■ CRITICAL RISK ■ Transportation & Logistics
Largely, yes. Matching loads to trucks, quoting rates, and tracking shipments are exactly the problems digital freight platforms were built to eat, and they're eating. The agents who survive will be the ones handling freight weird enough to confuse an algorithm.
“Shipping logistics is just math with boxes.”
Our AI replacement risk score — how we score jobs
A freight agent lives on the phone and in the TMS: sourcing carriers for shippers' loads, negotiating rates against a volatile spot market, booking the truck, chasing the driver for check calls, and smoothing over the inevitable — late pickups, breakdowns, a receiver who won't unload after 3 PM. The core value proposition has always been the Rolodex and the hustle: you know carriers, you know lanes, and you can get a load covered by Friday afternoon when nobody else can.
The problem is that a Rolodex is just a database with worse uptime. Digital freight matching platforms — Uber Freight, Convoy before its collapse, DAT's automated tools, and the brokerages' own in-house systems — now quote rates instantly from live market data, match loads to carriers algorithmically, and track trucks by GPS instead of check calls. Large brokerages have publicly reported automating huge fractions of their transactions end to end, no human touching the load. AI handles the email quoting too: systems now read a shipper's inbound request, price it, and respond before an agent finishes their coffee.
What's left for humans is the freight that fights back: oversized loads, hazmat, multi-stop projects, cold chain with tight windows, claims disputes, and relationship-heavy accounts where the shipper wants a name, not a portal. That book of business is real but shrinking as shippers get comfortable with self-serve platforms, and it supports far fewer agents than the industry currently employs. A 76 risk score reflects a job whose median version is automatable even as its hardest version stays human.
Automatability: our editorial assessment of current and near-term AI capability
This is a right-now story. Major brokerages already automate large shares of load matching and quoting, and AI email agents are answering shipper requests today. Expect steady headcount compression through the late 2020s as platforms absorb standard dry van and reefer freight, with the surviving agent roles consolidating around complex, high-touch freight by the early 2030s. The spot-quote-and-track version of this job doesn't have a next decade.
Viable, but narrower. The transactional middle of the job — quote, match, book, track — is being automated aggressively by digital freight platforms and brokerage AI. What remains viable is specialized and relationship-driven freight: complex loads, dedicated accounts, and problem-solving that requires judgment and accountability. Fewer seats, higher skill requirements, and commission structures that reward exactly what software can't do.
More than most outsiders realize. Large digital brokerages have reported automating the majority of standard load transactions end to end, and AI systems now handle inbound quote emails, carrier matching, and GPS-based tracking without agent involvement. The automation concentrates in dry van and standard reefer freight; specialized and exception-heavy freight still routes to humans.
Move up the complexity ladder. Learn oversize permits, hazmat regulations, cross-border customs, or cold chain requirements — freight categories where a mispriced load costs real money and shippers pay for expertise. Simultaneously, deepen a book of accounts that value your judgment. And use the automation tools rather than racing them; the productive agent of 2030 supervises software doing the routine work.
Convoy's collapse was a business-model failure in a brutal freight recession, not proof the technology failed. Its matching tech was acquired and lives on, Uber Freight continues, and traditional brokerages quietly built the same automation internally. The lesson isn't that algorithms can't broker freight — it's that they can, at margins so thin they crushed a well-funded startup. That margin compression hits agent commissions too.