Freight OS
A single operating system for 240 vehicles, 14 depots, and the clients who need to know where their freight is.
- Client
- Atlas Logistics Group
- Industry
- Logistics & supply chain
- Discipline
- Enterprise
- Duration
- 9 months, 2026
The problem
Atlasrananationalfleetonfourdisconnectedsystemsandasetofspreadsheetsthatonlytwopeoplefullyunderstood.Dispatchersrekeyedthesamebookinguptothreetimes,andnobodycouldansweraclient'sstatusquestionwithoutthreephonecalls.
Month-end reconciliation took eleven working days. Growth had stopped being a commercial question and become an operational one: the business could not take on more volume without adding administrative headcount at the same rate.
What we did
We spent six weeks in depots and dispatch offices before writing production code, mapping the real process rather than the documented one. The resulting platform put bookings, allocation, proof of delivery, and invoicing behind one domain model with a single source of truth for load state.
Dispatch got a control tower with live exception queues instead of a list of everything. Drivers got a lightweight mobile flow that works without signal and syncs when it returns. Clients got a tracking portal, which quietly removed most of the inbound status calls.
Everything runs on Terraform-managed infrastructure with tracing across services, so an incident is diagnosed from a dashboard rather than reconstructed from memory.
Services applied
Technology
- Next.js
- TypeScript
- Node.js
- PostgreSQL
- Redis
- Kubernetes
- AWS
- Terraform
Gallery
Inside the build.
Results
What changed, measured after launch.
- Less admin time per booking
- 0%Less admin time per booking
- Days to close month-end
- 0 → 2Days to close month-end
- Volume handled, same headcount
- 2.4×Volume handled, same headcount
- Platform uptime since launch
- 99.98%Platform uptime since launch
“They understood our operation better than the consultants we had used for years. The difference is that they then went and built the thing.”
Thandiwe NkosiChief Operating Officer, Atlas Logistics GroupRelated work
Next case study.
Sentinel Migration
Fourteen years of on-premise insurance systems moved to Azure, without a single unplanned outage.
Vantage Field Ops
Work-order automation and predictive scheduling for 380 field technicians.
Underwriting Copilot
An evaluated AI assistant that reads submission packs and drafts the first underwriting opinion.
Similar problem?
If this sounds like your operation, let's talk.
We will tell you honestly whether the approach transfers, and what it would take.
Or email hello@bonangtech.com