AI throwing wrench into arbitrage machine for forwarding labor

Freight forwarders don’t own cargo, and they are largely reliant on the assets of ocean, air, and road capacity providers. That liminal state puts them right at the center of the debate about whether artificial intelligence will transform how logistics is conducted.

Beyond the rhetoric of whether AI can functionally replace hundreds of thousands of people currently employed in operational, sales, development and administrative roles at forwarders, there’s a more basic metric that’s worth tracking. That is, how will AI — or, more specifically, automation enabled by AI — impact a forwarder’s cost structure? Because until recently, the money a forwarder spent per shipment boiled down in large part to how much it spent on labor to perform core tasks along the shipment lifecycle.

“For a digitally mature forwarder with decent data and workflow discipline, AI should be able to cut routine file-handling costs by roughly 15% to 35% over the next two to three years,” said Amit Maheshwari, CEO of Softlink, a provider of freight management solutions for forwarders. “On a poor-quality, exception-heavy import file, the savings will be lower. On a clean, repeatable lane with structured inputs, the savings can be higher.

“AI changes the math by shrinking the clerical piece of the work, but it does not erase local import knowledge, customer escalation, customs risk, claims handling, or delivery trouble-shooting,” Maheshwari added.

A simplistic way to view this challenge is that all the tasks a forwarder needs to execute on each shipment can be visualized on a spectrum of fully manual to fully automated. But that spectrum doesn’t account for the reality that forwarders have long tried to minimize their labor costs outside of automation. Those efforts sometimes come in the form of handing off work to business process outsourcing (BPO) providers, overseas agents, or employees in lower-cost regions.

That interplay of optimizing what work is done where, to maximize skills and customer engagement while making sure costs are in line with business expectations, is being upended by AI.

Wider margins, higher growth

JP Morgan, in a February report, posited that, “with successful deployment [of AI], the top European players can gain up to 200 [basis points] of market share and reduce staff costs by 20%, driven by higher productivity and monetizing natural attrition.”

That, in turn, has the potential to help large, publicly traded forwarders reach margins of 30% to 45% and double-digit operating profit growth by 2028, the bank said.

For arguments’ sake, let’s say that a forwarder using an origin office to handle all administrative aspects of a shipment is paying $120, while that same work handled at destination would cost $180. That 50% premium is the target for AI-enabled workflow automation.

“The origin-destination spread is real,” said Jamie Andrade, senior vice president of product management at SEKO Logistics. “A lot of the client-facing work, i.e. the service and differentiation bits we don’t want to offshore or automate away, sits on the destination side, and the labor differential is material. That said, it is getting eroded as costs rise [at origin]. We’re seeing that in our own operation and with our offshore partners. And honestly, you can only shift so much to origin before service starts to suffer.”

One example, from a software vendor who asked not to be identified, cited an example of a forwarding customer with nearly 100 IT and key account support staff, and 400 more employees handling basic freight forwarding operations at origin.

“They don’t have a plan to reduce labor costs [via automation] because of the complexity of dealing with origin activities,” and because labor costs remain sustainable, the vendor said.

“I think we will first see AI improving productivity and reducing manual labor more at the destination than at the origin,” said Hans Elmegaard, CEO of forwarding software vendor Moddule. “At destination, people manually coordinate operational activities between customers, carriers, customs offices, trucking companies/rail services, and share information with [distribution centers] and warehouses across various fragmented datasets. AI-backed orchestration services will help freight forwarders reduce labor costs and increase accuracy for their customers.”

Differentiator or table stakes?

A broader question is the extent to which automation drives differentiation in a crowded market over the long term, and whether a theoretical lower cost to serve bleeds into customer pricing.

Forwarding software WiseTech Global, in completely restructuring its pricing model last fall and later announcing it was reducing headcount by 30% over two years due to AI productivity gains, essentially signaled to the market that forwarders would have to reduce their cost to serve and that their customers would expect those gains to be passed through.

But the danger, argues Graham Cousins, former chief strategy officer at Vanguard Logistics Services, is that automation undermines rather than enhances margin.

“The challenge is that, at scale, lower cost-to-serve is not a core differentiator, and perhaps a risk if passed on in customer pricing,” Cousins wrote in a paper on how the international logistics industry was transforming.

“AI is not the bolt-on to make operators’ lives and workflows easier,” he wrote. “Customer service processes must be restructured to be AI-native and ensure data coverage, not to augment existing manual operations. Procurement and low cost-to-serve are baselines, not differentiators. If it ends up in pricing, the rest of the market will follow.”

So, at its core, the story of AI in forwarding is ultimately about how logistics services providers balance their current cost structure against one where automation plays a bigger role.

What can be automated?

“Most things along the shipment lifecycle are automatable,” said Robert Petti, CEO of forwarding software provider Prompt. “[WiseTech’s new pricing model] should have pushed people to do more automating. And the vast majority doesn’t require new tech. Integrated systems and good [standard operating procedures] should reduce labor cost 30% to 50%. That should get you closer to the origin costs and maybe even lower.”

Dan Bailey, CEO of forwarding workflow automation provider Nexcade, said it’s important for a forwarder to focus on a specific time horizon.

“We’ve seen two times file-per-head improvements in the specific workflows we target, but we’re also heavily focused on areas with commercial benefit, like procurement and quotation, not just [operating expenditures],” he said. “On simpler workflows it’s higher, but on an aggregate basis across the full lifecycle of all shipments, exceptions are still a primary driver of operational time and cost.”

Bailey said those exceptions are impacted less in the short run by AI but are “squarely in strike zone in the medium term,” adding that exception management was also largely unaddressed by business process outsourcing (BPO) efforts in the past.

“With automation levels of 60% to 80% of well-scoped processes, there can easily be a three- to four-times cost advantage for that specific process vis-à-vis lower origin per-shipment costs,” Fabian Struck, chief commercial officer of forwarding automation vendor Zauber, said. “This advantage is respectively higher for high-wage geographies. In the long run, we can only speculate what will ultimately be possible. For now, I do not see any reasons why total end-to-end shipment costs should not go down as far as three to five times, or even more.”

Low end of the curve

JP Morgan’s report cautioned that the forwarding industry still “sits low on the digitization curve,” with a reliance on manual processes and data that poses a risk to fully leveraging the benefits of AI for cost reduction.

“We do not expect AI to disrupt the fundamental prospects of the industry, given its exposure to physical transport and the movement of goods, with freight forwarding increasingly tapping into more complex services and moving further away from commoditized vanilla brokerage services,” the report said.

The other factor for forwarders to consider is that investments in technology to automate are not likely to stay static, but in fact are likely to move in a linear fashion with usage. That may prompt some to build internally.

“That’s part of why I’m leaning into building some of this ourselves for the easier use cases,” Andrade said. “If I own the tooling, my costs stay flatter as we scale rather than growing in line with transaction volume. It’s not the right answer for everyone, and you need [developer] resources and the appetite to maintain what you build.

“But we’ve got that capability, and the new AI coding tooling makes it faster to stand things up than it used to be,” she added. “So that’s where I’m placing my bets, at least on the BPO and automation side of the house.”

This article was originally published in the Journal of Commerce on May 12, 2026.