BIM
Scan to BIM Services
Converting point cloud and laser scan data of existing buildings and structures into accurate, working Revit models.
Reality → data → model
Existing · illustrative
Scan stage
A Point Cloud Records the Building. A BIM Model Makes It Usable.
Renovation, retrofit and asset-documentation projects often start with a scanned building and no usable model — the scan data alone isn't something a design team can work in directly.
A point cloud is an extremely accurate record of what exists, but it's also an enormous, undifferentiated dataset that has to be interpreted and turned into discrete, modelled building elements before anyone can actually design against it.
We convert point cloud and laser scan data into coordinated Revit models — architectural, structural and MEP as required — giving design teams a working BIM model of an existing building or structure to design from.
Point cloud
Measured points
BIM model
Walls · floors · columns · doors · windows · MEP · structure
Design team
Works with the existing building as a model
No scan density or measurement accuracy is claimed here — both depend on the scan data and the project.
From Reality to Working BIM Model
Follow one building through the whole transformation: the existing building, the scan, the point cloud, registration, interpretation, the Revit model and what it's then used for.
The existing building, as it stands.
Switch Between the Measured Data and the Model
The point cloud and the model are two stages of the same transformation — same building, same geometry. Switch between them, or overlay one on the other.
From Measured Points to Building Elements
Reading the data is the modelling work: surfaces become walls, clusters become floors, vertical structures become columns, gaps become openings and visible services become MEP elements.
Not every visible object becomes a BIM element — the final scope depends on project requirements.
The Model Is Only as Good as the Registered Scan Data It Starts From
Scan registration and accuracy are checked before modelling starts, because a model built from poorly registered scan data will carry that error forward invisibly — it's much cheaper to catch a registration problem before modelling than to discover it later when a wall doesn't line up correctly between floors.
Yes, provided the individual scans share a common registration or control reference, we can combine data from multiple sessions into a single coordinated model, flagging any registration uncertainty we find between sessions.
Multi-Storey Buildings Need More Than Floor-by-Floor Modelling
Multi-storey buildings scanned in separate sessions carry a specific registration risk worth flagging: if each floor's scan was registered independently rather than tied back to a shared site-wide control, small misalignments between floors can go unnoticed until the model is assembled as a whole. We check for this specifically rather than assuming floor-to-floor consistency by default.
We check specifically for floor-to-floor registration misalignment in this scenario, since scans registered independently per floor rather than to a shared site-wide control can carry small offsets that only become apparent once the full model is assembled.
How Much of the Existing Building Should the Model Preserve?
Modelling from scan data is a judgement exercise as much as a technical one — real buildings are rarely perfectly plumb, level or square, and we agree with you upfront how much of that as-built irregularity to capture versus how much to idealise for design purposes, since modelling every deviation exactly as scanned can make a model needlessly difficult to design against.
Real buildings are rarely perfectly plumb, level or square. The decision isn't always to copy every deviation — it's a balance between as-built fidelity and design usability, agreed per project.
What does the project need?Often favours keeping measured irregularity. Neither end is automatically right — the choice is agreed per project.
When Existing Character Matters
For heritage or architecturally significant buildings, we can model to a higher level of geometric fidelity where the project calls for it, capturing decorative or non-standard elements that a purely functional as-built model would typically simplify away.
Yes, for architecturally significant or heritage buildings we can capture decorative and non-standard elements to a higher geometric fidelity where the project calls for it.
- Decorative moulding
- Irregular opening
- Ornamental element
- Non-standard geometry
Required fidelity is project-specific. No heritage certification or conservation approval is implied.
Extract the Measurements the Structural Team Actually Needs
Where the point cloud is being used to inform structural assessment — checking for out-of-plumb columns or deflection in an existing structure, for instance — we work closely with the structural engineer to confirm which measurements from the scan they actually need extracted and to what precision, rather than delivering a generic model and leaving that extraction work for later.
Yes, where the model is being used to check for out-of-plumb columns or existing deflection, we work directly with the structural engineer to confirm which specific measurements they need extracted from the scan, and to what precision.
- Out-of-plumb column
- Deflection
- Alignment
- Existing geometry
Where the scan informs structural assessment, the required measurements are coordinated with the responsible structural engineer, who decides what they mean.
Architectural, Structural and MEP — From the Same Point Cloud
Yes, where the scan data supports it, all three disciplines can be modelled from the same point cloud.
For ongoing Revit model creation and updates see Revit modelling; for multi-discipline federation and clash detection, BIM modelling & coordination.
- Walls
- Floors
- Doors
- Windows
- Roofs
- Spaces
Scope depends on project requirements and on what the scan data actually captured.
Give the Design Team a Model of What Already Exists
This is commonly used for renovation and retrofit projects, heritage documentation, and facilities teams that need an accurate as-built model of existing assets.
Yes, an accurate as-built model is a common and valuable starting point for adaptive reuse projects, where understanding the exact existing condition shapes what's actually feasible in the new design.
Retrofit Starts With Knowing the Existing Condition
Existing structural frame, services and walls form the base; the proposed intervention is drawn against them. Existing and proposed are shown with different line styles — not bright colours.
- Existing — solid line
- Proposed — dashed line
Retrofit starts with knowing the existing condition; the as-built model is the foundation the intervention is designed against.
Turn Existing Assets Into a Working Digital Record
- 1Existing facility
- 2Scan
- 3BIM model
- 4Asset documentation
A working digital record of existing assets. Full facilities-management functionality is not assumed unless separately scoped.
Dense Industrial Environments Need a Different Kind of Existing-Condition Model
The Model Goes Back Against the Scan
Model checked against the scan for dimensional accuracy before delivery.
Overlay is a conceptual view of how measured data and model geometry relate. Automated deviation analysis is only included if a project specifically calls for it.
What You Get
Revit models from point cloud / scan data
As-built architectural models
As-built structural models
As-built MEP models
Model-derived 2D drawings
Coordinated federated models
Scan registration & accuracy verification notes
Where Scan to BIM Is Used
Renovation and retrofit projects
01
Heritage building documentation
02
Facilities and asset management
03
Existing structure documentation for design teams
04
Plant and industrial facility as-built modelling
05
From Scan Data to a Working Revit Model
- 1Scan data
- 2Registration check
- 3Modelling scope
- 4Revit modelling
- 5Scan comparison
- 6As-built model
- 7Drawings / coordinated model
01
Scan data intake
Point cloud or laser scan data supplied, along with the modelling scope and LOD target, and a check of scan registration accuracy before modelling begins.
02
Scope & quote
Discipline coverage and deliverables confirmed before modelling starts, along with how much as-built irregularity should be captured versus idealised.
03
Modelling
Revit model built directly from the registered scan data, with judgement calls on idealisation documented as they're made.
04
Review
Model checked against the scan for dimensional accuracy before delivery.
05
Delivery
As-built model delivered, with drawings produced on request.
Software We Use
Industries We Support
Related BIM Projects
Illustrative examples, marked as such. They will give way to verified case studies as client work is cleared for publication.
Scan to BIM FAQs
This service covers conversion of supplied point cloud/scan data into a BIM model. Let us know if you need scanning arranged and we can advise on scope and suitable providers.
Accuracy depends on the quality and density of the supplied scan data, and on the registration accuracy we verify before modelling starts — this is discussed during scoping against your intended use for the model.
Yes, where the scan data supports it, all three disciplines can be modelled from the same point cloud.
This is agreed with you upfront — real buildings are rarely perfectly square or level, and we model as much of that irregularity as your intended use for the model actually requires, rather than defaulting to one extreme or the other.
Yes, for architecturally significant or heritage buildings we can capture decorative and non-standard elements to a higher geometric fidelity where the project calls for it.
We check specifically for floor-to-floor registration misalignment in this scenario, since scans registered independently per floor rather than to a shared site-wide control can carry small offsets that only become apparent once the full model is assembled.
Yes, where the model is being used to check for out-of-plumb columns or existing deflection, we work directly with the structural engineer to confirm which specific measurements they need extracted from the scan, and to what precision.
Yes, provided the individual scans share a common registration or control reference, we can combine data from multiple sessions into a single coordinated model, flagging any registration uncertainty we find between sessions.
Yes, where original design drawings are available alongside scan data, we can produce a comparison highlighting where the as-built condition has diverged from the original design.
Yes, an accurate as-built model is a common and valuable starting point for adaptive reuse projects, where understanding the exact existing condition shapes what's actually feasible in the new design.
We can work with common point cloud formats such as E57 and RCP/RCS, along with registered scan data exported in other standard formats — confirm your specific format during scoping.
Yes, where MEP elements are visible and captured in the scan data, they can be modelled alongside architectural and structural elements as part of the as-built model.
This depends heavily on building size, complexity and the LOD agreed for the model, and is confirmed as part of your quote once scan data and scope are reviewed.
Yes, scope can be limited to whichever floors or areas your project actually needs modelled, rather than requiring the whole building to be captured and modelled at once.
Where a previous scan or survey exists for comparison, we can produce a deviation analysis highlighting change over time, which is useful for monitoring structures suspected of settlement or movement.
Yes, an accurate as-built model of an existing tenancy shell is a strong starting point for fit-out design, particularly where existing services and structural elements need to be worked around precisely.
This varies by intended use, but LOD 200–300 is common for general design and renovation purposes — higher LOD is scoped specifically where a project needs fabrication-level as-built accuracy.
Yes, an accurate as-built model can support a condition assessment exercise, though the assessment judgement itself would sit with the relevant qualified professional reviewing the model.
Yes, we can supply a lightweight point cloud viewer file alongside the Revit model so your team can independently spot-check the model against the original scan data at any time.
Yes, staging modelling delivery across a large facility, prioritising the areas your project needs first, is a practical way to keep an ongoing large-site project moving without waiting for the entire facility to be modelled at once.
Yes, where the scan reveals a structural element noticeably out of plumb or level beyond what would normally be expected, we flag it for the structural engineer's attention rather than silently modelling around it.
Yes, scan-to-BIM conversion extends to civil and infrastructure structures such as bridges, provided suitable scan data is available.
Yes, a short summary noting anything notable found during modelling — significant as-built deviations, areas of lower scan confidence, or structural observations worth flagging — is a useful accompaniment to the model itself.
Yes, an accurate as-built model produced during a renovation project has ongoing value for facilities management well after construction finishes, provided it's structured with that longer-term use in mind from the start.
Yes, scope can be as small as a single room where high-precision as-built information is needed for a specific design decision, rather than requiring a whole-building modelling exercise.
Yes, where the scan reveals a physical condition inconsistent with available design records, we note the discrepancy so your team can decide how to address it, rather than silently reconciling the two.
Yes, dense industrial and process plant environments are a common scan-to-BIM use case, and modelling depth is agreed based on which systems your project actually needs represented.
Start a project
Turn Your Existing Building Into a Working BIM Model
Tell us what you need and our team can review the scan data, modelling requirements and project scope.