Artificial Intelligence Bursts Limits
Artificial intelligence is the greatest invention since the steam engine—at the least. Everyone talks about it, but it is far from being used extensively in logistics and especially in air freight. Here are highlights of two forums on the topic at transport logistic 2019.
A survey by the magazine “Logistics Today” has shown the following: 90 out of 100 companies hope to gain a better market position from artificial intelligence (AI). It remains lip service for many. Only 26 of the respondents are actively working on applications. More than one in two (54 percent) lack know-how for it. This already starts with the definition. Previously, we fed data to a computer and it calculated with a preset algorithm. Today, AI recognizes the way from data input to output data itself. Robotics, speech recognition, and—also as part of the field—machine learning work based on this.
Where the greatest potential is
Companies see the most potential in AI for demand forecasting and demand planning (62 percent), production optimization (51 percent) and transport optimization (50 percent). Data and algorithms enable predictions using very different parameters. Even short-term information such as emerging traffic jams, sudden weather changes or waiting times at customers saves time and costs. The forum participants at transport logistic see benefits for event-based and dynamic tour planning. Examples show how tank tours can be organized with filling levels, such as electric vehicles under the influence of temperature, topography and traffic, drones can send pictures for damage checks on containers, or how machine vision helps to distinguish symbols for dangerous goods. Neural networks can already recognize 27 symbols.
Integrate data first
It is important in the first step to identify applications for AI in general. It is a question of integrating data long before the algorithm. There are currently 20 large and small use cases at DB Schenker, some of them in research. The focus is on dynamic offers and prices, demand and forecasting as well as capacity planning and autonomous vehicles. AI can also help to read from the booking behavior when a client appears to be switching to another provider or how goods are efficiently packaged in containers, a kind of 3D Tetris.
The added value of AI is the self-learning supply chain
Phoning, emailing and faxing has been used until today. Prices are negotiated orally, and static data provide orientation. With so many combinations of time, travel and resources, no one can derive the mathematical optimum. The goal and added value of AI is the self-learning supply chain. It avoids having products come from production when no trucks are there, and consequently the customer has to wait and business goals are not achieved as a result. This only works without silo mentality and rigid functional limitations.
Who will be the first Amazon of the airfreight industry?
“Artificial Intelligence: Next Level Air Cargo” was one of the forums at transport logistic and showed the interest of air freight spoiled by success. Its problem is increasing customer demands; the passenger area sets the benchmark with online portals and the ticketing service Etix on all terminals. The participants see benefits in a cloud or at least in standards for data and processes for the complex value chain. The major players in the International Air Transport Association IATA are working on this. The customary mindset is the biggest challenge. The industry still shies away from sharing its data, but it has to do this. “There is a killer out there,” Thorsten Friedrich warned boldly. He introduced e-billing at Lufthansa. His outlook for the future: Whoever controls data on Amazon level first will shake up the air cargo industry.