Cargo’s Crystal Ball: The power of machine learning

Consulting the crystal ball

Even though machine-learning technologies cannot visualize the future with 100 percent certainty, they can support more accurate predictions about it. As a result, pricing and forecasting are two areas where algorithms will dramatically improve insights.

Much of a freight forwarder’s success depends on its ability to purchase freight at the right rate.  Luxembourg-based integrated IT solutions provider CHAMP Cargosystems uses machine learning to analyze historical pricing data, and the factors that influence pricing to help forwarders calculate spot rates.

On the carrier side, Digistics’ supports dynamic pricing, based on evaluation of historical pricing and shipper characteristics. “All the electronic air waybill [e-AWB] data moves through our cloud, and we can make meaningful representations from the data,” said Unisys’ Kohli. Data from the e-AWB, paired with characteristics of the cargo and shipper – like monthly volume, sector, geography, profile – enables the system to calculate pricing trends and offer a price range at the time of booking, or contract renewal.

Moving into future, the consensus among experts in logistics analytics is that machine learning will continue to optimize and improve deliveries. JD.com’s Yan says the next frontier for his division is to “look at ways that machine learning can be used for route planning for delivery personnel, and route optimization for transport trucks.”

Resilience 360’s Kamal added that it will complement customized results for individual customers. “We see the value of adding machine learning technology when it comes to really specific risks for each customer.”

Back to Top