Automation can produce a draft bill of quantities from BIM models or 2D drawings, provided the source files are structured, measurement rules are agreed in advance, and a quantity surveyor reviews the output before it goes anywhere near a tender. Two routes exist: extraction straight from a BIM model, or AI-assisted vector extraction from PDF drawings. Either way, the practical first move is the same: agree your measurement rules and review workflow before you extract a single quantity.
TL;DR:
- Use native IFC or Revit for model extraction, and DWG or vector PDFs for drawing extraction; converting scanned PDFs first reduces OCR errors.
- Layered walls, floors, and roofs often create mismatches, so send compound elements, ambiguous classifications, and duplicate counts to a quantity surveyor for correction.
- Agree NRM2 categories, level of detail, and file formats with the design team before extraction, then map quantities to BoQ rules.
- Record each quantity’s source view, revision, applied rule, reviewer, and decision, then manually sample area and count results before tender sign off.
Table of Contents
- Technical approaches to automated takeoff: BIM extraction, vector parsing and AI
- Where automation succeeds and where it fails
- Measurement rules and information governance for automated outputs
- Five-step audit-ready workflow to adopt automated quantity takeoff
- Checks, exception management and what an auditable output looks like
- Evidence and practitioner guidance on automation's role
- If you want to try a draft BoQ generator
- FAQ
- Sources
Technical approaches to automated takeoff: BIM extraction, vector parsing and AI
Automated takeoff tends to follow one of two technical paths, and knowing which applies to your project changes how much checking you need to budget for.
BIM extraction pulls quantities directly from semantic objects in an IFC or Revit model. Because each wall, slab or door carries defined properties, software can query the model and return counts, areas and volumes without measuring anything manually. This is the method behind most 5D workflows, where quantities link straight through to cost planning. A 2024 study on automated BIM-based takeoff confirms that this approach can extract and visualise quantities automatically, though it stresses that accuracy depends entirely on how complete and correctly classified the model is.
Vector and PDF extraction works differently. Rather than reading object properties, it reads the geometry drawn on a sheet, whether that's native vector data from a CAD export or shapes recovered from a scanned PDF through OCR and vectorisation. Rules are then mapped onto that geometry to classify elements and generate counts.
Hybrid pipelines increasingly combine the two: raster PDFs get converted to vector geometry first, then measurement rules are applied to the resulting lines and polygons. This reduces the OCR errors that plague direct raster extraction.
When preparing files for any automated route, we'd prioritise:
- Native IFC or Revit exports over flattened PDF renders of a model.
- Vector PDFs or DWG files rather than scanned or raster PDFs.
- Drawings with consistent layer naming and object classification.
- Current revisions only, clearly labelled, to avoid extracting from superseded information.
The method you choose should match what's actually available for the project, not what looks most sophisticated on paper.
Where automation succeeds and where it fails
Model completeness and classification consistency decide whether an automated output is close to right or badly wrong. A wall modelled as a single generic object behaves predictably. A wall built from multiple layers, each meant to be measured and priced separately, is where extraction tools most often lose accuracy.
Comparative research on BIM-based takeoff for compound elements identifies exactly this: compound elements such as walls, floors and roofs are a leading source of mismatch between modelled and constructed quantities, producing either excess or missing quantities unless specifically corrected for.
Compound elements are a common source of mismatch between modelled and constructed quantities, unless the extraction method specifically accounts for them, according to comparative BIM research.
BIM extraction tends to be the stronger route when the model is well developed and consistently classified, particularly on larger projects where a design team maintains the model collaboratively. A 2D-to-AI approach is often more pragmatic for smaller projects, refurbishments or anywhere a usable model simply doesn't exist.
Common failure modes to watch for in extracted schedules:
- Duplicate counts where overlapping objects or linked models double up on the same element.
- Missing quantities for elements excluded from the model but shown only on 2D annotations.
- Misclassified elements that get mapped to the wrong BoQ category.
- Compound build-ups collapsed into a single quantity instead of their constituent layers.
None of these are reasons to avoid automation. They're reasons to build checking into the process from the outset.
Measurement rules and information governance for automated outputs
Extracted counts are not BoQ items until they've been mapped against proper measurement rules. The RICS NRM guidance sets out the detailed rules and templates for preparing bills of quantities, and it remains the reference point for how item descriptions, preliminaries and non-measurable work should be structured, regardless of whether the raw quantities came from a tape measure or a model query.
Before extraction starts, it's worth agreeing level of detail and data exchange requirements with the design team directly, rather than discovering gaps once the model arrives. Our guide on why architects use quantity calculations covers why this early alignment matters for the design side of that conversation.
Traceability rules matter just as much as the measurement rules themselves. For every extracted quantity, it's worth recording:
- The source sheet or view reference it came from.
- The drawing revision ID at the time of extraction.
- Which rule or classification was applied.
- Who reviewed the figure and what decision they made.
Pro Tip: Agree your LoD and file format requirements with the design team in writing before extraction starts, not after the first draft BoQ throws up gaps.
Five-step audit-ready workflow to adopt automated quantity takeoff
A structured process turns automated extraction from a shortcut into something you can defend at tender stage.
- Agree measurement rules and exchange files at project start. Settle on NRM2 categorisation, LoD expectations and which file formats will be delivered before any modelling or drawing work is extracted.
- Prepare and verify source files. Check that IFC, Revit, DWG or vector PDF exports are current, correctly layered and free of superseded revisions.
- Run extraction and map outputs to BoQ structure. Apply measurement rules to the extracted geometry or model data and route the result into your standard BoQ format.
- Review exceptions in a queue. Flag compound elements, ambiguous classifications and anything outside expected tolerances for manual correction.
- Produce a draft BoQ, log the audit trail and sign off. Record what was reviewed, by whom and with what outcome, before the document moves to tendering.
| Step | Main output | Who's responsible |
|---|---|---|
| Agree rules and files | Measurement and exchange brief | QS and design team |
| Prepare source files | Verified IFC/DWG/PDF set | QS or BIM coordinator |
| Run extraction | Raw quantity schedule | Extraction tool |
| Review exceptions | Corrected quantities | QS |
| Sign off draft BoQ | Audit-ready BoQ | QS |
For a more detailed breakdown of each stage, our eight-step audit-ready takeoff workflow expands on checklists and common pitfalls at each point.
Checks, exception management and what an auditable output looks like
Exception queues are what separate a usable draft from a liability. Rather than reviewing every line, configure the system to flag items outside expected tolerances and route them for manual attention, while simple, repeatable counts can move through with lighter checking. The AI4QS report identifies exception management as a reliable automation target: systems should route ambiguous or compound items to a QS queue and only auto-accept straightforward, repeatable counts.
Sampling checks work well alongside exception queues. Reconcile a sample of area and count figures against the source drawing or model manually, rather than trusting the full output on faith.
When inspecting flagged items, pay particular attention to:
- Openings that may or may not have been deducted correctly.
- Finishes schedules, which often sit separately from the structural model.
- The build-up of compound elements such as layered walls or floors.
Pro Tip: Keep a minimum audit trail for every quantity: source sheet or view, object ID, the rule applied, reviewer initials and the decision made. UK guidance on information management for built assets sets out comparable principles for recording revision IDs and reviewer decisions across a project's information.
Evidence and practitioner guidance on automation's role
The evidence consistently points the same way: automation assists quantity surveying work, but its reliability depends on the quality of what goes in. The AI4QS 2025 report frames AI in quantity surveying as an opportunity that requires transparency, accountability and continued human oversight, not a replacement for professional judgement.
In practice, that means treating automation as assistance: the QS defines the measurement rules, sets the exception thresholds and signs off the final BoQ.
- AI tools can speed up extraction and flagging of likely errors.
- They cannot currently substitute for a QS's judgement on ambiguous or compound items.
- Early rule-setting reduces downstream correction work more than any extraction algorithm does.
Our piece on AI for quantity surveyors looks at what current tools can and can't do in more detail.
If you want to try a draft BoQ generator
We've built QuantiFlow around this exact workflow: it reads construction drawings and produces a draft bill of quantities for a quantity surveyor to review and sign off, rather than trying to remove that review step. It maps directly onto the five stages above, from file preparation through extraction, exception review, draft BoQ and audit log.
QuantiFlow is still in development, with Solo, Business and Enterprise plans available to explore if you'd like to see how a draft BoQ fits into your own review process. If your workflow involves printing drawings for site reference or sign-off, large format print suppliers can also handle that side of document handling.
FAQ
Can automated takeoff replace manual measurement entirely?
No. Automated extraction can produce draft quantities from BIM models or drawings, but accuracy depends on model completeness and classification, so a quantity surveyor still needs to review and sign off the output, as the AI4QS report sets out.
What file formats work best for automated extraction?
Native IFC or Revit exports suit BIM-based extraction, while vector PDFs or DWG files work better than scanned or raster PDFs for drawing-based extraction. Converting raster PDFs to vector geometry before measurement reduces OCR-related errors.
Why do compound elements cause problems in automated takeoff?
Compound elements such as layered walls, floors and roofs are a recognised source of mismatch between modelled and constructed quantities, according to comparative research on BIM-based takeoff. They often need manual correction rather than automatic acceptance.
How does NRM2 relate to automated quantity extraction?
Extracted counts still need mapping against proper measurement rules before they become usable BoQ items. The RICS NRM guidance sets out the rules and templates for structuring item descriptions, preliminaries and non-measurable work.
What does QuantiFlow actually do?
QuantiFlow reads construction drawings and produces a draft bill of quantities for a quantity surveyor to review and sign off. Pricing runs from £69 per month for Solo to £299 per month for Business, with Enterprise pricing available on request.
Sources
For readers who want to check the sources behind this article or dig further into the underlying research, the following cover measurement rules, extraction accuracy and AI adoption in quantity surveying.
Recommended
- Audit log best practices for UK quantity surveyors
- Six Steps to an Audit Ready Cut and Fill Takeoff for UK QSs
- Best quantity takeoff software for UK QS firms: 2026 guide
- Audit Ready NRM2 MEP Quantity Takeoff: Eight Steps for UK SME QSs
This article is for general information only and is not professional, legal or commercial advice. Quantity surveying decisions should be taken by a qualified professional with reference to the specific project, drawings and contract in question. Content is produced with AI assistance and reviewed before publication. QuantiFlow Ltd accepts no liability for reliance on it.