C.H. Robinson uses in-house AI to expand margins amid revenue drop
C.H. Robinson has used internally developed AI agents to achieve a 45% productivity gain, demonstrating how targeted automation can drive double-digit earnings growth even during a severe shipping slump.
C.H. Robinson Worldwide has delivered double-digit earnings-per-share growth since 2023, despite a 34% drop in revenue caused by a post-pandemic slump in global shipping. The freight broker achieved this by deploying hundreds of AI agents that have boosted employee productivity by 45% since 2022. CEO Dave Bozeman applied Lean management principles to map workflows, eliminating tasks that did not add value and automating those that were repetitive.
By building systems in-house with 450 engineers rather than using third-party vendors, Robinson keeps token costs below $2 million while extracting hundreds of millions of dollars in value. “It provides us not just productivity,” Bozeman said. “This is revenue growth, margin expansion, productivity as well as customer advantage.”
This efficiency has fundamentally altered the company's labor model. Robinson experiences a natural employee turnover of 11% to 14% annually, but it is no longer backfilling all of those roles. For automated tasks like quote generation, which now take 31 seconds instead of 20 minutes, headcount is now largely divorced from volume.
Bozeman attributes the successful execution to cross-functional teams that combine engineers, legal staff, and operational experts. These teams use a rigorous risk-analysis framework to anticipate how AI systems might fail before they are deployed. This operational discipline has proven essential to scaling the technology without incurring unexpected costs.
Strategic pivot
The cost savings are funding a move up the value chain. Robinson is transitioning from a traditional freight forwarder to a supply chain consultant, aided by human specialists now freed from routine work to advise on shifting tariff regimes. Bozeman’s ultimate vision is what he calls “supply chain in a box,” aiming to convince clients they no longer need their own internal logistics departments.
The company is also using AI-assisted sales reps to recapture market share among small and medium-sized enterprises. Bozeman views the integration of domain expertise with proprietary AI as a formidable barrier to entry for competitors. “This is a deep, wide moat,” he said. “We calculated that if you wanted to replicate what we’re doing here, you would have to partner with 15 to 20 different entities to do that.”
