Shipment Record

Shipment details live in the dispatcher's head and on whiteboards. When a customer calls asking 'where's my load?', someone walks over to the dispatch board or calls the driver's cell phone to find out.

None — AI cannot track, predict, or optimize shipments because no shipment record exists in any system.

Shipments are tracked in a spreadsheet or emailed load confirmations. The ops coordinator enters some loads but not all — LTL moves and spot market loads often fall through the cracks. Finding last Tuesday's shipment means searching Outlook.

AI could scan email attachments and spreadsheets to assemble a partial shipment history, but cannot reliably calculate lane volumes or carrier performance because coverage is inconsistent and fields vary by person.

All shipments are entered into a TMS with consistent fields. Dispatchers can pull a report showing last month's volume by lane or carrier. But the TMS is an island — warehouse management and customer order systems don't connect to it, so matching a shipment to a customer order requires manual lookup.

AI can generate lane volume reports, identify top carriers by spend, and flag late shipments against planned delivery dates. Cannot correlate shipment cost to customer profitability or optimize across warehouse and transport because systems are siloed.

Shipment records are complete and connected — each links to the customer order, carrier contract, rate agreement, and warehouse receipt. A planner can query 'show me all shipments for Customer X in Q4 with carrier Y that exceeded transit time' and get a trustworthy answer within seconds.

AI can perform lane-level cost optimization, carrier performance benchmarking, and transit time prediction using historical shipment patterns. Cannot yet react in real-time to delays or reroute in-transit shipments because status updates arrive in batch, not streaming.

Shipment records are schema-driven entities with formal relationships to every connected object — carrier, route, rate, load, appointment, invoice, and emission record. Each milestone is timestamped and linked to its source (GPS ping, EDI message, driver app scan). An AI agent can query 'what is the current status and predicted ETA for PRO 78421?' and receive a structured, API-accessible answer.

AI can autonomously optimize carrier selection, predict ETAs with confidence intervals, trigger exception alerts, and recommend rerouting based on real-time conditions. Full autonomous execution is possible for routine shipment decisions.

Shipment records are living digital threads that auto-generate from connected systems — carrier tender acceptance creates the record, GPS feeds update location, EDI status messages log milestones, dock scans confirm delivery, and freight invoices auto-attach. The shipment record documents itself from creation through final settlement without human data entry.

Fully autonomous freight management is possible. AI agents monitor, decide, act, and learn in real-time — adjusting routes mid-transit, rebooking failed pickups, and settling invoices automatically. The shipment record is a continuous stream, not a static document.

Ceiling of the CMC framework for this dimension.