Your Drywall Sub Spent 48 Hours This Week Counting Walls. The AI Did It From Photos.

By Jake Kowalski, August 30, 2026, Construction Technology
360-degree camera on tripod in framed house with drywall installation progress and AI cost code overlay visualization

Two project engineers at Performance Contracting Group used to spend every Thursday and Friday doing the same thing, walking jobsites with clipboards to count which rooms had framing, which had board hung, and which had tape finished.

That ritual ate two to three days a week for two people, or about 48 hours that never showed up as installed work, only as a spreadsheet that was stale by Monday morning when crews had already moved.

PCG is not a small shop testing a gadget. They are the nation's largest drywall and ceiling contractor, handling about 12,000 projects a year, and in May they signed a multi-year enterprise deal with Track3D to automate that counting with photos and AI.

Since deployment started, Track3D has tracked $413 million in active project value inside PCG's portfolio, monitoring 65 unique drywall and ceiling elements tied directly to PCG's internal cost codes rather than treating drywall as one generic line item.

For a trade that lives on labor productivity and billing milestones, that level of granularity determines when you get paid and whether your crew is making money on the phase that is actually underway.

Why drywall reporting eats your week

General contractors track drywall as a broad category, which makes sense from 30,000 feet where it appears as one bar on the schedule, maybe two if someone split hanging and finishing.

Trade partners cannot work that way because framing, hanging, and taping involve different crews, different labor rates, and different billing triggers that disappear inside a single percent-complete number.

Framers run $35 to $45 an hour, hangers $30 to $38, and finishers $32 to $40 in most markets in 2026. Framing is typically 30 percent of drywall contract value, hanging 40 percent, and finishing 30 percent, so a single number tells you nothing about which phase is actually behind.

PCG's CFO Brett Dahmer said tracking that granular detail tied to specific cost codes is not something legacy platforms were built to handle, and that Track3D built on how PCG already operates instead of asking PCG to fit its work into their model.

The result was automated reporting that gives people time back for field execution and proactive decisions, which is the real pitch here. Time with crews, not more dashboards to check.

38 percent and climbing

Walls and Ceilings magazine reported August 25 that 38 percent of wall and ceiling contractors now see measurable business impact from AI, more than double the 17 percent who said the same a year earlier, which signals adoption moving from experiment to standard tool.

According to the magazine, companies use AI for estimating, documentation, customer service, drywall tracking, scheduling, and jobsite productivity. Performance Contracting Group selecting Track3D for drywall tracking was cited as an example of that shift toward aligning production data with how specialty contractors actually manage work.

Track3D's pitch is straightforward. Their Reality Intelligence Platform integrates AI with reality capture you already collect, usually 360 photos from a camera walk, to track progress and flag delays before they turn into rework.

They already count Hensel Phelps, one of the largest general contractors in the US, as an enterprise customer, and say they are in active discussions with several of the largest mechanical, electrical, and concrete contractors in North America.

PCG co-founder and CEO Chaitanya NK described it as field-first development, saying work with PCG pushed Track3D to track at the specificity their work demands, which shaped how the platform evolved for trade use.

The math nobody published

Two team members spending 2 to 3 days per week on manual reporting equals 32 to 48 hours per week. At a loaded cost of $75,000 to $95,000 for a project or field engineer including benefits, that is $36 to $46 per hour, or about $1,640 per week when you average 40 hours at $41 per hour.

Annualized, $1,640 times 50 weeks is $82,000 per year per project cluster, not per company. For PCG running 12,000 projects, even if only 10 percent get this level of tracking, that is 1,200 projects times 40 hours, or 48,000 hours that could shift from counting to crew support, quality checks, or catching billing early.

More useful for you: 48 hours equals six site walks at 8 hours each, or 12 crew huddles at 4 hours with travel, or three billing cycles you could catch a week early. On a $2.5 million residential project with a 12 percent margin, one early billing cycle is worth about $1,057 in carrying cost at 8.5 percent construction loan rates on a $650,000 average balance.

Break-even is simpler than vendors suggest. If you spend more than 5 hours a week on manual progress reports, you should automate, because a 360 camera costs $400 and tools like OpenSpace, Track3D, or even Polycam with iPhone LiDAR can turn a 20-minute walk into a mapped report tied to floor plans.

Even at 85 percent accuracy you save most of the counting time, and the question becomes whether you trust an 85 percent automated count plus 15 percent spot check more than a 100 percent manual count that is already two days old.

Methodology: labor uses Bureau of Labor Statistics mean wage for construction managers plus 30 percent benefits, PCG project count and $413 million value and 65 elements from GlobeNewswire May 18 2026 release, 48 hours and 2 to 3 days from same release plus CFO quote, 38 percent versus 17 percent from Walls and Ceilings August 25 2026, wage splits from three Northern California trade interviews March 2026, carrying cost uses $650,000 average balance at 8.5 percent equals $151 per day times 7 days.

Where the demo ends and the job site begins

PCG can afford enterprise pricing and you probably cannot, because Track3D publishes no pricing and tools in this class typically run $3,000 to $5,000 per month per project cluster, which is thin return for a custom builder doing 6 to 8 homes a year.

Reality capture still requires someone to walk the site with a camera. If two people previously spent 2 to 3 days compiling reports and now spend one day walking sites and uploading, net time saved is less than gross 48 hours, because vendor math rarely subtracts the new work that replaces the old work.

Sixty-five elements tied to cost codes sounds impressive until you realize PCG had to collaborate to build trade-specific dashboards, meaning Track3D did not have this out of the box and similar customization will be needed for a mid-size contractor with 15 cost codes.

Drywall is the easiest trade for this because it is flat, visible, and color-coded. Mechanical, electrical, and plumbing hide behind walls and in ceilings where cameras cannot see without opening things up, so translation to complex trades remains unproven.

Thirty-eight percent reporting measurable impact is self-reported survey data, not audited financials, and the definition of measurable impact is vague enough to include $100 saved on phone time or $100,000 saved on rework. Walls and Ceilings did not name the underlying report, sample size, or question wording.

Strongest counterargument

The best case against AI progress tracking is not that it fails but that it succeeds just enough to make you trust a dashboard you did not build and cannot audit when something goes wrong.

When PCG's CFO says Track3D did not ask them to fit work into its model, that is enterprise language for months of integration you did not see. Your cost codes are not PCG's cost codes and your crews are not PCG's crews, so accuracy on commercial open floors with good lighting will not match accuracy on a custom home with vaulted ceilings and skylights.

If you stop verifying manually because the dashboard says 80 percent complete, you stop catching the screw pattern that misses code, the corner bead not installed, and the joint taped but not sanded. Those are $200 fixes that become $2,000 callbacks when paint shows every flaw.

A general contractor who tracks drywall as one line already hides which phase is behind, and an AI that tracks 65 elements but gets three wrong still hides which three, so you trade one kind of blindness for another that feels more precise.

Residential needs more judgment, not less, because custom homes vary more than commercial, with same house, different lot, different light, different crew, different day, and software averaging across 12,000 projects might miss the one detail that matters on your project this week.

Limitations

This analysis uses Track3D and PCG claims from a single press release May 18 2026 without independent verification. We did not test Track3D on residential drawings or jobsites and could not verify whether 65-element tracking accuracy is 95 percent or 70 percent for framing versus hanging versus taping.

No third-party audit of the 48-hour claim exists. All performance claims are vendor or chief financial officer quotes, and pricing uses estimates from similar enterprise tools because no list price is published.

Walls and Ceilings 38 percent versus 17 percent is reported without underlying survey methodology, sample size, question wording, or industry report name, so we could not assess whether the sample was representative of residential contractors.

Residential applicability is unproven because all case studies are commercial or multi-family. No single-family custom home case study exists for Track3D or PCG deployment, and small builder adoption of scheduling software is under 20 percent per NAHB 2024, suggesting even basic digital tracking has not penetrated residential.

What to do with this

If you are a general contractor running $2 million to $5 million in residential, stop tracking drywall as one line item and break it into framing, hanging, and finishing, each tied to a cost code and a location like room or floor.

Even a manual spreadsheet with 3 phases times 10 locations gives you 30 data points instead of one, which is 80 percent of Track3D's value without any AI, and it forces your subcontractor to report phase and location instead of vague percent complete.

If you are a specialty contractor doing drywall, paint, or flooring, pilot reality capture with one $400 camera and a trial of OpenSpace or Track3D, then measure time saved versus manual counts for two weeks to see if the math holds in your market.

If you save 10 hours a week at $45 per hour loaded you freed $450 per week or $23,400 per year, which means break-even on camera is less than one week even before you count avoided rework.

Ask your progress tool one question: does it tie to my cost codes or generic categories, because generic means double entry, once for the tool and once for accounting, and that is where the 48 hours goes.

PCG's win was not artificial intelligence but tying installed work directly to specific cost codes automatically so project engineers could support crews instead of counting walls.

For homeowners, ask your general contractor what phase of drywall is done in which room, because framing done means 30 percent, hanging done means 70 percent, and taping done means 100 percent but still needs sanding check before paint, and that question alone saves a week of confusion.

Track3D and PCG proved enterprise can automate trade-level tracking that used to take 48 hours a week, and you can steal the principle without their budget by tracking phases, tying to cost codes, and using photos, not memory.

Sources