Two of the largest names in freight logistics just told investors exactly what AI is worth to their productivity. Kuehne+Nagel says AI agents will add an annualized CHF 100–150 million to profit by the end of 2027 — roughly 5% across more than 25,000 white-collar roles. C.H. Robinson already has the results, and they just got fresher: both their trucking brokerage (NAST) and freight forwarding divisions have improved productivity more than 60% since the end of 2022 — including another 15% jump in this quarter alone — powered by hundreds of AI agents running quoting, order-entry, and tracking.
Strip away the currency and the headcount, and what's left is a plain productivity percentage — and productivity percentages don't care how big you are. If you're running a forwarding operation with twenty operators, or two hundred, that number applies to you exactly as much as it applies to them.
Run that math at your own size. Most forwarders are nowhere near that level of productivity per operator today. Closing even part of that gap isn't a number that only exists in a Fortune 500 earnings call, it's the same shift C.H. Robinson already made, just applied to your headcount instead of theirs.
K+N's number is a projection, and to their credit they didn't oversell it: management called it a gross estimate, said they'd track it against real operational metrics, and flagged that adoption and rollout could move the number. A company that size doesn't put a figure like that in front of shareholders unless the internal case is already real.
Not because operators don't want the gain. Because most of the industry runs on a patchwork: an aging core system, a rate tool bought separately, a visibility portal bolted on, spreadsheets filling every gap between them. You can't layer real AI on top of that — there's no single place for it to read from, act on, or learn from.
It shows up hardest exactly where the operation gets more complex: intercompany billing between branches, a controller reconciling the same shipment across entities and currencies, an offshore back-office team keying the same booking twice because the systems don't talk, agent-network settlements still resolved over email. A five-person shop might not recognize any of that. A forty-operator forwarder running three offices sees it immediately — and it's exactly the part a patchwork of point solutions can't fix, no matter how many of them you buy.
Here's the part worth sitting with: K+N said as much themselves. Buried in their own investor materials is a line about why they believe integrated AI will become their competitive moat, and the reason they give is owning their own tech stack: "ownership of TMS and in-house IT development enable deeper AI integration, strategic differentiation, and speed to market."
C.H. Robinson's own chief strategy officer made almost the identical case, independently, the same week: "Our AI agents are powered by proprietary data, deep logistics expertise, and an engineered context layer that simply cannot be purchased or built overnight." Two giants, two different earnings calls, six days apart. Neither one is arguing that AI is about scale. Both are arguing it's about what you own and control.
A CHF-12-billion-a-year company and one of the largest logistics providers in North America are independently telling their own shareholders the same thing: AI only compounds if it's built into something you actually own, not bolted onto someone else's disconnected system. That's not a scale problem. It's an architecture problem, and it's as solvable at forty operators across three branches as it is at twenty-five thousand.
K+N put its AI numbers in front of investors on July 23. Six days later, C.H. Robinson went further on its own Q2 call: not a projection, but AI already delivering measurable results. The same week, two other major forwarders, Logwin and Geodis, reported their own results and barely mentioned AI at all. Separately, a post went around misquoting K+N's guidance raise as AI-driven, which isn't what they said.
That's the pattern worth watching: these numbers are moving quarter over quarter, sometimes week over week, and it's easy for last quarter's figure — or someone else's summary of it — to keep getting repeated as if it's still accurate. We're checking each one against what the company actually filed, every earnings season, and sending the accurate read to anyone who wants it.
The gap opening up in freight forwarding isn't between giants and everyone else. It's between operators running on one connected platform built for this — where quoting, operations, and accounting all feed the same system, and AI has something coherent to act on — and operators running on a patchwork held together by consultants and tribal knowledge, where AI has nowhere real to plug in.
K+N and C.H. Robinson just told their investors which side of that gap they're planning to be on. The question worth asking is which side your business is on — and whether that's a choice you're making on purpose, or one that's been made for you by whatever system you happened to end up on.
That's the gap GoFreight was built to close. One platform for quoting, operations, and accounting, built so AI has something real to work with instead of a patchwork to work around.