(Bloomberg) -- With cargo shipments falling in the post-pandemic era, some shippers have turned to artificial intelligence to boost efficiency. The freight unit of Uber Technologies Inc. has tapped its vast data files to use AI to predict late shipments — and make suggestions for alternative routes.

On Thursday, Uber Freight also introduced Insights AI, which it says will allow customers to make queries that could normally take days for workers to compile, such as routes with the worst on-time performance and the causes. The company also has AI embedded in its service that will allow shippers to estimate arrival times more precisely.

The AI offering is the culmination of Uber Freight’s transformation from a startup digital cargo broker in 2017 to a full-service transportation handler. The change was accelerated with the $2.25 billion acquisition of freight manager Transplace in 2021 and a $120 million investment to upgrade its systems to tap into data from the $18 billion worth of annual cargo movements it manages.

The goal, Uber Freight Chief Executive Officer Lior Ron said in an interview, will be for users to know “what’s happening across their organization in terms of service, cost, the operation; and then be able to take action on that in a very intuitive way.”

Uber Freight’s parent company doubled down on its broader commitment to technology on Wednesday with its announcement that a semiconductor executive, Prashanth Mahendra-Rajah, will become its new chief financial officer. 

Read More: Uber Taps Semiconductor Executive as New CFO to Replace Chai

Logistics are getting more C-suite attention than ever after supply chains strained to handle a flood of imports during the pandemic, causing transportation costs to soar even as on-time performance deteriorated. 

Spurred in part by software from startups including Convoy Inc., the freight industry is capturing more data than ever, which AI can quickly analyze. The technology is expected to cut warehouse wait time for drivers, increase on-time deliveries, reduce movements of empty trailers and address other inefficiencies that arise from the complexity of connecting rail, truck and maritime networks through hundreds of thousands of trucking companies and thousands of freight brokers.

The world’s biggest parcel companies, FedEx Corp. and United Parcel Service Inc., have been leaders in building in-house AI tools for customer service and to optimize how packages are routed. AI large language models — neural networks that are trained using massive amounts of text and data — are being adopted even by small companies to help reduce data-entry work in filling out order forms.

Uber Freight is now providing these AI tools to shippers, not only to attract more cargo business but to increase sales from software as a service. Customers, including chemical giant BASF SE and Nestle SA, will be able to analyze their own data and compare it to a peer group that doesn’t reveal individual shipping company data.

“When you have a supply chain that’s expansive and complex as ours, cutting-edge technology isn’t a nice-to-have. It’s essential,” said Randy Cooper, director of transportation at Del Monte Foods, in an Uber Freight statement. He added that the Uber Freight partnership “has become our secret sauce as we serve our customers.”

The freight market right now is in recession and Uber Freight is losing money. In the second quarter, the loss before interest, taxes, depreciation and amortization was $14 million, compared with a gain of $5 million in the same period last year. 

With the backing of Uber, the freight unit is still investing to build out its business, Ron said. It faces competition from legacy freight managers such as C.H. Robinson Worldwide Inc. and Expeditors International of Washington Inc. and brokers, including RXO Inc., which are also adopting digital systems.

Selling AI

Uber Freight will face challenges selling its new AI tools. Supply-chain executives tend to be cautious about signing up for new technology because one horrible snag can cost them their job. 

All the talk about AI and its ability to increase productivity may help. Large language models “are revolutionizing everything” because the technology is easy for people to use if it is set up properly, said Daniela Rus, director of the Computer Science and Artificial Intelligence laboratory at Massachusetts Institute of Technology. But the technology is still a work in progress, she said.   

“You need to have significant computing resources and you need to have the data, the resources, and the in-house knowledge in order to use these systems,” Rus said. 

Nonetheless, even simple software solutions can help small companies. CSA Transportation, a trucker with about $100 million in annual revenue that hauls goods mostly between the US and Canada, is testing AI to handle clients’ e-mails, which is typically done manually. The software, created by qBotica, sends a follow-up email automatically to customers if information is missing, said Herman Fallick, CEO of CSA Transportation.

“If you can take away all that manual data entry, I can free up those employees to do real customer service like following up on orders and supporting sales,” he said. 

(Adds Del Monte Foods comment in 10th paragraph. An earlier version corrected Lior Ron’s title in the fourth paragraph.)

©2023 Bloomberg L.P.