When Swiss International Air Lines began to use A.I. technology to boost efficiency, the carrier was able to better optimize more than half the flights in its network. It also saved 5 million Swiss francs ($5.4 million USD) last year.
Lufthansa had a similar experience. The German-based airline is using A.I. to better predict the winds that blow from the northeast to southwest Switzerland, which can result in flight delays and cancellations that can reduce capacity by up to 30% at Zurich Airport. A.I. is helping the airline more accurately forecast wind patterns, resulting in a more than 40% relative improvement in accuracy.
Swiss and Lufthansa are each leaning on A.I. forecasting models developed by Google Cloud, helping airlines model various scenarios and account for more “what if” disruptions that can delay or cancel flights.
A.I. tech that is being used by airlines today is improving “everything from how to actually schedule my people in a more efficient way, to understanding the weather conditions and being able to use that to predict when planes are going to land, and how much fuel they’re going to use,” says Warren Barkley, senior director of product management at Google. A.I. technology, Barkley says, “has the ability to look at hundreds of millions of data points and take factors that it never thought of before, or never could be used before, to predict what’s going to happen.”
The recent holiday meltdown experienced by Southwest Airlines highlights just how important it is for airlines to invest in tech that can help them avoid the chaos that can ensue when thousands of flights are canceled.
For carriers to run their operations effectively, they have to plan for a lot of turbulence. What’s the route of the plane? Is the crew at their destination? Is the plane early or is it late? What does any of that mean for connecting flights? And what does it mean for the luggage?
“If you look at some of the fascinating parts of how A.I. can be applied, a lot of it has to do with the flexibility to predict things,” says Barkley.
Delta Air Lines, JetBlue, and American Airlines are among the carriers investing in A.I. today. The sector is finally enjoying a bright tailwind after the COVID-19 pandemic badly bruised demand. Last year, the global airline industry’s revenue grew 44% to $727 billion from 2021, according to the International Air Transport Association, which is also projecting a return to profitability in 2023.
Courtesy of JetBlue
JetBlue Ventures, a subsidiary of JetBlue that invests in early-stage travel startups, ascertains that the pain points of travel are due to industry fragmentation. Travel can broadly refer to airlines, hotels, and ground transportation, and those providers don’t tend to work together. JetBlue Ventures aspires to connect the dots and create a more seamless journey.
“There’s so much data within our industry, and we do such a poor job of using it,” says Amy Burr, president of JetBlue Ventures. “The industry has a tech stack that’s very old and antiquated. Systems are old. And plugging in new technology is very challenging and very time consuming.”
Burr says A.I. can help airlines improve. She is particularly bullish about the use of A.I. to improve flight operations, the cockpit, and potentially autonomous decisions within the cockpit. A.I. can also be used to address issues that arise with plane maintenance, or ground and airport operations.
There’s also a long runway to boost fuel management, which is a key focus of sustainability for the industry and also a massive cost saver, as fuel is one of the largest expenses for the aviation sector.
“With fuel management, what you can do with A.I. is the ability to predict very precisely what the route is that you need to take, where the fuel needs to be, and how much is being burned,” says Barkley. Burr has a similar view. “Using an A.I. tool that allows us to better manage our fuel and routing, and allows us to save fuel, is a really compelling use case,” she says.
JetBlue Ventures has invested in eight A.I.-related startups, including Beacon AI, Tomorrow.io, UrbanFox, and FLYR Labs. With FLYR, JetBlue has already unveiled a partnership to better forecast pricing, which can help the carrier maximize revenue growth.
Fetcherr is a rival to FLYR that uses A.I. to create more stable pricing. That democratization levels the playing field to ensure consumers are paying similar prices for the same trip—but it ultimately is a tool that’s meant to boost revenue for carriers. The technology gathers a wide variety of data points, including airline ticket and booking data, competitor flight schedules and pricing, as well as information from capital markets, oil futures, and other economic indicators that influence the market.
“What is the willingness to pay? What is the elasticity of the demand? We know how to predict for every flight and every seat and how many passengers are going to buy a ticket for every possible price point,” says Uri Yerushalmi, cofounder and chief A.I. officer at Fetcherr.
Currently, most prices are set by humans. And price analysts can spook easily and lower prices if there is a concern about a low load factor, which is the metric airlines keep an eye on to determine the percentage of available seats that could have been filled by passengers.
“A.I. is not afraid of a low load factor,” says Yerushalmi. “They know what actions to do to maximize the revenue for the airline.” Last fall, Fetcherr announced that Brazil-based Azul Airlines became the first to pilot the startup’s demand prediction and algorithm pricing technology.
On the horizon, industry experts say A.I. could be used to create autonomous ground equipment and more futuristically, autonomous flying. As is always the case with A.I., this type of automation raises questions about the human jobs that could be eliminated when this tech is implemented.
“We want companies to invest in A.I., but at the same time, we crave that human connection as well,” says Alison Angus, head of practice for innovation research at Euromonitor.
The research firm says 51% of businesses it surveyed plan to invest in A.I., more than the planned investments in robotics and automation as well as augmented reality and virtual reality.
“Companies need to be careful and balance their investments in A.I. and robots and automation along with that human element,” Angus says. “We need to keep that emotional bond.”