C.H. Robinson
C.H. Robinson's generative AI agents handle over 3 million freight tasks
C.H. Robinson, a large US freight broker, built generative AI agents that quote prices, enter orders, post trucks and book appointments. By April 2025 the company said the agents had performed over 3 million tasks, and many customers had loads accepted in under 90 seconds.
Key figures
Copied exactly from the sources. Copy any figure with its citation.
Over 3 million
Shipping tasks performed by the company's generative AI agents (as of April 2025)1
Well over 1 million
Price quotes delivered by AI (as of April 2025)1
1 million
Orders processed by AI (reached in March 2025)1
Over 5,200
Customers whose emailed loads are now accepted in under 90 seconds, instead of waiting up to four hours in a queue1
What it took
Cost, team size and timeline were not disclosed in the sources.
The problem
Freight brokers handle huge volumes of repetitive work. Customers email for price quotes and send load orders, carriers email about spare trucks, and staff must read, key in and schedule each item. Until recently a person had to get to each email in a queue, which could take up to four hours for a shipper waiting to have a load accepted.
What they built
C.H. Robinson built its own generative AI agents. An agent is a program that takes on a whole task, not just one answer. The company says they now give price quotes, process orders, find trucking capacity, set pickup and delivery appointments and check on loads in transit. It has also taken first steps toward answering tracking requests.
How they did it
The company says it has used AI at scale for more than a decade. In 2023 it broke what it calls a long-standing barrier to automating freight work with generative AI, and it announced its first agents during 2024. The quoting agent started with truckload freight and later added less-than-truckload (LTL) freight. An email classifier can spot when a shipper asks for both. The order agent first read emails from the biggest truckload customers and in 2025 was extended to small and medium businesses. A truck posting agent launched in January 2025 reads carrier emails and loads the available trucks into the company's capacity system. The company draws on data from its 37 million shipments a year to build its models.
Results
By April 2025 the company said its agents had done over 3 million shipping tasks, delivered well over 1 million quotes and processed 1 million orders. It said over 5,200 customers now get loads accepted in under 90 seconds. Its LTL quotes by AI rose by at least 30% every month after LTL was added. In February the truck posting agent uploaded nearly 10 times more trucks to the system. An executive said generative AI played a key role in a 30% productivity increase across 2023 and 2024, and that it lowers the cost to serve.
All reported figures
| Measure | Figure | Citation |
|---|---|---|
| Shipping tasks performed by the company's generative AI agents (as of April 2025) | Over 3 million1 | |
| Price quotes delivered by AI (as of April 2025) | Well over 1 million1 | |
| Orders processed by AI (reached in March 2025) | 1 million1 | |
| Customers whose emailed loads are now accepted in under 90 seconds, instead of waiting up to four hours in a queue | Over 5,2001 | |
| Productivity increase across 2023 and 2024, which the company says generative AI played a key role in | 30%1 |
Lessons for your project
Start with a repetitive, high-volume task that has a clear business value. C.H. Robinson began with truckload price quotes, then extended the same agent to less-than-truckload freight.
Build one agent per step of the process (quote, order, truck posting, appointment) and measure each separately, so you can see which step pays off.
Roll out to your biggest customers first, then widen to smaller ones. The order agent first handled email from the largest customers and later moved to small and medium businesses.
Track speed from the customer's point of view. The company reports the wait for an accepted load, not just the number of tasks.
Caveats
- All figures come from the company's own press release. No independent source was found.
- The 30% productivity increase is credited to several factors. The company says generative AI played a key role, not that it caused all of it.
- The model and vendor behind the agents are not disclosed. The company calls the tools proprietary.
- The release gives different customer counts (83,000 in one place, 74,000 in another), so no customer count is used here.
- The release does not give a go-live date for the agents. It says the company announced its first generative AI agents throughout 2024. The deployed date shown is the date of the release.
Sources
- At C.H. Robinson, Artificial Intelligence Has Now Performed Over 3 Million Shipping Tasks
C.H. Robinson · 2025-04-16 · Press release · chrobinson.com
Summarized in our own words from the sources above. Spotted an error? See corrections.