A point cloud to BIM service takes scan data someone else already captured and turns it into a coordinated Revit model. That distinction matters, because a lot of firms search for “scan to BIM” when what they actually have is a finished point cloud sitting on a drive and a modeling gap they need filled, not a scanning gap.

Surveying firms run into this constantly. So do contractors who commissioned a scan for one purpose and later need it modeled for another. If you already have the cloud, you do not need a company that also owns scanners. You need a modeling team that can pick up where the scan left off.

What the Service Actually Covers

Point cloud to BIM modeling means building a native Revit file from scan data across whichever disciplines the project calls for: architectural and structural elements first (walls, floors, roof framing), then mechanical, electrical, and any process piping, each modeled as its own linked file. The last step pulls every discipline into one coordinated container and checks the geometry against the source cloud.

What you do not get from a modeling-only service is the scan itself. If you need both capture and modeling, that is two services, even when one provider offers both under one roof.

What You Send: Formats That Actually Work

Revit reads point clouds natively in two formats: RCP and RCS, Autodesk’s own container and scan formats. Everything else, including the open, vendor-neutral E57 standard, has to run through Autodesk ReCap for indexing before Revit can link it. Send whichever format your scan came in. A modeling team that has done this before will index it themselves rather than bouncing the file back to you.

The Modeling Process, Discipline by Discipline

Here is where the actual work happens, and where the differences between providers show up.

Architecture and structure come first. Floor slabs, walls, columns, and roof framing get modeled directly against the cloud, because everything else in the building references these elements for position.

Mechanical, electrical, and process piping follow. Pumps, ducts, conduit runs, cable trays, and piping each get modeled as real Revit families with actual connection points, not generic boxes standing in for equipment. The level of detail here depends on what the model needs to do afterward: coordination-only work needs accurate connection points and routing; fabrication or procurement work needs dimensionally accurate, connected systems.

Coordination happens last. Every discipline gets linked into one federated file, georeferenced to the project’s coordinate system, and checked for clashes between systems: a duct crossing a structural beam, a pipe run with no clearance for a support hanger.

Software gets you through the mechanical parts of this. It does not tell you whether a wall reads out of plumb because it was built that way or because the scan picked up noise at that surface. That call, made correctly and consistently across a few thousand elements, is most of what separates a usable model from one that is technically accurate and practically useless.

Where the Modeling Hours Actually Go

An open floor plan with a handful of walls models fast. A mechanical room with dense piping, supports, and multiple systems crossing the same ceiling space takes meaningfully longer per square foot, because the hours scale with how much has to be interpreted and connected, not how much square footage the room covers. A quote that does not ask about MEP density before pricing the job is usually guessing.

QA: Checking the Model Against the Source

A finished model gets checked against the point cloud it came from, not just reviewed for whether it looks right in a 3D view. That means confirming geometry accuracy against the cloud, running a clash check between disciplines, and cleaning out unused families and views before delivery. Skipping this step is how a model ends up looking complete while quietly drifting from what is actually built.

What You Receive

A typical deliverable includes the native Revit file (.rvt), export formats like .ifc and .dwg for teams not running Revit, and often 2D sheets, plans, sections, elevations, generated directly from the coordinated model. Some projects also want a QA report showing the model was validated against the source scan rather than handed over on faith.

When This Is the Right Service to Look For

If you are a surveying firm that already owns the scan and the client relationship, a point cloud to BIM partner fills the Revit gap without you building a modeling team for volume that may not be steady. If you are a contractor or facility team with an old scan and a new use case, the same logic applies: the capture already happened, and what is left is turning it into something Revit can use.

If you are still deciding between a flat CAD deliverable and a full parametric model, that is a separate question worth answering first, since it changes what the modeling scope needs to include.

Have a Point Cloud Ready to Model?

If you already have scan data and need it turned into a coordinated Revit model, tell us about the project and we will scope it against the disciplines and LOD it actually needs.

A client sends over a scan and asks for “the drawings.” Another asks for “a BIM model.” A third just says “whatever you’d recommend.” All three might need completely different deliverables from the exact same point cloud, and the wrong guess costs real money on a project that size.

Scan to CAD and scan to BIM start from the same data and solve different problems. Here is what actually separates them, and how to tell a client which one they need before the quote goes out.

What Is Scan to CAD?

Scan to CAD takes a 3D point cloud and flattens it into 2D drawings, usually in AutoCAD. A common method is slicing the cloud horizontally at a fixed height, often around four feet above finished floor, then tracing walls, columns, and openings exactly where the scan recorded them. The result is a clean .dwg file: floor plans, elevations, sections, the same drawing types you would get from a traditional field survey, just measured to millimeter accuracy instead of a tape and a clipboard.

What you get is flat by design. A scan-to-CAD floor plan will not tell you if a beam sags or a duct runs at an angle above the ceiling. It shows what a slice through the building looks like at that height, nothing more.

What Is Scan to BIM?

Scan to BIM takes the same point cloud and builds it back up as a 3D parametric model, almost always in Revit. Every wall, pipe, and duct becomes an actual object with real properties: material, dimensions, connections to the elements around it. Nothing about a BIM model is flat, and nothing in it is just a line pretending to be a wall.

The Real Difference Is What the File Knows About Itself

A CAD drawing and a BIM model can look almost identical in a printed plan view. The difference lives underneath the geometry. A line in a CAD file is a line. A wall in a BIM model knows it is a wall: what it is made of, how thick it is, which room it separates from which, and what happens to it if you change the floor above.

That extra layer of data is exactly what makes clash detection possible, what lets a model generate an accurate quantity takeoff, and what a facilities team needs years later to plan a renovation without rescanning the building. None of that works on a file that only knows where its lines are.

When Scan to CAD Is the Right Call

CAD is the right deliverable more often than the BIM-first crowd likes to admit. It fits permit applications, lease and area calculations, simple as-built floor plans, and straightforward renovation layouts where nobody needs to coordinate mechanical, electrical, and structural systems against each other. If the project is a single-discipline space plan or a documentation requirement with a fixed format, a full parametric model is usually more than the job needs.

When You Need Scan to BIM Instead

BIM earns its cost on anything with real coordination risk: buildings with dense MEP systems, multi-discipline renovations, facility management programs that will reference the model for years, or any project where clash detection between trades carries real cost if it waits until construction starts. If the deliverable needs to answer questions beyond “where is this wall,” CAD will not get you there.

Cost and Timeline: What Actually Changes

CAD is faster and less expensive to produce, because tracing geometry takes less time than modeling intelligent, connected elements. BIM costs more upfront and takes longer, but that cost buys clash detection, quantity data, and a model that keeps paying off long after delivery. Neither is the “better” option in the abstract. The right one depends entirely on what the client does with the deliverable after they receive it.

You Do Not Have to Choose Before You Scan

The scan itself does not care which deliverable comes out of it. The same point cloud can produce a CAD package now and a full BIM model later, without sending anyone back out to rescan the building. If a client is not sure yet whether they will need full 3D coordination down the line, starting with CAD and upgrading later works well and costs nothing extra in rescanning.

Not Sure Which One Your Project Needs?

If you have point cloud data and are not sure whether the project calls for CAD, BIM, or both, tell us about it and we will help you figure out the right scope before you commit to either.

If you run a surveying firm, you have probably heard the phrase “scan to BIM” more often this year than in the last five combined. Clients are asking for it by name. Some RFPs now require it. And most surveyors already own the hard part: the point cloud. What they do not always know is what happens after the scan, or whether they should be the ones doing it.

This guide covers what scan to BIM actually is, how a point cloud becomes a working Revit model, what LOD and LOA mean and why they are not the same thing, and what it looks like to offer this service without hiring a modeling team from scratch.

What Is Scan to BIM?

Scan to BIM is the process of turning a 3D laser scan of an existing building into a Building Information Model, usually built in Autodesk Revit. The scan captures what is physically there. The model interprets it: walls become wall objects, pipes become pipe objects, and each one carries real data (material, dimensions, system type) instead of just a shape.

That distinction matters more than it sounds. A point cloud is measured geometry. A BIM model is geometry someone has read, understood, and rebuilt as intelligent building elements. Getting from one to the other takes thousands of individual modeling decisions across the length of a project.

From Laser Scanner to Point Cloud

Most scan to BIM projects start with a 3D laser scanner recording millions of individual points per second, each with its own X, Y, and Z coordinate. Stitched together, those points form the point cloud: a dense, accurate, but entirely raw representation of the building. Nothing in a point cloud knows it is a wall yet. It is just measured space.

The scanning team registers and cleans up the scan positions, then typically hands off the cloud in an open format like .e57, or as an Autodesk-native .rcp file that Revit can reference directly. Either way, the cloud itself becomes the ground truth every modeling decision downstream has to match.

Turning a Point Cloud Into a Revit Model

A modeler works through the cloud discipline by discipline. Architectural and structural elements come first: floor slabs, walls, roof framing, doors, windows. Mechanical, electrical, and any process piping follow, usually modeled as their own linked files. The last step is coordination: pulling every discipline into one federated model, checking for physical clashes between systems, and confirming the geometry actually lines up with the scan it came from.

This is where experience shows. Software can trace a point cloud. It cannot tell you whether a wall is out of plumb because it was built that way or because the scan has noise in it. That judgment call, made correctly, is most of what separates a usable model from a technically accurate but practically useless one.

LOD and LOA: The Two Numbers That Actually Matter

These two terms get confused constantly, including in RFPs written by people who should know better.

LOD (Level of Development) describes how much detail and information a model element contains, on a scale most commonly running from LOD 100 (rough massing, early planning) to LOD 500 (verified as-built, field-confirmed). A wall at LOD 200 might just be a generic wall type with approximate thickness. The same wall at LOD 400 carries real material layers, accurate dimensions, and enough detail to fabricate from.

LOA (Level of Accuracy), defined by the USIBD’s LOA specification, describes something completely different: how closely the model’s geometry matches the physical building, measured in a tolerance range. LOA 20 typically covers 15mm to 5cm, suitable for early feasibility work. LOA 30 (5mm to 15mm) is the common target for renovation and MEP coordination projects. LOA 40 (1mm to 5mm) is reserved for precision fabrication work where a few millimeters of error actually breaks something.

A model can be high-LOD and low-LOA (very detailed, but not verified against reality), or the reverse. Specifying only one and assuming the other is where a lot of scope disputes start.

What You Actually Receive as a Deliverable

A finished scan to BIM project typically hands over more than one file. Expect the native Revit model (.rvt), export formats like .ifc and .dwg for anyone not running Revit, and often the cleaned point cloud itself for archival or later verification. On top of the model, you can usually pull 2D drawings (plans, sections, elevations), quantity takeoffs, and a clash detection or QA report showing the model was checked against the source scan rather than just handed over as-is.

Why More Surveying Firms Are Adding Scan to BIM

Surveying and BIM used to sit in separate lanes. That line has been getting blurrier for a few years, for a simple reason: the client asking for a scan is increasingly the same client who needs a Revit model six weeks later, and would rather not manage two separate vendors to get it.

For a surveying firm, that creates a real choice. Turn down the BIM portion of the work and risk losing the whole engagement to a competitor who can offer both. Build a Revit modeling team in-house, which means new hires, new software licenses, and a learning curve that does not pay for itself on one or two projects a quarter. Or partner with a modeling team that handles the Revit side while your firm keeps the client relationship and the scan work you already do well.

Building In-House vs. Partnering

In-house modeling makes sense at real volume, once you have enough steady BIM work to keep a modeler or two busy full time. Below that threshold, the math rarely works: Revit expertise takes time to build, and idle capacity between projects costs money whether or not anyone is modeling that week.

A white-label modeling partner solves the volume problem without the overhead. You send the scan, a modeling team builds the Revit file, and it comes back ready to deliver under your own name. Your client never needs to know a second company touched the project.

Have a Scan Sitting on a Drive Right Now?

If you already have point cloud data and need it turned into a coordinated Revit model, that is exactly what we do. Tell us about the project and we will scope it, quote it, and model it, white-label, under your firm’s name.