Is a 3D scan a digital twin? What 9 practitioners say

At a glance

What practitioners mean by a digital twin

Barry Bassnett, a digital twin practitioner and cartographer, uses the strict definition: a twin lets you ask “what if”, it is interoperable, and information flows both ways. Kelly Watt, who led the operational digital twin programme at Dallas Fort Worth airport, follows the Digital Twin Consortium’s maturity levels: the point of a twin is a decision – predicted, simulated or automated.

Nikhil Sawlani of MultiSet AI puts the emphasis on “live”: a twin should update as reality changes and unlock more value than a 3D view. Filip Fejdi of WOOW Studio admits there are many definitions – and uses the word for something that doesn’t exist yet (see the last section).

Point cloud, mesh, Matterport tour, BIM model: twin or not?

Barry Bassnett’s verdict: not a twin, but a visual twin – useful for training and familiarisation, yet still a “3D postcard” until information is attached. Kelly Watt agrees a Matterport capture is not a twin on its own, even though the DFW programme used around 1,400 Matterport projects tied to GIS as one component. In Kelly Watt’s analogy, the scan is the bag of flour, not the cookie.

A BIM model on its own isn’t a twin either, per Kelly Watt – but overlaying design and scan for clash detection, adding attributes, linking project management and costs, then machine learning and what-if scenarios gets “damn close”. Nikhil Sawlani’s version: if you stop at the static point cloud, you are one step behind.

What turns a scan into a twin

The scan is the easy part, says Barry Bassnett; the work is people “harvesting” knowledge and tagging it to the model so things can be found – where an asset is, when it was captured, what it is and why it matters. Kelly Watt’s test is that combining the data must add meaning: one plus one has to be more than two.

Connection is the other half. Filip Fejdi’s twin syncs both ways with the developer’s CRM; Barry Bassnett links the model to the documents and emails about each pump. Kelly Watt warns not to assume the data exists: at DFW the team often had to install sensors first, and “real time” is itself a slippery term.

Start with the business problem, not the scan

Both main guests put the problem first. Kelly Watt’s checklist: what is the business problem, who is the stakeholder, which human workflow becomes digital, and can the ROI be quantified – some problems don’t need 3D at all. Barry Bassnett’s advice is to solve one problem, because that earns the right to solve the next.

Barry Bassnett’s paper-mill example shows how small that can start: a 360° tour made for marketing became a maintenance tool once every grease and oil point was tagged, and the company now wants it across its global network – for that one use. At DFW, Kelly Watt says, the twin targeted technicians’ wrench time and ended up used by around 350 people a day. Barry Bassnett’s sales line: you’re selling time saved, not a twin.

How accurate does a twin need to be?

Less than surveyors expect. Barry Bassnett says facility users mostly need to know which room an item is in, that a twin can even be built from 360° imagery, and that accuracy drives cost – though a multi-storey twin must not “twist”. Kelly Watt captured DFW’s plant rooms with Matterport because dimensional accuracy wasn’t needed there, and says training needs only visual accuracy.

Accuracy still matters where it matters: Kelly Watt recalls a capture that ended up 14 feet off because survey control wasn’t used properly. Rami Tamimi makes the same distinction from the surveying side – asset inventory doesn’t need the highest accuracy, design work does.

Keeping it alive

A twin that isn’t updated becomes a historic document, says Barry Bassnett – and for firefighters relying on an old layout, a potential fatality. Kelly Watt’s answer is to treat reality capture as a funded programme rather than a project, rescanning the areas that change instead of the whole plant, because project data otherwise ends up in a “data graveyard”.

Others describe where this is heading: Ben Williams of Exyn Technologies talks about continuous digitisation by autonomous systems that keep the model current, and Zlata Karpushina of FJ Dynamics says scanning’s long-term value for owners is a reliable digital record of the building.

When definitions collide

Does a twin need 3D? Barry Bassnett says no – it could be a piece of paper or a program, and 3D is just the most engaging interface. Kelly Watt says a twin without any visual representation is “absolutely not” a twin, but something like system monitoring.

In practice the word stretches further still. Filip Fejdi’s WOOW Studio builds twins of apartments that don’t exist yet – exact replicas synced with live sales data. Bilawal Sidhu calls a live layer of every plane on the planet a digital twin, with the 3D globe as a backdrop. Nikhil Sawlani defines a twin as live, yet uses existing scans as the twin MultiSet’s navigation runs on. Kelly Watt’s conclusion: not everyone is on the same page – which is exactly why buyers should ask what they’re getting. Arek Szadkowski adds Esri’s angle: the twin has to be “living” – updated, fed by real-time data and sensors – rather than a static model.

Questions

Is a point cloud a digital twin?

Not on its own, according to Barry Bassnett and Kelly Watt: it is a visual twin or an ingredient. It becomes part of a twin when information is attached, it is connected to business data and kept up to date.

Is a Matterport tour a digital twin?

Kelly Watt calls it a component, not a twin – yet used around 1,400 Matterport projects as part of the DFW airport twin. Barry Bassnett says it is fine for short-term jobs, but wants data ownership and an exit strategy for a long-term twin.

What is the difference between a BIM model and a digital twin?

Per Kelly Watt, a BIM model is an interpretation; combined with reality capture, attributes, project management, costs and what-if scenarios it gets very close to a twin. Barry Bassnett says facility managers rarely open BIM files at all.

How accurate does a digital twin need to be?

Only as accurate as the use case requires: Barry Bassnett and Kelly Watt both say most operational uses need far less than scanner accuracy, while survey control still matters at scale.