A cross-border logistics company was running its entire operation on Excel, group chats, and human memory. We didn't add software on top of that chaos — we redesigned the operational backbone around the order lifecycle, connected field hardware into the core workflow, and systematized hundreds of implicit business rules into a platform that six departments now share as a single source of truth.
Full-chain logistics system design · Order lifecycle architecture · Multi-department permission & collaboration system · PDA hardware integration · Business rule engine · AI customer service · Operational dashboard
We didn't start from pages and menus. We started from the order — and built the entire system around its lifecycle.
Most projects start from page layouts and menu structures — and end up as disconnected admin panels. We started from the core business object: the order. Every department's work, every status change, every financial action, every exception is attached to the order's journey through the system.
Six departments (Sales, Customs, CS, Warehouse, Finance, Management) each have different views into the same data. The design challenge isn't building six modules — it's ensuring they all operate on the same objects with consistent data, clear permission boundaries, and real-time sync.
Hundreds of edge cases — return-to-warehouse detection, lost package recovery, split-ticket sub-orders, balance-insufficient blocks, customer tier auto-calculation, release-requires-finance-approval — were maintained through experience and memory. We systematized every one.
PDA terminals aren't accessories — they're part of the core workflow. Scanning, printing, exception identification, and status confirmation on the warehouse floor feed directly back into the system. Without this, half the operation stays invisible.
The entire platform is structured around the order's journey: Customer → Order → First-mile → Customs → Warehouse → Last-mile → Delivery → Archive. Each department's actions attach to the relevant stage. Status, cost, exceptions, and permissions all derive from where the order currently sits in its lifecycle.
Role-based views for six departments: Sales sees customers and orders, CS sees exceptions and reply tasks, Warehouse sees receiving/QC/inventory/dispatch, Transport sees first-mile/last-mile batches, Finance sees fee structures and reconciliation. All updates sync in real-time across the entire organization.
We modeled the implicit rules that kept the operation running: return detection (last action = outbound → return; no record → new arrival; last action = inbound → duplicate scan), lost package recovery flows, split-ticket logic, balance checks before settlement, customer tier auto-calculation, and conditional release gates. These are now enforced by the system, not by memory.
We didn't just 'support PDA' — we embedded field terminals into the core business process. Scanning triggers status transitions, printing generates shipping labels / pallet labels / transit codes, exception identification (UNKNOWN flows, LOST recovery) happens at the device, and all operations write back to the system in real-time. Multi-warehouse state management ensures correctness across locations.
Three physical pipelines run through the system: the scan pipeline (inbound scan, exception detection, return judgment, lost recovery, overseas warehouse change confirmation), the print pipeline (box labels, pallet labels, pallet documents, transit codes, reprint flows), and the status write-back pipeline (device operations trigger state changes, multi-warehouse state correction, real API replacement of mock chains).
Quick search across orders, waybills, and customer profiles. Intelligent retrieval of knowledge base, SOPs, and exception handling procedures. Auto-generates customer communication scripts (bilingual), full-chain order summaries, and business suggestions for exception resolution.
Channel transport timing comparison, in-transit exception rates, order data completeness ratios, warehouse processing efficiency, last-mile performance and cost breakdowns, customer lifecycle value analysis. Business decisions shift from experience-driven to data-driven.