Predictive value

posted on 15th August 2019
Predictive value

Using artificial intelligence to anticipate events like a shipment failing its delivery time would improve air freight services significantly and offer a competitive advantage, William Hayes reports

Artificial intelligence (AI) has undergone a rapid transformation in recent years, from sci-fi buzzword to a significant reality in more forward-thinking industries – including within parts of the aviation sector.

While some parts of the air freight world have been slow to embrace digitalisation, data experts and cargo specialists met at this year’s Air Cargo Europe event in Munich to discuss the potential of AI within air cargo, and whether it would soon become a meaningful tool – or an industry necessity.

“I truly believe AI can transform the air cargo industry – at this moment the technology is there, the computing power is there, and I think we are becoming more open-minded towards adopting digitisation in our day-to-day industry,” said Sara Van Gelder, cargo development manager at Brussels Airport Company – who is also responsible for Brussels Airport’s open data-sharing platform Brucloud, designed to enable different stakeholders in the air cargo community to act as an integrated network.

She added: “The challenge for a lot of companies is still stepping into this vision of data sharing, which some companies are still super-reluctant to do.”

Thorsten Friedrich, head of eFreight global rollout at Lufthansa Cargo, agreed that AI can be beneficial for the industry, noting how his company is “collecting a terabyte of data for every flight that we are operating, and can predict the chances that individual components of an aircraft fail during the operation – so when the aircraft lands we can predictively and pre-emptively repair stuff that could fail, which drives down operational cost”.

Pre-emptive capability

This notion of pre-emptive capability was also picked up on by Manel Galindo, CEO of WebCargo by Freightos, who highlighted the Jeff Bezos adage: “The best customer service is if the customer doesn’t need to call you, doesn’t need to talk to you. It just works.” His reference to Bezos comes as the US multi-billionaire’s Amazon increasingly encroaches on the territory of the air freight industry, including via its $1.5 billion investment in a 300,000 sqm US air cargo hub.

“The appeal of AI, in my view, is to try to reduce the pain of every player in the chain, from the shipper, airport, freight forwarder, airline. The idea is if you bring in AI, you can make the whole chain more efficient,” said Galindo, whose WebCargo website receives more than one million air freight rate searches per month.

Adrian Kosowski, head of research and co-founder of NavAlgo, and the only panel member not directly involved in the air cargo industry, said services like Galindo’s exert pressure on cargo carriers like never before, by telling them: “You have to change, you have to be digital, or you will be left out of the game completely.”

Kosowski noted: “So, in the end what I expect will happen is that the different stakeholders in this ecosystem will have to be sharing data in such a way that nobody is left out of the profits, not just from the sharing, but also from the profits. The first stage is getting the platforms done, solving the local individual problems – we have to think that in the end, these things are going to be decided at a business level, from a business perspective in a completely different way than they are being done now.”

Seeing the value

But building a data-sharing ecosystem is easier said than done. Brucloud’s Van Gelder observed: “The biggest challenge, as always, is not the technology, it’s not the computing power – it’s changing the minds of the people.” It’s all about seeing the “added value”, she explained, which companies often don’t see when they have their own internal systems.

“For that reason, we started to develop collaborative applications which solve operational issues, where collaboration between the different stakeholders is necessary and which can only be optimised if they start working together,” said Van Gelder. “So, by using these apps to start sharing data, we are taking the first steps towards analysing this data, finding algorithms and then feeding these algorithms back to our customers.”

Friedrich agreed that the collection of data and predictive capabilities of AI could improve the levels of service in the industry, saying: “If we were able to collect enough data to develop an algorithm that could tell us the probability of a shipment failing its delivery time… that would improve our service quite a bit. If there were an airline out there that could master this exercise, it would have significant competitive advantages over other airlines.”

Thorsten Friedrich

The International Air Transport Association (IATA) has been attempting to bring together data in the airline world with it’s One Record standard, added Friedrich, which pretty much does what it says on the tin – combines customer information currently held in different systems into one location.

This is welcome, said Van Gelder, but doesn’t mean that “the industry should wait for the standards to start moving”.

What could really help to move things along, she continued, would be to “try to enter the data digitally as early as possible in the chain. Because each stakeholder is entering the same data elements which were already in the system of the previous player in the chain; so if we start entering the data in a digital way as early as possible, I think that could truly help us in re-using it and starting to use AI and machine learning.”

“All the data has to be there, in one place, interoperable,” agreed Kosowski. “And this is something to think about, even at the early stages of digitalisation.”