AI has helped reduce in-flight food waste by 63%. How did it do that?

How AI can help airlines reduce food waste on board

Reprinted from China Civil Aviation News

KLM uses artificial intelligence to reduce food waste on board

Cao Bin, China Civil Aviation News

KLM Royal Dutch Airlines is using artificial intelligence (AI) to reduce food waste on board. The AI ​​program allows KLM to more accurately predict the actual number of passengers who have already purchased tickets and calculate the exact number of meals required based on the expected number of passengers on each flight, thereby reducing food waste by up to 63%, saving more than 100,000 kilograms of food each year.

“Investing in digital technology is a top priority for KLM,” said KLM CEO Hermann Schmidt. “Using artificial intelligence helps us significantly improve the efficiency of our flight operations and make them more sustainable. Reducing food waste is a great example, saving tens of thousands of meals a year.”

KLM is working to use AI to determine the number of meals on board to reduce food waste. Not all booked passengers will actually board the plane. Depending on the class, 3% to 5% of booked passengers will not board (on time).

TRAYS is the first AI model specifically designed for KLM's catering business. TRAYS predicts the actual number of people on board based on historical data. The Meals on Board System (MOBS) receives the estimated number of people on board in three different cabins on each flight: World Business Class, Premium Economy Class, and Economy Class. The AI ​​model prediction starts 17 days before the flight and ends 20 minutes before the flight. This means that KLM can predict the actual number of people on board as accurately as possible, covering the entire catering process from procurement to assembly, reducing excess meal supply.

The AI ​​model TRAYS was launched by Kickstart AI at the end of last year. The project brought together talents from leading companies including KLM, BOLO, Royal Ahold Delhaize Group, Netherlands Railways and ING to develop the model.
Sander Stomph, CEO and founder of Kickstart AI, said: “We are delighted to contribute to this important project for KLM. Kickstart AI’s goal 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 possible.”

A three-month survey showed that using this model to predict food saved 63% compared to preparing meals for each passenger who booked a ticket. The effect was significant on KLM's intercontinental flights departing from Amsterdam Schiphol Airport, reducing the amount of food discarded by 2.5 portions (1.3 kg) per flight. All KLM flights provided with catering services at Schiphol Airport can save 111,000 kg of meals a year.

KLM is also promoting artificial intelligence in other business areas. Artificial intelligence is crucial to improving the intelligence level of aircraft maintenance. AI can also predict bad weather, which is conducive to making the best route planning in advance and continuing to operate flights. In addition, AI can provide passengers with customized travel guides after booking tickets.

(Cao Bin)