Digital twin development, starting from the decision it serves.
You have decided a twin is worth building. The questions left are who builds it, whether to build or buy, and what the first step costs. Here is how we answer all three — including when the answer is not us.
The pieces a twin is made of.
A digital twin is sensor data, a model and a view joined up. We have built each of those halves in production, for real clients.
Sensors to a live dashboard
EnerDuce MASHAN: IoT pressure sensors on LPG cylinders, calibration for environmental drift, a web dashboard for operators and alerts that schedule delivery before a cylinder runs out.
Read the EnerDuce case study →3D scenes, multi-tenant, real-time
ViXR: the multi-tenant backend behind an AR experience platform — 3D scenes and their assets normalised for Unity, and live collaboration inside 3D spaces — engineered 2021 through 2024.
Read the ViXR case study →Build a custom digital twin, or buy a platform?
A custom build is not always the right answer, and an agency that only sells builds will not tell you that.
Build a custom twin if
- Your asset or process is specific enough that an off-the-shelf platform would need heavy customisation anyway.
- The twin has to sit inside your own product or portal, under your brand, rather than in a vendor dashboard.
- Your data comes from a mix of sensors, legacy systems and spreadsheets that no single platform connects to out of the box.
- You want to own the model and the data pipeline, not rent them per asset.
Buy a platform if
- Your equipment vendor already ships a twin or a monitoring platform for it. Start there and extend it.
- You need an industrial platform your OT team already knows, and the use case fits its templates.
- You have no one internally to own the twin after launch. A managed platform will outlast a custom build nobody maintains.
What a digital twin build covers.
Scoped before anything is built, fixed-quoted before you commit.
The decision first
Which decision the twin exists to improve — a maintenance call, a delivery run, a capacity plan — and who acts on it. A twin with no decision attached is an expensive dashboard.
Data ingestion from what you already run
Sensors, PLCs, telemetry streams, ERP and maintenance records brought into one timeline. Most of the work in a twin is here, not in the 3D.
A model of the asset or process
State, rules and thresholds for the thing being twinned, calibrated against its real behaviour — so the twin disagrees with reality as rarely as possible, and says so when it does.
Visualisation sized to the job
A 2D operations view is often enough. 3D or XR is built when spatial context changes the decision, not because it demos well.
Alerts and actions
The twin raises the call before a person has to notice it — the same pattern as scheduling a delivery before a cylinder runs empty.
A proof of concept on one asset, fixed-quote
The first step is one asset or one line, end to end, for a fixed quote after scoping — before anyone commits to the whole estate. It surfaces the data problems while they are still cheap.
Frequently Asked Questions
Tell us what you want to twin.
The asset or process, where its data lives today, and the decision you want it to improve. We will tell you whether to build, buy, or start smaller.