Operations prototype

Clipboard rounds into work orders

Utility readings taken on paper, typed in later, and analysed never. I prototyped the whole chain — the technician's round, the alert, the trend, and the write-back into the maintenance system.

The technician in the plant and the executive in the office, in one prototype
2 audiencesThe technician in the plant and the executive in the office, in one prototype
The whole chain shown, not just the dashboard at the end
Reading → work orderThe whole chain shown, not just the dashboard at the end
The maintenance system dependency shown as a status surface, not assumed
Integration visibleThe maintenance system dependency shown as a status surface, not assumed

The client

A large dairy manufacturer where plant utility systems — the readings that predict a failure days before it happens — were captured on paper rounds and rekeyed, with a maintenance system that never saw them.

The engagement

A prototype covering both ends of the same process: the technician capturing readings in the plant and the executive seeing the consequence, with the enterprise maintenance integration shown as a status surface.

A working demo of this build exists.

It is not public yet — ask for access, and where the NDA allows I will send a link or walk you through it on a call.

Request demo access

The problem

Digitisation projects in manufacturing die between the clipboard and the dashboard. Someone builds a beautiful analytics layer, the readings still arrive on paper, and the insight never reaches the maintenance system where work actually gets scheduled. Demonstrating only the dashboard reinforces exactly the wrong assumption about where the difficulty lies.

What I did

I prototyped both ends and the join. The technician's interface exists because if capture does not change, nothing downstream matters — and it is designed for someone standing next to equipment rather than sitting at a desk. The executive surfaces exist because that is who funds it. And between them sits the alert centre and the integration status view, which is where I deliberately spent credibility: rather than implying that data flows into the enterprise maintenance system, the prototype shows that integration as a thing with a state, which is honest about it being the hardest and least glamorous part of the delivery. Data generation covers a realistic reading history so trends look like plant data rather than like a sine wave.

What was built

A prototype with two distinct surfaces — a mobile interface for the technician's round against an equipment catalogue with readings, and a desktop set for the plant and executive audience covering an operations dashboard, an alert centre and trend analytics — plus an integration status view for the enterprise maintenance system, so the write-back is visible as a real dependency rather than implied.

On the table at the end

  • Technician mobile round interface and equipment catalogue
  • Executive dashboard, alert centre and trend analytics
  • Maintenance system integration status surface
  • Mock data generators covering the full reading history

What it changed

Showed the complete chain rather than a dashboard: a reading taken on a round becomes an alert, becomes a trend, becomes a work order in the system maintenance teams already use — which is the only version of this story a plant director can act on.

How it ran

  1. 01

    Start at capture

    The technician's round interface built first, because an analytics layer over a clipboard changes nothing.

  2. 02

    Give the executive their view

    Operations dashboard and trend analytics for the audience that funds the programme.

  3. 03

    Put alerts in the middle

    An alert centre as the connective tissue between a reading and a decision.

  4. 04

    Show the integration as a dependency

    The enterprise maintenance connection given its own status surface rather than implied as automatic.

  5. 05

    Generate realistic history

    Reading data modelled to look like plant behaviour, so the trends are arguable rather than decorative.

Something similar on your plate?

Thirty minutes, no deck. I will tell you whether it is worth doing at all.