Artificial intelligence has made huge strides in the shipping industry, but where is it really being applied today? Jason Murray, chief executive officer and co-founder of Shipium, explains.

Over the last 10 to 15 years, artificial intelligence has become “a lot more sophisticated” in its ability to collect huge amounts of data to support decision-making in shipping, Murray says.

The latest iteration of AI is “game-changing on the order of the internet,” he says. While still in “the early innings,” it promises to enhance optimization through the training of forecasting models based on input from multiple sources. In the past, Murray says, “we couldn’t scale that, but now it’s really interesting.”

A second innovation in the technology is agentic AI, making possible “self-running” tasks carried out without the need for human intervention. Again, it’s in the early stage of development and application. “The ability to communicate with humans, and tell them how to do things, is going to be huge [when applied] to the supply chain space,” Murray says. “If something is wrong with a shipment, most companies can’t scale [a response] by a group of humans to solve it. But the agentic piece can call it out to the carrier.”

Machine learning is well along in its ability to work with large datasets and training models. Users can better understand transportation flows, assessing, for example, how particular lanes have performed historically. That intelligence can then be combined with other considerations such as seasonality of demand.

Generative AI, which has dominated the tech headlines in the last three to four years, offers even more enhanced capability to analyze information on cost structure based on input from many different sources, he says.

Further in the future lies self-driving trucks and other types of vehicles that can greatly reduce the cost of middle- and final-mile delivery, Murray says.