As the airlines increasingly let artificial intelligence determine prices in real time, opportunities for cheap tickets on busy flights are diminishing.
Those who are chasing cheap flight tickets will have more and more difficulty: the time when you accidentally ran into a cheap spot on a popular flight seems to come to an end, as airlines turn to artificial intelligence to determine their prices.
For years companies were based on analysts who developed billing rules, such as raising fares by 20% once they filled a quarter of the seats. AI now enables them to adjust prices much faster, assessing dozens of real-time variables. Thus they increase their income and at the same time close the billing gaps that formerly allowed travelers to find opportunities.
With a key incentive to launch operating costs, companies from Delta Air Lines to Virgin Atlantic increasingly adopt this technology. For passengers this probably means higher fares on busy lines: fewer seats will be sold below what companies estimate the traveler is willing to pay, with flights operating closer to their maximum completeness. Traditional rules and spreadsheets give way to predictive models that constantly adjust prices according to demand.
«Consumers must expect that airlines will become smarter at their pricing and take advantage of this opportunity to increase prices where they can and reduce them where they have room to stimulate demand», said Bryan Terry, analyst of the New York-based advisory firm Alton Aviation Consultancy.
«The airlines will understand these conditions more clearly, earlier and with greater certainty, which allows them to adapt pricing, both upwards and downwards, with greater dynamics.»
Technology can, however, also favour travellers on flights with less demand. On off-peak lines, passengers may find lower fares, as companies will use AI to fill the empty seats, according to Terry. The trend occurs at a time of increased costs in labour force, maintenance and fuel, which forces the industry to draw more revenue from each flight. Established software providers such as Amadeus and Pros now offer AI-based pricing tools.
Whether all this ultimately benefits the traveler remains controversial. «Profitable purchases favour passengers, as with smart or timely booking they can secure prices below average», said aviation analyst Guy Leitch. Critics object that companies operate with narrow profit margins and will use the strongest tools to maximize prices, completeness and revenue.
One of the companies leading the change is Israeli startup Fetcherr, whose platform is used by nearly twelve airlines worldwide, among them Canadian WestJet and Brazilian Azul Airlines. During the unrest caused by the conflict in the Middle East, its software immediately incorporated factors such as fluctuations in the price of oil, the departure of competitors from markets and cancellations of flights, to re-price flights around the world. On average, as the company reports, its technology increases revenue by filling more positions rather than raising prices.
«Our models analyze dozens, if not hundreds of categories of variables to reach prices. This can only be done now, thanks to AI», said Uri Yerushalmi, co-founder and head AI of Fetcherr. «The platform knows how to adjust offers directly whenever something changes in the market.»
AI helps companies win even after selling the tickets. Suppose someone closes a seat on a popular flight months earlier, at a low price, and a week before departure demand is launched. Volantio's platform, an Atlanta-based business information company, has access to reservations and locates passengers who may accept to move on a less crowded flight in exchange for a voucher. The position released can then be resold to a professional last-minute traveler for $1,000.
«The airline earns more money for the position and the passenger is rewarded for its flexibility», said Azim Barodawala, co-founder and CEO of Volantio, to whose customers Japan Airlines is among them.
The spread of AI in pricing, however, also causes concerns. Consumer protection organizations and American lawmakers warned last year that companies could end up using productive AI for «surveillance pricing», charging a different price to each client for the same location, based on personal data such as browsing history or income. Delta has stated that it does not use, nor plans to use, AI to determine fares based on personal data. Yerushalmi stressed that Fetcherr's models are based on aggregated market data rather than individual information, while Volantio announced that its offers for redistribution of positions are not personalized.
For airlines the reasoning is simple. «On the part of an airline, this is the ultimate secret ingredient», said Alton Aviation's Terry. «In current circumstances, it is easier to increase revenue than to reduce costs. And to achieve this, airlines continue to widen boundaries in the search for new sources of revenue.»
