1. News
  2. AI models & agents
  3. Freight · AI billing
Freight · AI billing3 min read

Shipping firm uses AI to clear billing disputes and free up $10 million

Worldwide Express used AI agents to sort out freight billing disputes, cutting open cases 75% and freeing $10 million in cash, with people still approving.

By the Dashline deskSource: AWS Machine Learning blog
Shipping firm uses AI to clear billing disputes and free up $10 million
Photo: USDAgov, Public domain

Worldwide Express, a freight company that is part of ShipStation Global, says AI now handles most of its routine billing disputes. Six months after launch, active disputes are down 75%. Customers are paid about four weeks faster. The company also reports a one-time $10 million boost to free cash flow.

The problem

Worldwide Express has about $5 billion in systemwide revenue, more than 75 freight carriers and over 130,000 customers. In freight, an invoice has to match the purchase order, the goods receipt and the carrier's bill. Roughly 14% of invoices fail that check, and each failure becomes a dispute.

Each dispute was worked by hand, one case at a time, with calls to customers and carriers. Every carrier has its own pricing rules and dispute steps. The company says $15 million in free cash flow sat frozen each month in unresolved disputes.

What changed

Worldwide Express worked with AWS and a partner called LTM to build a team of AI agents on Amazon Bedrock, an Amazon service that supplies the AI models. The agents plug into SpeedShip, the company's transportation management system. When a dispute appears, it goes to one of three specialized agents.

  • Each agent looks up the specific carrier's rules before making a recommendation.
  • The agents use the same system connections human analysts use, so they see the same contract rates and delivery proof.
  • After a case is resolved, the system writes a plain-language note explaining what it found and did.

The company says the same three agents handle every carrier without carrier-specific code. Older rule-based automation struggled here because these cases need judgment, such as reading a carrier's fee policy and comparing it with shipment details.

People still decide

A human reviewer approved the AI's recommendations at first, using the existing SpeedShip screen. Once accuracy passed 99%, bulk approval became the default. Dashboards track error rates and how often reviewers override the AI. Rejections are logged, and the team uses them to improve the agents' instructions and carrier information.

The results

  • 75% fewer active billing disputes six months after launch.
  • Time-to-cash improved by four weeks.
  • $1 million in annual savings.
  • A one-time $10 million free cash flow infusion.

The project moved fast. Worldwide Express first worked with AWS in June 2025 to identify 19 automation opportunities across its order-to-cash process. It went from that assessment to production in under six months.

What it means

These figures are reported by the company, through an AWS blog post, so they are not independently checked. Still, the pattern is useful for any business slowed by invoice disputes. Let AI handle the routine cases, keep a person approving, and track where the AI gets it wrong.

The takeaway

AI can clear routine invoice disputes faster, but this company kept people approving until accuracy passed 99%.

More articles

You should really read these.

Next step

Let’s talk