3D Scan to CAD: Workflow, Outputs, and When It Makes Sense

3D scan to CAD turns measured laser-scan data into CAD deliverables: DWG floor plans, reflected ceiling plans, elevations, sections, roof plans, and the PDF sheets that go with them. The point cloud is the measured source; the CAD package is the interpreted drawing built from it. The gap between them is drafting labor, not a button.

The question that scopes the job is not “can the building be scanned.” It almost always can. The question is what you need after the scan: 2D drawings, an intelligent Revit model, a mesh, or a reverse-engineered part. Scan-to-CAD is the right call when your downstream team needs clean, dimensionally honest drawings of existing conditions and does not need a coordinated BIM environment.

Start with the output, not the scan

The first decision is the final file, because it sets everything upstream. A team that needs DWG backgrounds for an architect, a permit set, lease plans, or facade linework wants CAD. A team that needs model elements, levels, schedules, or clash coordination wants scan-to-BIM instead. Picking wrong is expensive both ways: a flat DWG will never behave like a parametric model, and a full Revit model is more than most documentation jobs need. CAD is not the lower-tier choice; it is a different deliverable, and for much existing-conditions work it is the cleaner one. If you are weighing the two, scan-to-CAD vs scan-to-BIM walks the decision in one screen.

What the workflow actually looks like

A scan-to-CAD job runs in a predictable order, and most of the time is spent away from the part that looks like a drawing. It starts with scope (below), then field capture. WeAre Capture runs a Trimble X7 on site; a full scan day is up to roughly 12 hours depending on building size, occupancy, and how many scanner positions the space demands. Every column, soffit, and blocked doorway is another setup, so a multi-floor building with mechanical rooms needs far more stations than a single suite.

Next is registration: aligning every scanner position into one coordinate-consistent cloud. People underestimate this step because it happens before anything looks like CAD, yet if registration is loose the linework gets traced over evidence that has already drifted, and clean drafting cannot fix that. We register, export to E57 (and RCP/RCS for Autodesk workflows), clean stray returns, split the cloud by floor, and cut the plan and section slices the drafter will trace. The post on registered vs unregistered point clouds explains why a file can open fine and still be unusable for drawing production.

Drafting is interpretation. The drafter cuts plans and sections at the right heights, then draws walls, doors, windows, stairs, shafts, columns, and the in-scope fixtures on your layer standard. The cloud supplies measured geometry; a usable drawing still needs deliberate linework, annotation, and sheet organization. Internal QA then checks the DWG back against the cloud (wall alignment, opening locations, floor-to-floor heights, missing rooms), you review, and we deliver DWG plus PDF.

Step 1: scope the drawing package

Before anyone scans or drafts, the CAD scope has to be written down, because “scan to CAD” silently covers two different jobs.

For a building, scope means the floors and areas, the exterior and ceiling coverage, how much MEP should appear (one line of “MEP visibility” can swing a quote), drawing scale, layer and titleblock standards, file formats, and the deadline. For an object or component, it means target CAD format, surface detail, feature and hole locations, profile sections, tolerance expectations, and whether the result feeds manufacturing or reference. That second path is reverse engineering, a different discipline with different tools and review criteria; see reverse engineering services. Naming which job you have, in the first email, is the single biggest thing that makes a quote accurate.

Step 2: scan it, or hand us a cloud

The scan data comes from a field visit or from files you already have. When WeAre Capture captures it, we cover rooms, corridors, stairs, facades, roofs, mechanical spaces, and ceiling conditions; if you want the 3D laser scanning scope on its own, it can be billed as a field-only deliverable.

Client-provided clouds work well but need review before drafting, because the data inherits whatever the capture was built for. A walkthrough or visualization scan can look great on screen and still miss enough ceiling, facade, and room-corner geometry to break drawing production. So we check the file format, whether the cloud is registered, the support folder, the coordinate setup, the capture date, and known blind spots before we confirm scope, rather than build CAD on gaps.

How detailed should the CAD be?

There is no single right level. CAD lacks the LOD ladder that BIM has, but the principle from USIBD Level of Accuracy (LOA) still applies: scope to how the drawing will be used, not to a vague idea of “accurate.” A demolition background needs less than a facade-restoration elevation. The tiers below are typical scope, not a graded standard.

More overhead and equipment detail means more for the cloud to capture and more for the drafter to interpret, which is exactly where cost moves. There is no neutral published $/sqft here; price tracks element density and detail, not floor area. The honest ranges live in the as-built drawings cost post.

Where scan-to-CAD fits, and where it does not

Scan-to-CAD is a strong fit for measured floor plans, renovation backgrounds, lease plans, existing-conditions sets, facade documentation, roof plans, sections, and RCPs, and for any team in AutoCAD handing DWG and PDF to consultants, contractors, owners, or permitting. For renovation work especially, the CAD set becomes the existing-conditions base the next team builds on, so nobody redraws the building from scratch. The view-by-view breakdown is in point cloud to CAD services.

It is the wrong tool when you need intelligent geometry: Revit elements, levels and views, schedules, clash detection, or a model owners update over time. That is BIM, and when AutoCAD drawings are enough instead of a BIM model is worth a read first. If the answer is BIM, the cloud often links straight into Revit; our point cloud to Revit workflow covers that path.

A CAD scope does not include these unless you name them: hidden conditions, engineering or code analysis, design corrections, full MEP tracing, construction details, fabrication-ready CAD, unlimited revisions, and BIM modeling. If you need any, say so before pricing.

What to send for an accurate quote

Send the project address or scan files, the target deliverables and drawing types, approximate square footage and floor count, file formats and any CAD standards, the deadline, site-access notes, and an example of the drawing detail you expect. If you already have a cloud, note the format and whether it is registered. The closer your request is to a named drawing list, the closer the quote is to a real number. When you are ready, request a quote and we will scope the scan, registration, and drafting against the drawings you actually need.

FAQ

What is the difference between point cloud to CAD conversion and scan-to-BIM?
Point cloud to CAD conversion produces 2D drawings (DWG and PDF: plans, sections, elevations, RCPs) traced from the registered cloud. Scan-to-BIM produces a 3D model, usually a Revit RVT or an IFC, made of intelligent elements with levels, categories, and schedules. CAD describes the conditions; BIM models them.

What is scan to BIM?
Scan-to-BIM turns registered laser-scan data into a coordinated BIM model. A modeler builds walls, floors, structure, and selected MEP as parametric Revit elements to a defined level of detail, so the geometry carries data and can be coordinated, scheduled, and updated, which flat CAD linework cannot.

What are as-built drawings, and how do they differ from record drawings?
As-built drawings document the actual, measured state of a building as it exists, not the design intent; scan-based as-builts are traced from a point cloud, so walls, openings, and levels reflect what the scanner measured on site. Record drawings are the design set marked up with reported construction changes, so they depend on what was logged. When accuracy matters, the scan-backed as-built is the more reliable source.

How does 3D laser scanning work for a CAD job?
A scanner such as the Trimble X7 sends laser pulses and times their return from up to 500,000 points per second to build a dense, measured point cloud. Multiple positions are registered into one coordinate-consistent dataset, and the drafter cuts plans and sections from that cloud to draw the CAD deliverables.

---

Last reviewed: May 2026.