the way we buy plane tickets. As the technology continues to evolve, the lucky dip of stumbling on a cheap seat will become a nostalgic memory, and the humble act of booking a flight will become yet another arena where artificial intelligence and data rule the skies.