ShipRight Logistics AI

The problem

ShipRight moved freight across the US Southeast, working with eight different carrier relationships. Every morning, their operations team spent four hours manually reviewing the day's shipment queue, checking carrier rates, evaluating transit times, and assigning shipments to carriers. It was methodical, experienced work — but it was also perfectly automatable rule-following that didn't need humans to execute it.

As shipment volume grew, the four-hour morning window wasn't enough. They were making suboptimal carrier choices under time pressure, occasionally paying premium rates when better options existed, and the entire operation was gated on whoever did the routing being present and healthy.

What we built

An automated routing and carrier selection system that pulls from their TMS each morning, queries all eight carrier APIs for current rates and availability, and applies their routing rules to generate optimized assignments. The system considers: carrier rate, transit time, shipper history with that carrier (on-time delivery rates, damage frequency), and customer delivery SLAs.

For 85% of shipments, the system produces a fully automated routing decision and logs it to the TMS. For the remaining 15% — unusual dimensions, special handling requirements, high-value cargo — it generates a recommendation for human review with supporting rationale. The operations team spends 12 minutes reviewing exceptions rather than four hours routing everything.

A weekly optimization report shows which carriers are performing best and worst, with recommendations to adjust routing rules. The system gets smarter as carrier performance data accumulates.

Technology

• Python routing engine with custom carrier API integrations
• TMS integration via REST and EDI
• PostgreSQL for routing history and carrier performance tracking
• React operations dashboard with exception review queue
• Scheduled jobs on AWS ECS

Results

Daily routing time reduced from 4 hours to 12 minutes. Logistics costs reduced 22% in the first six months, primarily from better carrier selection eliminating premium-rate decisions made under time pressure. The operations team redirected the recovered capacity to customer service and exception management. Zero routing-related delays in the first quarter post-launch.