Mid-market international freight forwarder with a large overseas agent network

We automated event ingestion, entity resolution, and TMS milestone updates so operations teams could keep customer-facing visibility current.

Impact

22 min average event lag after implementation

Operating scale

22,000 active ocean and air shipments per month across North American import and export lanes.

Context

Shipment updates were arriving through unstructured emails, PDFs, portal notices, and malformed EDI messages. Coordinators had to interpret each update manually, match it to the right shipment, and key it into the TMS before the data went stale.

What was broken

• Status information arrived through fragmented channels with no consistent document format across agent and carrier partners.
• Event processing lag averaged 3.7 hours, leaving customer service and operations teams to work from stale data.
• Peak-season and disruption periods caused unread queues to spike, producing missed events and dark shipments in the TMS.
• A single shipment could carry multiple bills, bookings, container IDs, and agent references, making record matching error-prone.

What Twenty built

• Unified ingestion across email, portals, EDI, and fax-derived PDFs, with event classification by document type.
• Extraction and normalization of milestone type, date, location, equipment identifiers, and exception data.
• Multi-key entity resolution tuned for freight workflows, including container, house bill, master bill, booking, and lane-level signals.
• Delta detection and impact analysis so operators could see what changed before approving updates to the TMS.

Operational result

The visibility team spent less time chasing inboxes and more time handling true exceptions. Event lag fell, missed updates dropped, and customer-facing data stayed current.

‍