Commentary Trading Places

By Peter Tirschwell

It doesn’ t take much to envision AI using the full suite of carrier data to optimize a network for maximum profits.

It is reflexive in a conservative industry like shipping to assume that change will come slowly. Hundreds of years of experience supports that conclusion.

AI will not go along with that idea. The technology that is getting smarter by the day, causing tremors in the stock market, driving economic growth due to data center buildout and altering job prospects for millions, will be ruthless in upending the shipping industry.

The question is not how quickly change will come; it will come fast. The real question is where the impact will be felt most.

It will not structurally disrupt or reorder the industry. The top tier of ocean carriers who control assets, shipyard slots, and market share will not be replaced by AI.

If the recently announced acquisition of Zim Integrated Shipping Services by Hapag-Lloyd showed anything, it’ s that market concentration will only increase as time goes on. That is a 30-year trend that will not be reversed. In forwarding, technology advances by one player are quickly matched by others, neutralizing competitive advantage.

What AI will do without question is drive down costs, which in hyper-competitive and largely commodified container shipping means savings swiftly passed along to customers in the form of lower rates, versus making providers more profitable.

The carriers and forwarders able to drive down costs the fastest by substituting AI tools for headcount, for example, will be most competitive, while those slow on the uptake will get squeezed.

“ One issue that stands out is the opportunity to use AI to cut costs … the supply chain sector is ripe for major cost savings,” Benjamin Gordon, managing partner of Cambridge Capital, wrote in late January following his annual BGSA Supply Chain Conference.

“ Many carriers are indeed looking into AI as a measure of cost savings, but equally as a way to improve the way to serve customers,” Ulrik Sanders, BCG managing director and senior partner, told JOC. com.“ The area of focus of AI seems to be around customer service and internal‘ procure to pay processes,’ and here 30 % cost savings are realistic.”

Siddharth Vijay, founder of New York-based Lynk Labs, which is building AI agents to automate operational workflows at supply chain companies, estimates 40 % of ocean carrier customer service volume can be automated, with humans left to focus on complex exceptions.

“ I estimate this number can get up to 70 % in the next two years,” Vijay told the Journal of Commerce.

Assuming a $ 30,000 fully loaded headcount cost per customer service representative, if an ocean carrier has 2,000 staff, the savings would be around $ 24 million per year. If the staff count is 5,000, the savings could reach $ 60 million per year, Vijay estimated. Those numbers are too big to be ignored.

But for container lines and the container industry more broadly, there is a more profound impact AI could have in driving carriers towards greater long-term profitability. That is by taking capacity management to a whole new level.

Carriers have already made massive strides in capacity management by using blank sailings, vessel speeds and other maneuvering to offset overcapacity. Observers believe carriers can use such tools to prevent rates from freefalling, but not to squeeze capacity tightly enough to force rate upwards.

But the operational process of implementing capacity today is cumbersome, slow and lacks integration. Painstaking decisions only partially automated and disconnected from each other dictate which cargo to load, which bookings to accept, how cargo is stowed, what speed to sail ships, and when sailings get blanked, among others.

But the data that drives all those decisions is both robust and available. Thus, it doesn’ t take much imagination to envision the full suite of carrier data— bookings, rates, vessel locations and speeds, container locations and availability, and vessel turnaround times at ports— all loaded up into an AI engine instructed to run the network in such a way that maximizes profit.

Could AI one day soon, in other words, make the key tactical moment-to-moment decisions that collectively roll up to profit and loss based on a continuous stream of real-time data being constantly fed into the engine?

Such a scenario is“ very viable,” Vijay said. Others agree that scenario, based on the availability of multiple robust data streams already within carrier systems, is not implausible.

“ Everything that implies computation of large amounts of data to come to optimized outcomes can be radically improved by AI,” Sanders said.“ This indeed includes most core trade management functions, pricing and key customer interfaces.”

email: peter. tirschwell @ spglobal. com

86 Journal of Commerce | March 2, 2026 www. joc. com