Counting window frames, measuring floor areas and calculating wall quantities. For estimators, quantity take-off is a time-consuming but essential part of cost estimating. A small dimensional error or an overlooked item can have a significant effect on the final estimate. AI is increasingly used to support this process. But how reliable is it?
From construction drawings to quantities with AI
Preparing a construction estimate means turning drawings into specific quantities: the number of doors, square metres of wall finish or the amount of insulation required, for example.
Traditionally, an estimator gathers this information by studying drawings, counting items and determining dimensions. This takes considerable time, especially on large projects with dozens of drawings.
AI can support some of this preparation. Depending on the quality of the drawings and the software’s capabilities, AI can:
- Recognise elements: identify doors, window frames, walls and other building components.
- Collect quantities: extract quantities from drawings, dimensions and related schedules.
- Structure the work: turn identified elements into clearly organised estimate items.
- Flag missing information: indicate where dimensions, specifications or other assumptions are unclear.
This reduces manual information gathering and produces an initial overview of the required work sooner.
Practical example: how much wall finish is needed?
Suppose you are preparing an estimate for fitting out an office building.
The floor plans show several types of internal wall. For the estimate, you need to determine the required square metres of wall finish. AI can help identify the different wall types and gather the lengths available in the drawings. But that does not complete the calculation. The drawings may not specify the wall height for every room. It may also be unclear whether door and window openings should be deducted and whether both sides of the walls need finishing. Instead of filling in these gaps itself, an AI system can ask: “No consistent wall height was found for these internal walls. Which height should be used? And should door and window openings be deducted from the finishing area?”
The estimator assesses the situation and supplies the correct assumptions. AI can then help develop the quantities further. Its value therefore lies not just in recognising quantities automatically, but also in identifying the information needed for an accurate estimate.
Which quantities should you check yourself?
Not every construction drawing is equally suitable for automated analysis. A digital drawing with clear dimensions offers different possibilities from a scanned PDF with hard-to-read text and lines.
AI can also misinterpret components, count the same elements twice or miss important details. Estimators should therefore check four things:
- Completeness: have all relevant components, storeys and drawings been included?
- Dimensions: do the dimensions used match the original drawings?
- Assumptions: have wall heights, openings and deduction rules, for example, been applied correctly?
- Traceability: can you identify the drawing or specification a quantity came from?
The last point is especially important. A quantity of 250 m² means little if you cannot see how it was derived. How to verify a quantity in a draft estimate.
From wall dimensions to finish area
- Source
- Record the drawing, revision, sheet and wall type.
- Calculation
- Wall length × finish height × number of sides to finish.
- Deductions
- Record which openings you deduct and the agreed basis.
- Scope
- Record which rooms and construction phases are included or excluded.
Missing a height or deduction rule? Keep it open until you have checked it. A measurable wall is not yet a complete finishing item.
How does Baudex support this process?
Baudex helps construction companies analyse project documentation and prepare estimates. It identifies work and relevant quantity information in the available project documents and helps structure these into a draft estimate. Baudex can take the company’s own estimating rules and work descriptions into account. Where information is missing or unclear, it can flag issues and ask the estimator for further input. The supporting information remains traceable to the project sources used.
This creates a working relationship in which AI supports the preparation and the estimator focuses on assessing and checking the quantities. Less time spent on manual take-off, more attention to a reliable estimate. Want to see how it works? Discover how Baudex supports cost estimating.









































