How BuildSmart Saved 1,200+ Hours Monthly by Replacing Manual Takeoffs with AI

In commercial construction, the margin between winning and losing a bid is rarely the price. It's the timing.

BuildSmart (name changed) is a 45-person general contractor based in the Pacific Northwest, specializing in multifamily residential projects in the 20–80 unit range. By most measures, they were competitive: experienced estimators, solid subcontractor relationships, and a track record that should have put them in the running for most projects they pursued.

Losing Bids You Should Have Won

But in an eight-month stretch leading into 2024, they lost four consecutive competitive bids — not because their numbers were wrong, but because they submitted late. In each case, a competing GC had submitted a complete, accurate estimate two to three days earlier. Two of the four owners had already moved forward with the early bidder before BuildSmart's submission arrived.

The root cause was straightforward: their estimating process was manual. Two senior estimators, working from digital PDFs, spending 60 to 80 hours per bid manually counting doors, windows, linear feet of framing, plumbing rough-in points, and dozens of other quantities across multi-story drawings. On a complex 60-unit project, a complete takeoff took the better part of two weeks.

"We knew exactly what the problem was," the operations director said. "We just didn't think software could actually solve it yet."

The Decision to Change

After the fourth lost bid, the operations director pulled together a cost analysis. The four projects represented a combined contract value of approximately $28 million. At the firm's typical margin, winning even two of them would have covered the annual cost of almost any estimating software on the market many times over.

The firm began evaluating AI-powered takeoff tools in Q1 2024. After a trial period, they standardized on Togal.AI for quantity extraction, with the plan to layer in scheduling and site monitoring tools once the estimating workflow was stable.

The Tools That Changed the Workflow

AI estimating: Togal.AI for quantity takeoffs

The core change was replacing manual plan counting with AI-driven quantity extraction. Estimators upload digital drawings to Togal.AI, and the system automatically detects and measures building components — doors, windows, rooms, wall lengths, and dozens of other quantities — directly from the PDF plans.

What previously took a senior estimator 60 to 80 hours now produces a complete first-pass takeoff in four to six hours. Estimators spend the remaining time reviewing the AI output, checking flagged items, and applying unit costs — the judgment-intensive part of their job — rather than spending most of their time counting.

The conversational search feature (Togal.CHAT) also changed how the team worked with unfamiliar plan sets. Instead of manually searching through 80-page drawing sets to find where a specific detail was called out, estimators can ask the system directly and get an immediate answer with a reference to the relevant drawing page.

AI scheduling: ALICE Technologies

Once the estimating workflow was running, the firm introduced ALICE Technologies for schedule generation on larger projects. Rather than manually sequencing trades and building a schedule from experience and intuition, ALICE analyzes project constraints — site access, trade dependencies, crew availability — and generates optimized schedule options.

For a 72-unit project completed in Q3 2024, the AI-generated schedule identified a sequencing option that reduced the projected timeline by 23 days compared to the schedule the project manager had drafted manually — by optimizing how concrete pours, framing, and MEP rough-in overlapped across the building's vertical phases.

AI site monitoring: OpenSpace

The third layer was OpenSpace for site progress tracking. Project managers mount a 360° camera to their hard hat during weekly site walks, and OpenSpace automatically maps the captured footage to the project's floor plans and BIM model. The result is a visual, timestamped record of actual site conditions that can be compared directly to the planned schedule.

This proved particularly valuable for subcontractor coordination. Rather than relying on verbal progress reports or scheduled walkthrough meetings, project managers could review site conditions remotely, identify discrepancies between actual and planned progress, and address issues before they affected the critical path.

The Results, Six Months Later

The 1,200 hours per month figure reflects the combined time savings across the estimating team on all active bids. The $350K annual savings estimate comes from three sources: reduced estimating labor cost, increased bid volume leading to more contract awards, and reduced schedule overrun costs on active projects.

The firm submitted eight bids in the first full month after stabilizing the new workflow — more than double their previous monthly volume with the same estimating team. Three were awarded contracts.

What This Means for Construction Firms

The BuildSmart story is primarily about capacity, not cost-cutting. The same two senior estimators, doing the same quality work, can now cover more than twice the bid volume. The AI handles the counting; they handle the judgment.

The scheduling and site monitoring tools compound this: projects are more likely to finish on time, which protects margin and builds the kind of track record that wins more bids in the first place.

The tools don't make us better estimators. They make our estimators available to actually estimate — instead of spending most of their day manually counting things a computer is faster at anyway.

For firms still running manual takeoffs, the math is straightforward. Every hour an estimator spends counting is an hour not spent on the analysis, scope review, and subcontractor coordination that actually requires their expertise.