AI agents now quote freight. A human still presses “book”.
The booking button is the last thing AI will be allowed to press
Fura Freight, a fast-growing US truck broker, says its AI agent now handles more than 40% of its load bookings. The agent is called Donna. FreightWaves reported on 11 November that Donna makes and receives calls, negotiates rates with carriers and books loads. CEO Jeff Dangelo expects the share to reach 90% by the end of next year.
A day later, on 12 November, California forwarder Freight Right Global Logistics launched a plugin that puts its quoting and tracking tools inside ChatGPT.
Get The Freight Brief by email
What's going on in freight, in your inbox every morning. Free.
Put the two side by side and a clear pattern shows. Our view: AI in freight has moved past answering questions. It now does nearly all the work around a booking: finding rates, haggling, checking compliance, filling in the details. But at both companies a human still owns the moment of commitment. We think that line will hold for longer than the hype suggests, and that forwarders who plan around it will do better than those who ignore it or try to skip it.
What Fura’s agent actually does
Fura’s numbers are big. Dangelo told FreightWaves the company grew from a $10 million revenue run rate to more than $100 million in two years, with about half of that organic. FreightWaves says “the company currently generates $2.5 million in revenue per employee.”
Donna sits on top of three systems, according to the article: Fura’s pricing algorithm, its TMS (transportation management system, the software that plans and tracks loads), and Highway, a carrier compliance tool. “Donna is doing all of that work from compliance to negotiating to booking,” Dangelo said.
Read on, though, and there is a human in the loop. “We do have her handing off the finished sort of negotiation to a team lead,” Dangelo said. FreightWaves reports that full autonomy is a longer-term goal, because each new acquisition brings new lanes and carrier relationships the agent has to learn.
Those acquisitions are the real business model. Fura has just closed its eighth and largest, Road Dog Logistics of Cincinnati, at just over $30 million in revenue. It plans 6 to 12 more deals next year and is targeting $1 billion in revenue within three years. It says it can move an acquired brokerage onto its systems in about two weeks, because integration is planned before the deal closes.
So the agent is not only a cost tool. It is how a buyer of small brokerages turns each purchase into more volume without hiring in proportion.
What Freight Right’s plugin can and cannot do
Freight Right’s launch uses MCP, the Model Context Protocol. MCP is an open standard that lets AI assistants such as ChatGPT or Claude connect to outside software and use it on a user’s behalf.
According to the company’s release, a shipper signed in to Freight Right’s Shipment Manager platform can, from inside ChatGPT:
- compare ocean FCL (full container load) and LCL (less than container load), air freight and US/Canada LTL (less-than-truckload) rates, with the lowest rates listed first;
- look up port and airport codes;
- send custom quote requests, such as full truckload or Mexico trucking;
- prepare a booking;
- track by reference, bill of lading, container or purchase order number;
- pull shipment documents.
The release lists five built-in skills covering pricing, quote requests, booking preparation, account and billing, and tracking.
The limits matter as much as the features. “The plugin can prepare bookings, but it cannot confirm them,” the release says. The shipper gets a link to a page where they review and request the booking, and Freight Right then confirms it. Payments are not taken in ChatGPT. Rate checks count toward the account’s monthly limit, and tracking shows the latest recorded status, which the company warns can lag the freight itself.
Founder and CEO Robert Khachatryan framed the launch around habit, not automation. “Many of our customers already use ChatGPT in their daily work,” he said. “Customers should be able to price a container, check on a late shipment or see what’s arriving this week just by asking,” he added.
Why the last click stays human
The reasons are partly legal. On 14 May 2026, the US Supreme Court ruled 9-0 in Montgomery v. Caribe Transport II. It held that state lawsuits against brokers for carelessly choosing a trucking company fall under the “safety exception” in federal trucking law, so they are not blocked. In plain terms: if a broker hires an unsafe carrier and someone is hurt, the broker can be sued.
Dangelo connects this directly to his deal pipeline. He said the ruling “changed the math risk overnight.” FreightWaves reports that small brokers are now looking to exit.
It also explains why an AI agent does not get the final say. When the carrier you pick can land you in court, someone with a name and a job title needs to have approved the choice. Fura’s agent runs compliance checks through Highway, but a team lead still confirms the deal.
For forwarders, the risk is different but just as real. A confirmed ocean booking ties up container space, triggers charges and sets dates for customs filings. A wrong commodity code or a bad cut-off date costs money. Freight Right’s choice to stop at “prepare” is a choice to keep that risk with its own staff, not with a chatbot.
This is not one company’s experiment
Larger players have been moving the same way. C.H. Robinson says “hundreds of connected AI agents are now embedded” in its operations. The company says its agents deliver price quotes in 32 seconds, process orders in 90 seconds and automate appointments across 42,000 locations. It also says AI-recommended loads are booked four times faster. Its line on jobs: “The goal is not to replace people.”
The common thread across Fura, C.H. Robinson and Freight Right is speed on the routine work, with people kept on the decisions that carry liability. Fura is the most aggressive of the three, and even it has not removed the human sign-off.
The market backdrop raises the stakes. Dangelo cited spot cost-per-mile increases of 31% for dry van, 32% for reefer and 28% for flatbed. Spot rates are one-off prices for a single load, not contract rates. When capacity tightens like that, fast quoting wins freight, and a slow email reply loses it.
What this means for forwarders
First, being findable by AI tools is becoming part of sales. If shippers start asking ChatGPT “what does a 40-foot container from Shanghai to Long Beach cost this week”, the answer will come from forwarders whose rates a machine can read. Those who send PDFs by email will not appear.
Second, the human step is worth keeping, but it has to be quick. If a shipper gets an instant quote in a chat and then waits a day for confirmation, the speed gain is lost. The forwarders who do well will promise a fast human confirmation, measured in minutes.
Third, data quality becomes a commercial issue. Freight Right openly warns that tracking may lag. Any forwarder plugging into an AI tool will have stale milestones and missing references exposed in front of the customer, without an account manager to explain them.
Fourth, roll-ups will speed up in US brokerage. Fura’s model, which is to buy small brokers and move them onto an agent stack in two weeks, works better when owners want to sell. Montgomery gives them a reason to.
For shippers, the near-term gain is easier price checks and tracking. The risk is relying on a chatbot’s rate list as if it were a firm offer. Under Freight Right’s terms, it is not one until the forwarder confirms it.
Our prediction
Our call: by 30 June 2027, at least three more US forwarders or freight brokers will launch quoting inside ChatGPT or Claude through MCP or a similar connector. None of them will let the AI confirm a booking without a human approval step. We also expect Fura to report AI handling a clear majority of its bookings, above 50%, by the end of 2027, with a team lead still signing off.
Sources
Get The Freight Brief by email
What's going on in freight, in your inbox every morning. Free.
Read us on Google? Add Forwarder News as a preferred source to see more of our stories in Top Stories.