C.H. Robinson created proprietary technology incorporating Generative AI (GenAI) to automate transactions across the entire lifecycle of a freight shipment.
“We announced in May that we’d been using our new tech for emailed price requests. Within a few short months, we created new models to automate more shipping steps and have already implemented them at scale,” says Arun Rajan, chief strategy and innovation officer. “This a major efficiency breakthrough for the industry and for supply chains around the world. When you think about retailers that need hundreds of different products on their shelves or automakers that rely on just-in-time delivery for the 30,000 different parts in a car, saving hours and minutes on every shipment matters.”
Key takeaways:
- This technology reads incoming email and then replicates the tasks a person would do, whether that’s giving customers a price quote, accepting a load, setting appointments for pickup and delivery, and/or checking on the load in transit.
- This technology automates more than 10,000 of the routine transactions per day.
- After starting with price quotes, C.H. Robinson applied Gen AI to increasingly complex tasks. That required infusing the technology with the company’s market knowledge, specialized expertise in nearly every kind of supply chain and the specifics of each customer’s unique needs.
- C.H. Robinson’s new automation tech is being used for emailed price requests, emailed load tenders, emailed appointments and in-transit visibility.
“An emailed load tender might only say, ‘I have a load for Tuesday’ because the shipper knows we know what they ship on Tuesdays. Or it could contain thousands of words about 20 loads in a PDF attachment with handwritten notes on it,” says Mark Albrecht, VP, artificial intelligence. “Our tech can connect details in different parts of the email, discern what’s missing, go fill in the blanks and take action. We’ve even built it to determine things like which shipments are best for less-than-truckload and how different commodities should be palletized. We can do that like no one else, because we have the competitive advantage of the largest dataset in the industry and because our generative AI tools continually learn from our experts.”
“Once a person got to the email in their inbox, it still took an average of seven minutes to manually enter all the shipment details into our system, and that’s for a single load,” Albrecht says. “If the email tendered us 20 loads, a person would be stuck manually entering the information one load at a time. With generative AI, we can process all 20 loads simultaneously in the same 90 seconds. That’s an enormous time savings, especially when you consider we’ve scaled this to thousands of shipment orders per day just since June.”
