Players in logistics and global trade have yet to scratch the surface when it comes to making artificial intelligence work for their businesses.
A recent Freightos survey illustrates that AI, the common shorthand for the rapidly advancing technology, is still in its infancy. An overwhelming 96 percent of logistics professionals expect to leverage AI, while just 14 percent say they are already using or piloting solutions.
More from Sourcing Journal
In a recent webinar, Dan Gardner, co-founder and president of freight forwarder and customs broker Trade XCelerators, tried to demystify some AI use cases for U.S. importers.
From demand planning and forecasting; purchase order (PO) management; lead time management; automating processes at ports and terminals; optimizing air freight consolidations; or chatbots for customer interaction, AI can help importers streamline their tasks.
For example, an AI-driven purchase order management system could automate PO-invoice mapping. This would reduce disputes while increasing renewals.
“When an importer creates a forecast and they know what they think they’re going to sell for the balance of the year—what do they do then? They start placing purchase orders [with] vendors overseas,” said Gardner. “The application of AI to PO management can automatically sense and identify the last change date or the last sailing date on an upcoming PO.”
Gardner noted that cumulative lead time management is an area within trade and logistics where AI can help, particularly since current ocean freight visibility tools often focus on transportation, but not the order processing component. Cumulative lead time refers to the time between the importer placing a PO and getting the merchandise in hand.
He said AI-based cumulative lead time tools can leverage data to make predictions and decisions that shorten processes that last as long as 120 days.
Like many other industries, such as e-commerce, AI-based solutions can help deliver recommendations for shipment quoting, so that an importer isn’t just offered a price, but a recommendation based immediately on current availability, shipping route and capacity.
Businesses also can automate appointment making for drayage, which is vital in an industry where ports such as Los Angeles and Long Beach have a combined 12 container terminals.
“Humans are going in scouring the websites of every terminal and trying to make appointments,” said Gardner. “But it’s like trying to get Taylor Swift tickets when you’re trying to make appointments and they become available at four o’clock in the morning on a Monday.”
He’s skeptical of some of the applications he has seen touted in the industry—specifically calling out that predictive analytics for ocean freight shipping, saying “I don’t know if it lives up to the hype.”
The logistics industry is bullish on AI, despite the generally low adoption right now.
Only 5 percent of logistics professionals think AI is all hype, with more than half anticipating AI to have a major impact on the industry and 43 percent expecting a more limited impact, according to the 55 respondents surveyed by Freightos.
Most logistics providers and managers expect to use AI for pricing (64 percent) and customer service (56 percent) automation in the future, with about half anticipating using AI in operations (51 percent) and others for sales functions (35 percent) or software engineering (24 percent), the survey said.
While the conversation over automation in logistics itself has gotten testy over concerns of job losses, particularly at ports, AI and automation shouldn’t automatically be tied together, Gardner said.
“Don’t confuse AI with automation,” Gardner said. “AI is about interpreting data, applying what’s been interpreted and making predictions and decisions. We want to focus on how humans really think and work—the cognitive process, how we perceive our environment. How do we understand and experience different things, use our judgment, consider options, envision outcomes and apply what we learned? AI only knows what’s in the data.”
Gardner urged importers to carefully consider what AI solutions claim to offer.
“What specific predictions do the solution make? What decisions can it make without human intervention and without a program explicitly programmed to make that decision?” he said. “What level of ops experience have? You have brilliant technologists that wouldn’t know an airway bill from a telephone bill.”