Worldwide Express works with greater than 130,000 clients and greater than 75 less-than-truckload (LTL) carriers, which fill a trailer with freight from a number of clients, AWS mentioned in a case research revealed Oct. 30. Each service bill is checked towards the acquisition order and the supply file earlier than the cargo is billed.
About 14% of these invoices didn’t match, the case research mentioned. The line gadgets had been flawed, the value differed from the contracted fee or an surprising price appeared, comparable to a liftgate cost. An outsourced workforce labored the pile one case at a time by cellphone and e mail. By June 2025, the unresolved circumstances held $15 million in frozen money move.
Three synthetic intelligence brokers constructed on Amazon Bedrock now work these disputes. Within six months, the backlog of energetic exceptions had fallen from $15 million to $4 million, the case research mentioned. That freed about $10 million in money move, lower time to money by 4 weeks and saved greater than $1 million a 12 months.
Worldwide Express, a part of ShipStation Global with about $5 billion in systemwide income, earned a 60-fold first-year return on the challenge, in keeping with the case research. It doesn’t say how that return was calculated.
Three Agents Split the Disputes by Type
Each exception goes to one in every of three brokers, the case research mentioned. An invoice-matching agent finds the cargo file and flags mismatches. A pricing-variance agent compares the service’s cost to the contracted fee and decides whether or not the hole is legitimate. An accessorial-validation agent checks whether or not further charges, comparable to liftgate or residential supply fees, match what occurred at supply. Before deciding, every agent reads the service’s tariff, surcharge definitions, dispute steps and the corporate’s variance thresholds. The similar three brokers cowl each service with out carrier-specific code.
The brokers attain the corporate’s SpeedShip transportation system by means of the identical interfaces its analysts use, the case research mentioned. They see the identical contract charges, cargo particulars and proof of supply. After every decision, the system writes a plain-language observe on what it discovered and what it did. At first, human reviewers accredited each advice. Bulk approval grew to become the default as soon as accuracy handed 99%. Rejections are logged and used to refine the brokers’ prompts and reference materials.
“With AWS, we’ve automated more than 40% of weekly billing exceptions,” Nathanael Pearson, Worldwide Express’ vp of FinOps, mentioned within the case research. The outcomes part places the determine at about 50%, or roughly 10,000 exceptions every week, with an error fee underneath 1%. The challenge went from an evaluation in June 2025 to manufacturing in underneath six months.
Unilever Automates the Order End of the Same Cycle
At the opposite finish of the order-to-cash cycle, Unilever’s AI brokers lower preliminary order processing from 15 to twenty minutes to about 90 seconds and settle refund claims twice as quick in pilot markets, PYMNTS reported final week.
Middle market finance chiefs are transferring the identical approach. According to the PYMNTS Intelligence report “Time to Cash: What Two Years of Data Say About Faster Cash and Rising Risk,” CFOs now automate about 73% of their accounts receivable processes, up from 55% a 12 months earlier. The report surveyed 100 U.S. CFOs at firms with $250 million to $2.5 billion in annual income in July 2026. Integration complexity is the impediment they cite most to automating the remaining, named by 90%, up from 67% in 2025, the report discovered.
Worldwide Express averted a lot of that integration work by plugging its brokers into the interfaces its analysts already use, and it’s extending the method to different order-to-cash processes, the case research mentioned. “Today, we release approximately $10 million more in customer billings each week,” Pearson mentioned.
For all PYMNTS AI protection, subscribe to the each day AI Newsletter.
