Work

A sample of recent engagements. Several clients asked us not to use their names, so we describe the sector and the problem instead.

Logistics

Dispatch platform for a regional freight operator

62%
faster load assignment: 340
drivers on the app

Next.js
React Native
PostgreSQL
AWS

The problem

Three planners were assigning loads from a whiteboard and a shared spreadsheet that only one person could edit at a time. Drivers phoned in for their next job, and about a fifth of those calls came from places with no signal.

What we did

We ran a week on site watching the assignment process, then built a dispatch board that updates live for every planner and a driver app that queues everything locally and syncs when the connection returns. Rollout went depot by depot over five weeks.

Where it stands

Replaced a whiteboard and three spreadsheets with a live dispatch board and a driver app that works without signal. Planners now assign a load in seconds instead of minutes.

Manufacturing

Shop-floor visibility across four plants

4
plants on one system: 11 min
average time to spot a stoppage

Node.js
TimescaleDB
Grafana
Azure

The problem

Four plants each had their own way of recording output. Head office reconciled it weekly in a spreadsheet, which meant a stalled line could run unnoticed for most of a shift.

What we did

We put a small collector next to each machine controller, streamed readings into a time-series database, and built one dashboard with a single definition of what counts as running. Alerts go to the supervisor on shift, not to a mailing list.

Where it stands

Machine data, order status, and quality checks pulled into one dashboard. Supervisors see a stalled line the minute it stalls, not at the end of the shift.

Healthcare

Clinic migration off ageing on-premise servers

0
hours of unplanned downtime: 38%
lower monthly infrastructure cost

Terraform
Kubernetes
Azure
Veeam

The problem

Patient scheduling and records ran on two ageing on-premise servers with an expired support contract and a backup nobody had restored from in three years.

What we did

We inventoried every dependency, rehearsed the restore twice against a copy, then moved workloads across six scheduled weekends with a tested rollback at each step. Clinical staff kept working throughout.

Where it stands

Moved patient scheduling and records to a private cloud over six staged weekends, with a tested rollback at every step. No appointments were lost.

Your problem probably looks like one of these

If something in your business runs on a spreadsheet, a server nobody wants to touch, or one person's memory, we have seen it before. Tell us about it.

Start a project Email us directly