Freight forwarding has plenty of AI demos. The harder test is whether the technology can handle the daily mess of the quote desk: emails, PDFs, spreadsheets, carrier portals, agent replies and rates that change before anyone has time to clean the data.
That is where Sumeet Trehan, founder of Starboard, believes the market is moving. His argument is not that AI will replace small and mid-sized forwarders. It’s that those forwarders are too important to global trade, and too close to their customers, to be pushed aside simply because larger players have better tools.
Trehan came to that view after a career through Maersk, BCG and Flexport, where he helped build the company’s Canadian business. At Flexport, he said that he began to question whether digital freight was solving the right problem.
“Small and mid-size freight forwarders are essential for global trade,” Trehan said. “The operational expertise they have cannot be replaced by a digital freight forwarder.”
Starboard’s pitch is that local forwarders should keep the customer relationship and operating knowledge, while the platform gives them some of the advantages normally associated with larger competitors: better technology, stronger procurement and, eventually, financial infrastructure.
Forwarding remains relationship-driven, local and exception-heavy. The person who knows which carrier office to call, or which routing will fall apart, still matters. But the pressure on smaller forwarders is real. In developed markets, Trehan said, forwarding has become a thin-margin business, often operating at net margins of 1-2%. After the freight rate itself, labour is the biggest cost.
“It’s almost an existential crisis for a large number of freight forwarders,” he said. “How can they cut down cost? And it’s not a five-year strategy roadmap. For a lot of our customers, it needs to get done tomorrow.”
Starboard’s first product focuses on quoting. In spot-heavy forwarding markets, importers and exporters may send the same request to several forwarders at once. A small or mid-sized forwarder might receive 20-30 quote requests a day, Trehan said, with each taking one or two hours to build. Average response times can stretch to 24-48 hours.
That delay can lose business. The customer wants to know what it will cost to move cargo from A to B, and the forwarder that replies first often has the advantage.
Starboard plugs into the inbox of a forwarder’s sales or pricing team. When an RFQ arrives, its AI agent reads the request, extracts the cargo and routing details, breaks the job into legs, searches contract and spot rates, and can reach out to agents or co-loaders for missing pieces. The company says users have reduced response times from a day or two to roughly one or two hours in many cases.
Trehan argues the benefit is not just speed. Forwarders often receive rates from dozens of sources in different formats. A human may rely on memory, assuming yesterday’s best option is still today’s best option. In volatile markets, that assumption can be expensive.
“The idea is not only to quote fast, but also always quote at the cheapest and best rate you can give to your customers,” he said.
The industry’s data problem is old. Freight has spent decades trying to digitise around documents and messages that rarely line up cleanly between parties. EDI may be everywhere, but standardisation remains uneven. Trehan sees AI as useful because it can turn messy inputs into structured data that systems can use, without forcing forwarders to replace every legacy system first.
He is also cautious about AI hype. Starboard spent more than two years working with around 25 design partners in the US and Canada to train its agents on real quoting workflows.
“You can show magic in a demo,” he said. “But actually having something which can start replacing human work takes years and years of effort.”
For forwarders evaluating AI, Trehan’s advice is to start with the business problem, not the technology. Faster quote turnaround, better win rates, higher profitability per quote and less manual work are measurable. A vague AI pilot is not.
Trehan said customers have typically seen air quote response times fall to about an hour and ocean quotes to two or three hours. He also claimed average quoted rates can drop by around 5% when the system finds better options inside the forwarder’s own rate sources. Starboard does not name those customers publicly, citing confidentiality agreements.
His final advice is simple: choose the partner carefully.
“At this stage the technology is so nascent and early, you need to invest in the team over the brand or the company,” he said.
For smaller forwarders, that may be the useful AI story: not a future where the local operator disappears, but one where the operator answers faster and prices more intelligently.
This post was sponsored by Starboard.
Comment on this article