AI agents in freight brokerage sales: faster quotes, automated orders and what shippers should ask

A freight broker sells one thing above all: a fast, reliable answer to the question "what will it cost to move this load, and who will move it?" For decades that answer came from a person working through an inbox, a phone and a spreadsheet of lane rates. When the inbox backed up, quotes went out late and the business went to whoever replied first.

That sales desk is where AI agents have made the clearest progress in logistics. An AI agent, in this context, is software that reads a request, decides what to do and completes the task in the company’s systems, such as pricing a load or entering an order, without a person handling each step. This deep dive looks at what two of the largest truckload brokers in North America have disclosed, what the numbers do and do not prove, and what leaders on both sides of the deal should do next.

The sales problem in freight brokerage

Freight brokers sit between shippers, who have goods to move, and carriers, who have trucks. Much of the sales work is repetitive and urgent. A shipper emails a spot quote request with an origin, destination, equipment type and pickup date. A rep checks market rates and recent costs, prices the load, replies, then re-keys the details into an order if the shipper accepts. Each step is small, but together they decide who wins the freight.

Before automation, delays were long. In 2025 C.H. Robinson said some customers had waited up to four hours for a person to reach their shipment in an email queue ³. In a market where volumes have been weak, that lag costs business. C.H. Robinson’s own 2026 planning assumes market volume growth of minus 3% to plus 1% for the year ¹, so brokers are competing for a flat pool of freight with fewer people.

How AI agents change the quote-to-order workflow

The shift is easiest to see in C.H. Robinson’s investor materials. Its third-quarter 2026 earnings presentation lists a "fleet" of AI agents, each with one job: quote agents that give customers transactional quotes, order agents that build and update orders in seconds, and others that contact carriers for tracking updates, book appointments, post available capacity, recommend loads to carriers, collect documents and make sure carriers are paid on time ¹. The same presentation lists email classification, email quoting and email order entry as uses of large language models, the technology behind tools like ChatGPT ¹.

The volume is large. By April 2025 the company said its agents had performed more than 3 million shipping tasks, delivered more than 1 million price quotes and processed 1 million orders ²³. It also said more than 5,200 customers were having loads accepted in under 90 seconds, and that monthly quote volumes rose at least 30% after less-than-truckload (LTL) freight was added to the quoting system ³. These are company figures, reported to trade press.

RXO, another large US broker, has described a similar path in its 2026 filings. In the first quarter it reported automating more than 500,000 phone calls, expanding an AI agent that answers spot quote requests by email, a more than tenfold improvement in time-to-bid for carriers using its digital platform, and a more than 30% quarter-on-quarter rise in digital truckload quote opportunities ⁴. In the second quarter it reported about 50% more calls handled by AI agents than the quarter before and five times more spot quote emails processed per day by its own AI agent ⁵.

In plain terms, the new workflow looks like this:

1. Intake. An agent reads the shipper’s email, identifies a quote request and pulls out the lane, equipment and dates.
2. Pricing. A pricing model suggests a rate from market and cost data. Older machine learning still does much of this work.
3. Reply. The agent sends the quote, often in seconds, within limits the broker sets.
4. Order. If the shipper accepts, an agent builds the order in the system, with no re-keying.
5. Exceptions. Unusual requests, large customers and loads outside set limits go to a person.

What the evidence shows

The strongest evidence is in productivity and earnings, and it comes from the companies themselves. C.H. Robinson reports that productivity in its North American surface transportation and global forwarding businesses improved by more than 15% year over year in the third quarter of 2026, and by more than 65% since the end of 2022 ¹. Average headcount in its North American surface transportation unit fell 11.4% year over year in the quarter, from 5,187 to 4,597 ¹. Adjusted income from operations rose 26.1% to $291 million ¹. RXO reported that brokerage productivity, measured in loads per person per day, rose 15% over the twelve months to the first quarter of 2026 ⁴.

The market has taken notice. In October 2025, after C.H. Robinson credited AI agents for productivity gains, its shares rose more than 20% in a day, the biggest intraday gain since 2007 ⁶. In October 2026 it agreed to buy RXO in a deal valued at $5.8 billion, expected to close in the first half of 2027, and it is counting on its "fleet of AI agents" to help deliver about $300 million in cost savings ¹.

Three cautions apply. First, these are company-reported metrics, defined by the companies, and not independently audited studies. Second, productivity gains also reflect process changes and cost cutting, not only AI. Third, the leaders are the largest brokers, with large in-house technology teams. Most of the industry is far behind. Statistics Canada found that just 1.8% of transportation and warehousing businesses used AI to produce goods or deliver services in the year to spring 2025, among the lowest of any sector, against 12.2% for all businesses ⁷. By the second quarter of 2026 the all-business figure had reached 19.2% ⁸. In the US, 19.8% of businesses reported using AI in any business function as of May 2026, according to the Census Bureau ⁹. The two surveys ask different questions, so the figures are not directly comparable, but both show that most firms are early.

Limitations and risks

Speed does not equal accuracy. An agent that quotes in seconds can also misprice in seconds. Brokers need limits on what an agent can quote without review, and regular checks of won and lost quotes against the market.

Fraud is rising alongside automation. Verisk CargoNet recorded 2,646 confirmed cargo thefts across the US and Canada in 2025, up 18% from 2024, with estimated losses of nearly $725 million, up 60% ¹⁰. It expects theft groups to focus more on misdirecting shipments tendered to legitimate carriers and notes that many schemes rely on taking over carriers with clean load histories ¹⁰. In Canada, the Canadian Trucking Alliance has been gathering member input on cargo theft, motor carrier identity theft and "chameleon" carriers that reappear under new names ¹¹. An agent that books carriers faster must also check them as carefully as a person would. RXO has introduced an AI agent aimed at fraud protection ⁴, a sign that brokers see the risk.

Relationships still matter. Large shippers buy on service, reliability and problem-solving, not only on response time. Automating routine quotes frees reps for that work, but only if the company redeploys them to it.

Human oversight that works

The programs described above are a split of work, not a handover. Agents take the high-volume, rules-based tasks, and people handle judgment calls. Good practice includes:

• Clear limits. Set price, customer and lane thresholds above which a person must approve.
• An audit trail. Log every quote and order an agent creates, with the data it used.
• Carrier vetting outside the agent’s control. Keep identity and safety checks as a separate gate that no agent can skip.
• Exception queues that people own. Assign named reps to review what agents escalate, with response-time targets.

Practical next steps

For brokers:

• Measure the baseline. Track quote response time, quote-to-win rate, orders keyed by hand and loads per rep before changing anything.
• Start with email intake. Classifying and extracting quote requests is lower risk than automated pricing.
• Let agents quote inside narrow limits first, such as a set of known lanes and customers, then widen them as accuracy holds.
• Pair automation with fraud controls from day one.

For shippers:

• Ask your brokers how quotes are produced, which steps are automated and how exceptions reach a person.
• Compare speed and service. A fast quote matters less than on-time pickup and honest updates.
• Check that carrier vetting has not been thinned out to keep up with faster booking.

Sources

1. C.H. Robinson Worldwide (October 28, 2026). Q3 2026 Earnings Presentation. Form 8-K exhibit filed with the SEC. Company-reported metrics.
2. Transport Topics, Connor D. Wolf (April 16, 2025). C.H. Robinson Expands AI Agents Across Shipment Life Cycle. Company-reported figures.
3. FreightWaves, Grace Sharkey (April 16, 2025). C.H. Robinson delivers on AI, but investors still skeptical. Company-reported figures.
4. RXO (May 7, 2026). First Quarter 2026 Results, earnings presentation. Form 8-K exhibit filed with the SEC. Company-reported metrics.
5. RXO (August 6, 2026). Second Quarter 2026 Results, earnings presentation. Form 8-K exhibit filed with the SEC. Comparisons are to the prior quarter.
6. Transport Topics, Matthew Griffin of Bloomberg News (October 30, 2025). AI Pays Off Big for C.H. Robinson.
7. Statistics Canada (June 16, 2025). Analysis on artificial intelligence use by businesses in Canada, second quarter of 2025. Canadian Survey on Business Conditions, 9,103 responding businesses.
8. Statistics Canada (May 27, 2026). Business conditions in Canada, second quarter of 2026. No industry breakdown published.
9. U.S. Census Bureau, Grundy, Breaux and Khatiwoda (May 26, 2026). AI Use at U.S. Businesses. Business Trends and Outlook Survey, December 2025 to May 2026, "in any business function" wording.
10. Verisk CargoNet (January 21, 2026). Cargo Theft Losses Surge to Estimated $725 Million in 2025. Covers the US and Canada. Industry data provider.
11. Canadian Trucking Alliance (August 4, 2026). Reminder: Share Your Input: CVSA Cargo Theft, Motor Carrier Identity Theft and Reincarnated Carriers Associate Member Survey. Industry association.