Beyond the algorithm: How Uber Freight is using AI to redefine logistics

Artificial intelligence has become the loudest conversation in logistics, but at Uber Freight, the technology has been part of the company’s DNA from the beginning. Speaking with Supply Chain Management Review at the CSCMP EDGE conference, Steve Barber, vice president of product at Uber Freight, said the company’s evolution from a digital-first freight platform to an AI-powered ecosystem is well underway, and it’s happening faster than even the company expected.

“The good thing about Uber Freight is that we've been dealing with big transportation data for a long time,” Barber said. “We’ve got $18 billion of freight in our network … and so we’re very comfortable with the fact that our data has to be clean, it has to be solid. Our customers need their reporting, their analytics to work the way that it should.”

From chat to conversation

Uber Freight’s first step into generative AI began where many others did, with a chat interface. But the company quickly turned it into something far more powerful.

“The first place we started was the place everybody started,” Barber recalled. “We built a really cool AI-powered generative chatbot. It basically allowed an executive, a director, somebody that typically would not know where to go get the data that they need, to ask a question in plain English and get an answer back that was formatted and clean and usable.”

Over time, that evolved into a system that doesn’t just answer questions—it anticipates them.

“Now we’ve got situations where folks are no longer logging in and asking a question,” he said. “They’re logging in and being presented with things that the AI model believes they should be interested in.”

Those insights might flag a lane underperforming on cost, identify a carrier with falling service levels, or even recommend converting lanes to intermodal.

“It’s the kind of thing that when you start to use it, most folks go, this is just the way it should have been all along,” Barber said.

AI Agents that talk (and listen)

Uber Freight has now taken that conversational AI further with autonomous “voice agents” that can negotiate rates directly with drivers.

“It’s really amazing,” Barber said. “I can’t tell which person on the phone is the human and which person is the AI … they go through the full negotiation process of ‘I want to haul it for this rate’ … ‘Let me check and see if we can do that.’ It’s really, really cool.”

But, that doesn’t mean Uber Freight has cut out the human agent.

Barber explained that the AI handles early, low-risk interactions, freeing humans for the final steps of verification.

“We think humans in the loop is a very important strategy,” Barber said. “There’s a point where we don’t want AI to be making that decision for us … so we’ve got to mitigate the risk by having humans.”

The payoff is significant. According to Barber, using AI for the initial back-and-forth has reduced driver hold times“98%, almost a hundred percent.”

Guardrails and governance

Still, Barber was candid about the risks. “There’s too much risk in letting AI book it all the way through,” he said. “As far as we’re concerned, your robot is you.”

Uber Freight’s safeguard is simple: automation can accelerate decisions, but humans still own them.

“The cool thing the AI is doing is not setting price,” Barber said. “It’s having a conversation.”

That conversation relies on a deeply tested pricing algorithm that’s been part of Uber Freight’s operations for years.

“We automatically bid on rates and auction platforms,” Barber explained. “That algorithm is well established, super trusted, and highly monitored.”

Fraud, security, and the human layer

Fraud prevention has also become a major focus area. Barber said Uber Freight has built new tools to flag suspicious behavior in real time.

“Fraud is a big problem,” he said. “We saw a 72–73% fraud reduction over the last 12 months as some of the [new] tools were implemented.”

He compared the current wave of theft and fraud to the early days of what happened as e-commerce exploded.

“It’s happening again,” Barber warned. “It’s happening with chemicals and textiles and other things like that—but it’s happening again.”

Tying technology, network, and people

Ultimately, Barber said, Uber Freight’s goal isn’t just automation, it’s orchestration.

“We are trying to build a platform that ties together the technology, the network, and the people,” he said. “We want the technology to be an enabler … to meet people where they’re at so that we can leverage our network to their best benefit.”

He believes that’s where AI’s true value lies, not in replacing humans, but in amplifying their judgment.

“The human side of our business is alive,” Barber said. “And we believe it will be for a very long time.”

SC
MR