Artificial Intelligence for Sustainability

KLM Has Wasted 60% Less Inflight Food, Thanks to AI

Not all flights booked result in a passenger on board said flight. By predicting the number of passengers based on historical data, AI is helping the airline waste significantly fewer inflight meals.

Published 10 months ago | About a 3 minute read

Image: KLM

KLM says it is wasting fewer inflight meals, thanks to artificial intelligence. Newly adopted AI programs now enable the airline to better predict how many passengers who have booked will actually board a flight — this allows the exact number of required meals to be calculated, which has resulted in up to 63 percent less food waste based on the expected number of passengers per flight. On an annual basis, this amounts to a saving of more than 100,000 kg of meals.

“Investments in digital technology are a priority for KLM,” said CEO Marjan Rintel. “The application of artificial intelligence contributes enormously to improving our flight operations and making them more sustainable. Combatting food waste is a good example of this — resulting in tens of thousands fewer meals being wasted on our flights each year.”

When it comes to air travel, not all flights booked result in a passenger on board said flight; depending on the class, between 3 and 5 percent of booked passengers do not show up or miss their flight. Because of strict regulations regarding airline food service to avoid spreading food-borne illnesses, all in-flight food that passengers do not finish is considered contaminated and is immediately disposed of upon landing. An estimated 20 percent — or $4 billion worth — of all inflight food is wasted per year.

To address this, KLM — which has encountered some bumps along its journey to create a more sustainable future for aviation — adopted Kickstart AI’s TRAYS model, which was specifically developed for the airline’s catering activities. It predicts the number of passengers on board based on historical data. The Meals on Board System (MOBS) receives the expected passenger numbers per flight with separate forecasts for Business, Premium Comfort and Economy classes. The prediction using the AI model starts 17 days before departure and continues until 20 minutes before the flight departs — enabling the most accurate prediction of the possible number of passengers for the entire catering process from purchasing to loading, thus preventing a surplus of meals.

The TRAYS development process assembled talent from companies including KLM, bol, Ahold Delhaize, NS and ING.

“We are pleased that we have been able to make a valuable contribution to this important project for KLM,” says Kickstart AI CEO and co-founder Sander Stomph. “Our goal with Kickstart AI is to accelerate the adoption of AI in the Dutch business community and we look forward to working closely with Dutch companies to make this happen.”

A three-month analysis shows that 63 percent less food has been wasted compared to catering for every booked passenger. The largest improvement can be seen on intercontinental KLM flights from Amsterdam’s Schiphol airport — where an average of 2.5 fewer meals (1.3 kg) need to be thrown away per flight. On an annual basis, this amounts to a saving of 111,000 kg in meals across all KLM flights that are catered from Schiphol.

KLM is also working on the application of artificial intelligence in other parts of its business operations — including making aircraft maintenance smarter. In addition, AI programs are used to simulate predicted bad-weather days, making it clear in advance which schedule would be best to allow flights to continue as much as possible. AI can also help passengers by giving personalized travel tips after booking a flight.