2026, August 7th – Faster, Smarter Takeoffs Inside Contractor Foreman

August 7th, 2026
Hosted by: Taha, Trainer at Contractor Foreman
Topic: The Takeoff Tool — From Drawings to Quantities to Estimate

Webinar Description

In this advanced training session, Taha walks through the Takeoff tool in Contractor Foreman — currently available in preview — covering how it works, how to verify what the AI produces, and how the output feeds into the estimating workflow. Drawing on a decade as a construction project manager and his background as a licensed mechanical engineer, Taha frames the session around a practical question: where does this tool actually save time, and where does your own judgment still have to do the work?

Before touching the software, Taha sets out three things that determine whether a takeoff will be usable. The first is a mindset point he returns to throughout: the AI expedites the takeoff, it does not take responsibility for it. He uses the analogy of a car that reverse-parks itself — the technology does the work, but your hands stay on the wheel and your eyes stay on the mirrors, and if something goes wrong, it’s your fault, not the car’s. The second is scale verification. Contractor Foreman now reads and sets the drawing scale automatically, but Taha is emphatic that this makes manual verification more important rather than less, because it’s exactly the kind of step people skip once they assume the system has handled it. The third is drawing quality — a back-of-the-napkin sketch will not produce a clean takeoff, and clearly labeled drawings let the AI categorize rooms and objects correctly on the first pass.

The session then moves into the tool itself, found under Financials → Takeoffs. Taha uploads a full drawing set live and shows the AI breaking it out page by page, grouping sheets by discipline — architectural, structural, mechanical — so they can be filtered later, and running a preliminary takeoff across the set as it uploads.

The centerpiece of the session is the Legends tab, which Taha calls the most important part of the tool. This is where everything the AI identified is stored: Spaces (including the building footprint and a Rooms folder broken out by square footage), Objects (doors, toilets, sinks, casement windows), and Walls. His key point is that the Legends tab is the AI’s entire knowledge of your drawing — when you later ask it how much tile the bathrooms need, it can only answer from what is stored here. He demonstrates this by finding a room the AI labeled “unknown” because the drawing didn’t name it, and relabeling it as the laundry room. Any room left unlabeled is effectively invisible to every question you ask afterward.

Taha then covers the manual measurement tools, which sit alongside the AI output rather than replacing it: draw area, draw lines, and the count tool for tallying repeated objects across a page. He gives particular attention to the Search Element tool, which lets you highlight a single item on the drawing — a wall type, a fixture, an outlet — and have the system find every matching instance. He demonstrates it by highlighting a W1 wall tag and returning six matches, and notes it’s the tool he’d reach for first on electrical or mechanical takeoffs, where the tedium of finding every GFCI or HRV buried in a busy sheet is where mistakes happen.

The second half covers the AI chat and how to get useful output from it. Taha starts with a broad question about flooring square footage, then progressively refines it — tile in wet rooms, LVT in dry rooms, carpet in bedrooms — to show how much the specificity of the prompt determines the quality of the answer. He introduces the Prompts Library, where reusable prompts can be saved and applied across drawings, and walks through a stored sheetrock prompt that specifies walls only, the columns to return, and the assumptions to apply. His argument for using saved prompts is consistency: the same prompt across every page means the same assumptions and comparable quantities. His prompting guidance is explicit — state what’s out of scope, supply variables the drawing doesn’t contain (an attendee question about wall heights surfaces that the AI was using an 8-foot assumption Taha had given it earlier in the conversation), and instruct the AI to ask before proceeding when a detail is unclear.

The session closes with the handoff into estimating. Takeoff data exports three ways — Excel, PDF, or a Contractor Foreman–formatted CSV that matches the estimate import template. Taha imports that file into an estimate live, and points out that everything transfers except unit cost and cost code, which remain yours to fill in. From there he covers the further uses: asking the AI for production rates based on crew size and shift structure to build out a scope of work and a preliminary project schedule, with the advantage that this AI has seen your actual drawings rather than working from a general description.

Taha was direct about current limitations throughout. The Search Element tool works one page at a time. Takeoffs don’t carry across drawing pages. Full mechanical, electrical, and concrete takeoffs aren’t supported yet — the tool can supply areas and linear feet, but the trade-specific work is still manual. Externally marked-up drawings can be uploaded and flattened, though clean sheets produce better results.

Attendee questions covered whether the tool handles French-language drawings, whether takeoffs can be combined across pages, where production rates come from and whether custom rates can be supplied, and where in the menu to begin. Attendees were offered a complimentary one-on-one training session, with an offer to review their own drawings directly.

This session is ideal for:

• Estimators who currently do takeoffs by hand or in a separate tool and want to see whether this one fits their workflow
• Anyone using the Takeoff tool who has been asking the AI questions without first checking the Legends tab
• Teams doing repeat takeoffs of the same type who would benefit from saved prompts and consistent assumptions
• Contractors weighing what the tool can genuinely do today against what is still on the roadmap