A mid-sized logistics firm ran on three disconnected tools. Staff spent hours each day copying data between them by hand. A thin integration layer made them a single system, without replacing anything.
A logistics firm with roughly forty staff ran its daily operations across three systems that had no connection to one another. The warehouse management system tracked inventory and pick lists. The courier API handled label generation, pickup scheduling, and tracking, but it was a separate third-party service with its own web portal. Invoicing lived in a homegrown database built years earlier by a developer who had since left the company. Moving a shipment from warehouse to delivery to invoice meant three people entering the same data three times.
The cost was not just the hours lost to manual data entry. Every re-typed field was a chance for error: a transposed tracking number, a mistyped address, an invoice sent to the wrong billing contact. Discrepancies surfaced days or weeks later, usually when a customer complained. The team knew the problem but assumed fixing it meant ripping out their existing tools and starting over: a project they could not afford to take on.
Rather than replacing any of the three systems, a thin integration layer was placed between them: a handful of serverless functions wired together with an SQS queue and an EventBridge bus. When a pick list is completed in the WMS, an event is published. A Lambda function picks it up, calls the courier API to generate labels and schedule pickup, and writes the tracking number back to the WMS. When the courier confirms delivery, another function pushes the completed shipment details into the invoicing database. A simple reconciliation dashboard queries all three systems and highlights any shipment where the data does not match.
The key design decision was to keep the integration layer stateless. It does not hold business data of its own; every function reads from and writes to the existing systems. If the integration layer goes down, the three systems continue to work independently exactly as they did before. The worst case is a temporary return to manual data entry, not a business-wide outage. This meant the team could deploy the integration incrementally, one shipment type at a time, without risking the operation they depended on.
Manual data entry between the three systems was eliminated entirely. Shipments now flow from warehouse to courier to invoice without a single copy-paste. The reconciliation dashboard catches mismatches in near real time; wrong tracking numbers, missed deliveries, invoices that do not line up; and flags them before a customer ever sees the discrepancy. The whole thing runs for roughly the cost of a couple of coffees a month in AWS charges. No new platform to learn, no migration to manage, no vendor lock-in to worry about.
Systems that need to talk? Write plainly.
info@seijakulabs.com