Insights — Digital twins
Digital twins: from concept to implementation.
“Digital twin” is one of the most used — and most abused — phrases in the built environment. Stripped of the hype, a digital twin is simply a digital representation of a physical asset that stays connected to reality, so decisions about the asset can be made from the model rather than a site visit. Here is how to get from concept to something that earns its keep.
The maturity ladder
Most organisations do not need — and should not start with — a fully instrumented, sensor-fed twin. Value arrives at every rung of a maturity ladder:
- Digital record: an accurate, navigable capture of the asset — point cloud, mesh or 360° tour. Static, but true.
- Structured model: the record rebuilt as an information model (BIM), with elements classified and data attached — maintainable and queryable.
- Connected twin: the model linked to live or regularly refreshed data — condition surveys, asset registers, IoT sensors — so it evolves with the building.
- Predictive twin: analytics layered on top, supporting predictive maintenance and scenario planning.
Each rung delivers usable value on its own. The commonest failure mode is aiming for rung four while standing on inaccurate drawings — which is rung zero.
Start with truth: the capture phase
Everything above rests on the base geometry being correct. Terrestrial laser scanning and UAV photogrammetry establish that truth quickly: professional-grade LiDAR captures one million points per second, and a combined aerial-plus-terrestrial campaign can record a full estate in days rather than months. Captured once and geo-referenced properly, that dataset serves the drawings, the model and the eventual twin.
Structure with intent: the modelling phase
Scan-to-BIM converts point clouds into intelligent models — but “model everything to maximum detail” is a budget-burning mistake. Definition should follow use: a facilities team planning maintenance needs different information from a design team planning a refurbishment. This is where ISO 19650 earns its place — defining information requirements up front so the model contains what its users need, no more and no less.
Connect with purpose: the operational phase
A twin becomes operational when it is wired into the workflows that manage the building: the asset register, the maintenance system, condition data, and where justified, live sensors. Two principles keep this phase honest:
- Single source of truth. The twin must be where people actually look — if answers still live in spreadsheets, the twin is decoration.
- A plan for currency. Buildings change. Without a process for recording alterations against the model, a twin decays into just another out-of-date drawing set.
What the return looks like
Measured benefits from our client work and the wider industry cluster in four areas: faster access to reliable information (we typically see 30–60% faster data insight once records are digital and centralised); reduced rework from decisions made on bad data; fewer site visits, because questions are answered in the model; and stronger compliance evidence, with the twin doubling as the Golden Thread record for building safety purposes.
A pragmatic route in
- Pick one building where poor information is causing visible pain
- Capture it properly — aerial and terrestrial as needed — and stand up a navigable digital record
- Define information requirements, then model only to the level those requirements justify
- Connect the model to your existing asset data before considering new sensors
- Review what changed operationally, then scale to the next asset
We have delivered every rung of this ladder — from virtual estate demonstrators for defence and manufacturing clients to condition-connected virtual records for commercial HQs. If you are weighing up where a twin would pay for itself in your estate, we can help you find out.