Innovation | Built | The Bluebeam Blog

Project Management & Efficiency, Technology & Innovation

Why AI Still Can’t Fix Construction’s Most Expensive Problem

Bluebeam’s 2027 AEC Technology Outlook asked 750 U.S. firms what AI is doing for them. On rework, the answer so far: not much.

Frank Kalman
November 12, 2026

Oregon Needs More Building Inspectors. This School Trains Them on Bluebeam.

Health, Safety, and Compliance, People & Project Spotlights
Frank Kalman

Inside the Oregon college program teaching future inspectors Bluebeam from the first term and getting about half of them hired before graduation

Dan Carlson was inspecting the new east grandstand at Oregon State University when he stepped into the contractor’s job trailer to check the plans. On a 60-inch monitor, the drawings were up and live in Bluebeam. His first thought: “Wow, what is that?” His second: “That looks super cool.” That moment in a job trailer now shapes how Chemeketa Community College trains the people who will sign off on Oregon’s buildings for the next generation.

Why is Oregon running short of building inspectors?

Every permit, every final inspection and every rebuild after a disaster runs through a code professional. When that workforce thins out, projects stall at the counter and communities wait. Oregon’s gap is a pipeline problem: retirements are outrunning the handful of programs built to train people for the seat.

The math is blunt. Three programs train code professionals in Oregon — the state Building Codes Division, Chemeketa and Portland Community College — and together they turn out fewer than 50 a year, according to figures Carlson tracks across all three. The Oregon Employment Department projects about 200 openings a year for construction and building inspectors. And the cliff has been visible for a decade: a 2014 survey that the National Institute of Building Sciences conducted for the International Code Council found that 82% of code professionals planned to retire within 15 years, including 31% within five.

It isn’t for lack of a good job. Construction and building inspectors in Oregon earn an average of $85,154 a year, according to Oregon Employment Department wage data. Chemeketa’s Building Inspection Technology program, in Salem, has offered a building inspection degree since 1978 and has produced more than 800 graduates, according to the college. Its cohort is capped at 30 students a year. Last year, 52 people applied.

How did Bluebeam end up in a building inspection classroom?

Code enforcement used to run on a red pen and a roll of drawings. Building departments don’t work that way anymore, and a program that pretends otherwise sends graduates out a step behind. Chemeketa made a deliberate call to teach the tool the job runs on today, not the one it used to.

Carlson knows the program from both sides of the desk. He graduated from it in 1996, spent 14 years as the building official in Corvallis and now serves as building official for the city of Wilsonville. When Chemeketa restarted the program in 2016, he came back to teach. Then the COVID-19 pandemic hit, and passing paper plan sets around a classroom stopped making sense.

“We had to get real about doing electronic plan review in the classroom,” Carlson said. “And Bluebeam was the answer.”

The state was heading the same direction. House Bill 2415, passed in 2021, directed the state to set minimum standards for electronic plan review, permitting and inspections, and required local building departments to meet them starting Jan. 1, 2025. Chemeketa runs on licenses provided through the Bluebeam Academic Program, and Carlson became a Bluebeam Certified Instructor to teach it.

Students take a one-credit Essentials of Bluebeam class in their first term. After that, the software never leaves: every plan review, assignment and group project over the two-year program runs through it, including Studio Sessions where classmates mark up the same set together. Nearly all classes meet live over Zoom three or four evenings a week, which is how the program pulls students from Oregon, Washington, Idaho, Montana, California and, once, Minnesota. The degree costs roughly $20,000, according to the college, with eight ICC certifications built in.

What does electronic plan review change for a building department?

Plan review is where a project’s schedule lives or dies. When every reviewer has to wait for the paper set to come around, the permit counter becomes the bottleneck. Putting all the disciplines on the same drawings at the same time changes who controls the clock.

Carlson estimates electronic plan review runs 30% to 60% more efficiently than paper. He has a favorite proof. In 2019, Wilsonville hosted the NW Natural Street of Dreams, a showcase of six new homes, and the city ran plan review differently.

“Builders have a schedule. They know when they’ll submit plans to the city and how long it should take,” Carlson said. So builders scheduled their submissions in advance, and the city blocked those dates on reviewers’ calendars. On review day, reviewers from several disciplines marked up the same plan sets at once in Bluebeam Studio. All six homes cleared review in a single day. Applicants either had a permit in hand the next day or knew exactly what to fix. The city’s community development staff called it a paperless beta test of the future of plan review, according to a July 2019 city report.

“It’s the type of streamlined permit process I wish we could do with every project,” Carlson said.

Can a career in the trades lead to building inspection?

Some of the sharpest inspectors don’t come from offices. They come off roofs and out of crawl spaces, carrying years of knowing how buildings go together. For a tradesperson whose body can’t keep up with the work, inspection isn’t a step away from construction. It’s a way to stay in it.

Heidi Kameroff spent about 20 years in the trades as a journeyman roofer before an injury took her off the roof for good. “I didn’t want to leave an industry that I had worked so hard to become a journeyman in,” she said. “But physically, there was no way for me to continue to do what I did. And you can’t really be a remote roofer.”

Vocational rehabilitation wanted to steer her into office jobs. She found Chemeketa online instead, and by the end of the program she “was suddenly passionate about something again.” Today she is a building inspector and plans examiner for Lincoln County on the Oregon coast, where she inspects homes being rebuilt after the 2020 wildfire that tore through the Otis area. “Nobody ever thought that the coast would catch on fire,” she said.

“Bluebeam didn’t become essential in my life until I became a code professional,” Kameroff said. “Now I open it every single day.”

Heaven Doherty took a different road in. She spent years with a general contractor and in local government finance before moving into a building department, and she completed Chemeketa’s program while working full time for the city of Pendleton, the Hermiston Herald reported. Now she inspects and reviews plans for the city of Hermiston in eastern Oregon, and the Oregon Building Officials Association named her its 2026 Building Inspector of the Year. “It’s always a privilege to serve your community and to make things safer and better for people,” she said. She credits Chemeketa for the technical depth behind that work — without the program, she said, “I couldn’t do my job at the level I do it.”

Why do building departments hire Chemeketa graduates?

A new hire who needs a year of handholding is a cost a short-staffed department can’t absorb. What employers are really buying from a program like this is time: inspectors who can carry a caseload from the first week instead of learning the code book and software on the clock.

About half of Chemeketa’s building inspection students are hired before they graduate. John Rodriguez, Lincoln County’s building official, hired Kameroff and saw the difference right away. “I didn’t have to do a whole lot of extensive training for plan reviews with my hire — she was very capable from the start,” he said. “You know you’re hiring a capable and well-versed person for your staff in codes and plan review.”

Carlson hears the same thing from the other side of the hiring table. “They’re getting someone who’s been exceptionally well prepared to leap right into the deep end,” he said. “They’re not having to learn Bluebeam on the job.”

The bottom line for the trades

Oregon isn’t waiting on the shortage to fix itself. Senate Bill 1086, passed in 2025, directs the Oregon Department of Consumer and Business Services to help the Oregon Building Officials Association develop a proposal for a building inspector apprenticeship program. The association has identified Chemeketa as its intended education provider and is targeting a 2027 launch. A night-and-online model that turns tradespeople and career changers into job-ready code professionals isn’t a uniquely Oregon answer either. Every state staring down the same retirement wave could use one.

For the people coming out of it, the job is bigger than a checklist. “We’re advocates, not just inspectors,” Kameroff said. “We’re advocates for safer buildings, safer structures.”

Read the full story of how Chemeketa built Bluebeam into its building inspection program in the complete customer case study on bluebeam.com.

Frequently asked questions

Do building inspectors use Bluebeam?

Yes. Building departments use Bluebeam for electronic plan review, drawing markup and real-time collaboration in Bluebeam Studio Sessions, where reviewers from multiple disciplines can mark up the same plan set at once. At Chemeketa Community College, building inspection students learn it in their first term.

How many building inspectors does Oregon need?

The Oregon Employment Department projects about 200 openings a year for construction and building inspectors. Oregon’s three training providers — the state Building Codes Division, Chemeketa Community College and Portland Community College — together turn out fewer than 50 new code professionals a year, according to figures compiled by Chemeketa instructor Dan Carlson.

Is electronic plan review required in Oregon?

Yes. Under House Bill 2415, passed in 2021, local building departments in Oregon must meet state minimum standards for electronic plan review, permitting and inspections, starting Jan. 1, 2025.

What is Chemeketa’s building inspection program?

Chemeketa Community College in Salem, Oregon, has offered a building inspection degree since 1978. Its Building Inspection Technology program is a two-year degree that costs roughly $20,000, according to the college, builds in eight International Code Council certifications, holds nearly all classes live online in the evenings and caps enrollment at 30 students a year.

See how Bluebeam handles plan review on your next job.

Start Free Trial

A Takeoff That Adds Up Can Still Be Wrong

Project Management & Efficiency, Technology & Innovation
Frank Kalman

The checks estimators run between measurement and pricing to catch scale errors, missed scope and double counts before a bid goes out

The takeoff is finished the afternoon before bid day. Every total ties out. The summary sheet is clean and the quantities are ready to go to pricing.

Then a second estimator turns on the takeoff layers and pans across the third floor. The east wing is blank. No color, no markups, no quantities. The corridor walls on that end of the building were never measured, and the total never noticed, because a total can only add up what’s in it.

Misses like that hurt because there’s so little room to absorb them. Direct rework costs average about 5% of total construction cost, according to a study of 359 projects published in the American Society of Civil Engineers Journal of Construction Engineering and Management. The average contractor kept a pretax net margin of 6.8% in fiscal 2025, according to the Construction Financial Management Association’s annual financial benchmarking survey. The two figures come from different studies and don’t measure the same thing. Read side by side, though, they make the point: an ordinary rework hit can eat most of a job’s profit, and a quantity error that rides into a winning bid only adds to the pressure.

Catching that error is the job of validation: the pass between measurement and pricing where estimators stop building the takeoff and start testing it. They check whether the scale is right, every area is covered once, the numbers make sense for this kind of building and someone else can trace each quantity back to the sheet. Construction Takeoffs: The Complete Guide lays out the full workflow. This is a closer look at the validation step and the checks experienced estimators run before a number goes anywhere near a price.

What does it mean to validate takeoff data?

Measuring produces numbers. Validation decides whether those numbers have earned a price. It’s a separate pass with a different mindset: the estimator stops building the takeoff and starts trying to break it, because a mistake caught at the desk is a correction and the same mistake caught in the field is a cost.

In a digital takeoff, every measurement lives in two places at once: as a markup on the drawing and as a row in the Markups List. That double record is what makes validation possible. The drawing shows where a quantity came from. The data shows what it adds up to. Validation works both sides and checks that they agree.

The cost-engineering profession draws the same line. The Association for the Advancement of Cost Engineering (AACE) International’s Recommended Practice No. 31R-03, “Reviewing, Validating, and Documenting the Estimate,” separates review, which checks whether an estimate covers the entire project scope and is free of errors and omissions, from validation, which benchmarks the numbers against outside metrics. It’s explicit about who has to do it: “Validation should be done even if the review team also prepared the estimate.” For a takeoff, that’s the difference between asking whether everything was measured and whether the measurements make sense.

Timing matters as much as method. The takeoffs guide places validation before pricing, “when changes are still cheap.” Once unit costs, labor and markup are layered onto the estimate, a quantity error is harder to see and more expensive to unwind. It’s also why estimators keep the takeoff to net quantities and apply waste downstream. A quantity takeoff, a material takeoff and a cost estimate are different jobs, and a number with waste baked into it can’t be checked against the drawing.

Is the drawing scale right on every sheet?

Scale errors are the most dangerous kind because they’re consistent. Every quantity on a mis-scaled sheet is wrong by the same ratio, so nothing stands out against anything else. The takeoff looks internally sound and is off from the first click, which is why calibration is the first thing validation reopens.

Calibration happens before measuring, but validation confirms it held. PDFs get resized, scanned and re-exported as they move between offices, and the printed scale on a sheet proves nothing on its own. The check is simple. After the takeoff, measure one known dimension on each sheet, such as a gridline spacing or a dimension string, and compare it with the number on the drawing. If a bay that’s supposed to read 30 feet comes back at 28, every quantity on that sheet carries the same error.

Give extra attention to sheets that arrived late, came from a different consultant or carry details at a different scale from the plan. Those are the places where a calibration that was right for the rest of the set lightly stops applying.

Did the takeoff cover all of the scope, and only once?

A total can’t tell you what’s missing. Only the drawing can. Coverage review swaps the question from whether the numbers add up to whether they match the building, and that’s where skipped rooms, forgotten floors and doubled counts finally show themselves.

This is the fastest check in the process. Turn the takeoff layers on, give each scope its own color and pan the full set. Measured areas carry color. Unmeasured areas stay blank. When a room, wing or floor has no markup on it, the takeoffs guide’s rule of thumb applies: “it probably wasn’t measured.” Two markups stacked on the same fixture or wall run point the other way, to a double count.

Automated counts need the same scrutiny. VisualSearch can find every instance of a symbol across a drawing set in minutes, but it can’t tell a light fixture on the plan from the same symbol in the legend, a detail or a fixture schedule. Filter the Markups List by page and subject, and check whether any counts landed on a legend sheet or a schedule. Treat automated results as a draft until someone confirms them.

That caution applies to newer AI tools, too, and more contractors are using them. In a late 2025 survey by the Associated General Contractors of America and Sage, 23% of firms reported using AI for estimating. The technology is good but not finished. A 2025 peer-reviewed study in the International Journal on Document Analysis and Recognition by Laura Jamieson, Carlos Francisco Moreno-García and Eyad Elyan reported a mean average precision of 79% for one AI method of detecting symbols on construction drawings and 83% for another. Accuracy in that range means some symbols get missed and some get counted that shouldn’t be, which is exactly what a coverage review is built to catch.

Then look up and down. Vertical scope is where coverage reviews miss the most: conduit drops, wall heights above the ceiling, sloped roofs and anything else that lives in elevation rather than plan. A plan view can look fully covered while material sits in the dimension the plan doesn’t show.

Some scope gaps start in the drawings, not the takeoff. Smart Review, currently in preview in Bluebeam Max, scans construction documents for design issues, scope gaps and discrepancies, which gives estimators a head start on the questions worth raising before quantities are locked.

Do the quantities make sense for this kind of building?

A takeoff can be measured perfectly and still be wrong, because a system was misread or a scope was missed on the drawings themselves. Order-of-magnitude checks test the takeoff against what experience says a building like this should need. They don’t prove a number is right. They catch the numbers that can’t be.

Experienced reviewers don’t start by measuring again. They start with ratios. In Bluebeam’s Modern Quantity Takeoff Playbook, IBEW journeyman electrician Chris Paup describes common sanity checks: weight per square foot for steel, volume-to-area ratios for concrete and conduit footage relative to device counts in electrical work. Reviewers often spot-check a single floor, bay or system, work out what it should produce and compare that with the full takeoff. Numbers that feel too clean, or land far outside historical norms, are a signal to go back to the drawings before pricing.

Published reference points help, as long as they’re used for what they are. The American Institute of Steel Construction’s conventional framing studies put a Minneapolis office prototype at 7.55 pounds of structural steel per square foot and a five-story Denver parking structure at 11.85 pounds. The Florida Department of Transportation tells bridge designers that, in the absence of better information, they can assume 110 to 225 pounds of reinforcing steel per cubic yard of concrete, depending on the element.

Check	Reference value	Context	Source
Structural steel, office	7.55 lb. per sq. ft.	AISC office prototype, Minneapolis; 728 tons over 192,730 sq. ft.	American Institute of Steel Construction, 2018
Structural steel, parking structure	11.85 lb. per sq. ft. (9.55 with castellated beams)	Five-story prototype, Denver; 40 psf garage live load	American Institute of Steel Construction, 2019
Rebar, bridge deck slabs	205 lb. per cu. yd. (standard); 125 (isotropic)	Design aid “in the absence of better information”	Florida Department of Transportation, 2002
Rebar, bridge piers	150 to 215 lb. per cu. yd.	Varies by pier height and column count	Florida Department of Transportation, 2002
Rebar, concrete box girders	165 to 225 lb. per cu. yd.	Typical segment vs. pier segment	Florida Department of Transportation, 2002
Design vs. independent check	Within 5%	Preliminary bridge quantities only	Colorado Department of Transportation, 2026

These are single-project reference points and design aids, not industry averages. Steel tonnage climbs with building height, bay size, seismic and wind demands and crane loads. Use them to catch order-of-magnitude errors, such as a missed floor or a bad scale, not to price work.

Even the agencies that publish ratios warn against leaning on them. The Colorado Department of Transportation’s Bridge Design Manual allows reinforcing estimates by weight per cubic yard at the preliminary stage, expects design and check quantities to land within 5% of each other and is blunt about the final numbers: “Do not use preliminary quantities based on volume (#/cy), area (lb/sf), or percentages for final quantities.” Benchmarks catch gross errors. Measurement still has to carry the bid.

For most trades, the best benchmarks aren’t published at all. Free, primary sources for ratios such as drywall per square foot of floor or conduit footage per device are hard to find, and many of the figures circulating online trace back to vendor blogs. That’s an argument for building your own. AACE ranks a company’s historical data from similar completed projects as an acceptable basis for validation, behind third-party published data. A firm that records quantity ratios from its closed-out jobs owns a check no one can sell it.

Custom columns in the Markups List let those checks live inside the takeoff instead of on a scratch pad. A formula column that converts a slab’s volume back into its implied thickness, for example, puts a ratio check on every row. A 6-inch slab works out to about 1.85 cubic yards per 100 square feet. If a slab subject implies 9 inches, either the drawing changed or the depth input did, and both are worth knowing before the bid goes out.

Check the takeoff against the specifications, too, not just the drawings. When the two disagree, the spec usually governs, as the seven-step takeoff workflow points out, and a quantity that matches the plan but not the spec hasn’t been validated.

Can another estimator trace every quantity back to the sheet?

Validation that lives in one person’s head isn’t validation. A reviewer can only challenge a number if they can see how it was built, and a number nobody can challenge hasn’t really been checked. Reviewability is what turns one estimator’s judgment into a number the whole team can stand behind.

Peer review starts with structure, not arithmetic. According to Paup, reviewers first check whether the takeoff is readable as a system: whether the same tools and naming conventions run across every sheet, quantities can be isolated by floor, zone or system and alternates and revisions are clearly separated from base scope. When that structure is missing, even accurate numbers are hard to trust because no one can tell what belongs where.

That’s the payoff of setting up a takeoff with review in mind from the start. If estimators begin “with the end in mind,” Paup said, “it saves a lot of redundant work. It makes the takeoff easier to review and easier to trust.”

Public owners write traceability into policy. The US Army Corps of Engineers’ technical review guidance for cost engineering lists it as a checklist item: “Major quantities supported by a quantity take-off document.” A quantity that can’t be traced to a markup on a sheet fails that test, however right it happens to be.

Status tracking makes the review visible. Moving items from in progress to reviewed to verified in the Markups List shows at a glance what has been checked and what hasn’t, and it keeps a junior estimator’s work from reaching pricing without a second look. That matters more as the workforce turns over. The Bureau of Labor Statistics projects about 17,600 openings a year for cost estimators through 2035, all of them to replace people who retire or leave the occupation, even as total employment slips 3%. A repeatable review is how a newer estimator’s takeoff gets the same scrutiny as a veteran’s.

Log open assumptions alongside the quantities, too. A clarification that never got asked is still a hole in the bid, even if every number around it is right.

On a large set, the Markups List itself can become a lot to read. Bluebeam Max connects Revu to AI through MCP, launching with Anthropic’s Claude, so estimators can run prompts that search project information and turn markup data into insights. The AI speeds up the search. The estimator still makes the call.

What happens to validated data when the drawings change?

Validation isn’t a stamp. It’s a state, and every addendum puts part of it back in question. Teams with structured, traceable takeoffs revalidate the slice that moved. Teams without that structure end up choosing between checking everything again and checking nothing, and neither is a good answer on a deadline.

Drawings change mid-bid on most projects. The validation job then is narrow: confirm the set is current, compare old and new sheets, measure again only what changed and rerun the checks on those quantities. Keeping the original quantities for comparison shows what changed and by how much, which is often the first question a project manager asks. Why most takeoffs fall apart when drawings change covers that discipline in depth.

Comparison tools speed up the first part. Smart Overlay in Bluebeam Max compares two versions of a drawing set and detects design changes across disciplines and drawing scales so reviewers can go straight to the sheets that changed most. Keeping quantities linked from the measurement to the estimate through features like Quantity Link means a revalidated number flows downstream without being retyped. Retyping is where a validated number can lightly stop being one.

Validated data is what makes a takeoff defensible

The blank wing at the top of this post doesn’t require a careless estimator. It only requires a total nobody tested against the drawing. Validation is the habit that closes that gap: confirm the scale, check the coverage, test the magnitudes, make every quantity traceable and repeat the checks when the drawings move.

None of it is slow once it’s built into the workflow, and all of it is cheaper than finding the blank wing after the contract is signed. That’s why takeoff errors matter more than most teams think, and it’s the thinking behind Bluebeam’s takeoff and estimation workflows.

For the full process, from drawing intake to revision management, read Construction Takeoffs: The Complete Guide to Quantity & Material Takeoffs in 2026.

Frequently asked questions

What does it mean to validate a construction takeoff?

Validating a takeoff means checking that its quantities are correctly scaled, complete, counted once, reasonable for the type of building and traceable to the drawings before they’re priced. In a digital takeoff, estimators check both the markups on the drawings and the quantity data behind them.

What’s the difference between reviewing and validating an estimate?

AACE International’s Recommended Practice No. 31R-03 treats them as separate steps. Review checks whether an estimate covers the full scope and is free of errors and omissions. Validation benchmarks the numbers against outside metrics, such as published data or a firm’s own historical projects, and AACE says it should be done even if the review team also prepared the estimate.

When should takeoff validation happen?

Before pricing. Once unit costs, labor and markup are applied, quantity errors are harder to spot and more expensive to fix. Validation should also be repeated for any quantities affected by addenda or drawing revisions.

How do you find missing scope in a takeoff?

Visual review is the fastest method. Turn on the takeoff layers, give each scope its own color and pan the drawing set. Areas with no color usually weren’t measured. Check elevations and sections as well, since vertical scope such as conduit drops and wall heights doesn’t show in plan view.

How do you catch double counting in a takeoff?

Look for overlapping markups on the drawing, and use consistent subjects and naming so the same scope can’t be measured under two names. For automated counts, filter the results by page to make sure symbols in legends, details or schedules weren’t counted as installed items.

Are there published benchmarks for checking takeoff quantities?

A few. The American Institute of Steel Construction publishes steel tonnage for prototype buildings, and some state transportation departments publish reinforcing steel per cubic yard of concrete for bridge elements. Most trade ratios, such as drywall or conduit per square foot, aren’t publicly available, so many firms build their own from closed-out jobs.

Can AI validate a takeoff?

AI tools can speed up counting, flag scope gaps in drawings and detect changes between drawing versions, but published research shows automated symbol detection still misses some items and miscounts others, and software doesn’t understand design intent. Automated results should be treated as a draft and confirmed by an estimator who remains responsible for the final quantities.

Put your next takeoff through the same checks.

Start Free Trial

The Trades Are Going Digital. This Union School Saw It Coming.

People & Project Spotlights, Technology & Innovation
Frank Kalman

Inside the Silicon Valley union school teaching electricians to run data center work in Bluebeam before they hit the job

A journeyman used to show up, grab paper prints off the table and go to work. That day is ending. Walk onto a data center job today and the foreman hands you a laptop, not a roll of drawings, and expects you to already know your way around software like Bluebeam. At the Electrical Training Alliance of Silicon Valley (ETASV), the union electricians building the Bay Area’s data centers learn it before they ever hit the site.

Why is a union electrician learning software at all?

The old line between “field” and “office” is collapsing. As data center work floods Silicon Valley, the electricians who move up aren’t just the ones who can bend pipe. They’re the ones who can read, mark up and manage a drawing set on a screen without missing a beat.

“Gone are the days of just showing up and working with your hands,” said Chris Paup, a journeyman electrician, ETASV instructor and business representative for IBEW Local 332. “Technology has implemented itself in this industry in so many different ways that you really have to take control of your own education.”

Paup would know. He built ETASV’s Bluebeam Fundamentals course from the ground up. And the reason the school leaned in isn’t a hunch about the future. It’s the work already landing on its doorstep.

How much of a data center is actually electrical work?

This is the number that reframes the whole trade. Data centers aren’t buildings with some wiring in them. They are, functionally, electrical machines the size of warehouses, and the crews who energize them carry the bulk of the job on their backs.

The IBEW estimates 45% to 70% of a typical data center’s construction cost goes to the electrical subcontractor. That’s not a rounding error, either; that’s the job. What’s more, the demand behind it isn’t slowing: the U.S. is looking at roughly 81,000 electrician openings a year over the next decade, even as older workers retire faster than new ones sign up.

ETASV Training Director Albert Lancaster doesn’t talk about AI in terms of chatbots. He talks about it in terms of copper and steel. “When I think of AI, I don’t typically think of the user interface,” he said. “What I think of is the backbone behind it, and the infrastructure needed to power the AI. That’s where our training center comes into play.”

How do you teach Bluebeam to someone who’s never touched it?

The hard part isn’t the software. It’s meeting a field worker where they are, then showing them the tool does far more than mark up a PDF. ETASV splits that job across two instructors and two courses, so nobody gets thrown in the deep end.

Paup’s Fundamentals course assumes zero knowledge. Step one is literally what the icon looks like and how to pin it to your taskbar. From there, students learn to navigate the interface, draw, assign data to their markups and build the custom columns that turn Bluebeam from a viewer into a calculator.

Then they walk into Chris Doyle’s advanced class. Doyle, a longtime journeyman and a detail manager who works an active data center project in Reno, Nevada, hands students a profile packed with a giant tool set and roughly 100 custom columns he’s built over years, then teaches them to design real projects with it. Paup builds the engine. Doyle teaches you to drive.

Doyle pulls real problems straight off the job and into the classroom. One first-day exercise hands students an actual EV charger plan with a genuine code error buried in it, a ground that wasn’t upsized, and tells them to find it. “Feel free to cheat and share as much as you want,” he tells them, “because that’s what you’re going to do in the field.”

What does a Bluebeam class earn a working electrician?

Training only sticks when it counts toward something real. At ETASV, the payoff is built into the trade’s own rules: the course doesn’t just make you sharper, it keeps your license current.

The Fundamentals course runs 54 hours of classroom instruction and application, well beyond the continuing education a journeyman needs to renew their California license every three years. “The purpose of continuing education is so you don’t get stale,” Paup said. “This helps them stay on the forefront.”

There’s a bigger prize on the horizon, too. When a contractor calls the union hall, they can ask for workers certified in confined space or high voltage. For Bluebeam, the best they’ve been able to ask for is “experience preferred,” because there was never a way to prove the skill. ETASV wants to change that: build out the full course series, add a real exam and issue a certificate a contractor can actually request by name.

Is construction software really built for the trades, or just the office?

This is the chip on the shoulder of every field worker who’s been told a tool wasn’t made for them. The people teaching it at ETASV reject the premise outright, and they’ve got the workflows to prove it.

“The concept of ‘I’ll never touch a computer as a journeyman in the field’ is going to be over pretty soon,” Paup said. “In the next five to six years, instead of getting handed a printout, paper’s going to be obsolete. You’re going to get an iPad on the job. You need to know how to navigate these things to stay relevant, to stay future-proofed in your own career.”

Doyle puts the stakes in numbers a contractor feels. “If there’s any mistakes in a UPS system, you’re talking about millions of dollars in downtime,” he said. “The bar is very, very high, not just because we as an institution want it to be, but because that’s what the industry demands.”

The bottom line for the trades

ETASV is the largest electrical apprenticeship in Northern California, and it’s not chasing the technology wave, it’s teaching ahead of it. The workers coming out of that building in Milpitas are the ones who’ll wire the infrastructure the AI era runs on. And thanks to two instructors, one who built a course on a military deployment and one who’s seen every boom and bust the trade’s had, they’ll walk onto the job already knowing the tools.

“We have this opportunity to really mold people for jobs that don’t exist yet,” Doyle said. “And we’re figuring that out as we go.”

Read the full story of how ETASV built Bluebeam into its curriculum in the complete customer case study on bluebeam.com.

Frequently asked questions

Do electricians use Bluebeam?

Yes. Electricians and MEP trades use Bluebeam for quantity takeoffs, material planning, drawing markup and field-to-office coordination. At ETASV, union apprentices and journeymen learn it as part of their formal curriculum.

What share of a data center is electrical work?

The IBEW estimates 45% to 70% of a typical data center’s construction cost goes to the electrical subcontractor, making electricians the core trade on these projects.

Does a Bluebeam course count toward electrician continuing education?

At ETASV, the 54-hour Bluebeam Fundamentals course counts toward the continuing education hours a journeyman needs to renew a California electrician license, which is renewed every three years.

What is ETASV?

The Electrical Training Alliance of Silicon Valley is an electrical apprenticeship in Milpitas, California, jointly funded by IBEW Local 332 and the NECA Santa Clara Valley chapter. It is the largest electrical apprenticeship in Northern California.

See what Bluebeam can do on your next job.

Start Free Trial

How to Walk into Your CFO’s Office and Make the Case for Changing How Your Team Handles Drawings

Construction & The Economy, Project Management & Efficiency
Frank Kalman

You already know the problem. Here’s how to explain it to the person who controls the budget.

The super finds the drawing eventually — version 7, when he needed version 9, buried somewhere in a shared drive that nobody has organized since the last PM left. Forty minutes gone, plus the calls that followed, plus the rework that came after. None of it shows up on the job cost report.

That’s the thing about document management failures: they don’t announce themselves as failures. They show up, instead, as a bad morning, a slow week, a job that just never quite found its rhythm, and by the time anyone connects the dots back to a folder full of outdated PDFs, the money is already spent and the schedule is already gone.

You know this, and so does your PM, but the question is whether the person who controls the budget knows it — and whether you can walk into that office and make them see it the way you do. Most people can’t, not because they’re wrong about the problem, but because they bring the wrong argument. They bring the frustration instead of the math, describe the workflow instead of the cost, and hand a CFO a complaint when what they needed was a business case.

This piece is the business case. It’s for the PM, VDC lead or superintendent who is done getting the sympathetic nod and needs to close the room.

If you want the full cost picture first, read The Hidden Cost of Document Chaos. What follows is what you do with it.

Lead with a Number, Not a Complaint

A CFO who hasn’t stood in a trailer hunting for a drawing revision isn’t going to feel the cost of that problem intuitively, which means describing it won’t get you very far. You have to quantify it, and you have to make the number specific to your firm.

Start with the industry baseline. Construction professionals spend more than 14 hours a week on what one major study calls “non-optimal activities” — hunting for project data, resolving conflicts, dealing with rework rather than building or managing — and across the US workforce, that wasted time runs an estimated $177.5 billion a year in labor cost.

Here’s how to translate that into something that lands in your office. According to BLS May 2025 wage data, construction managers earn a mean $59.79 per hour, while first-line supervisors — the closest SOC match to superintendents — earn $41.56. Add benefits, which BLS benefits data puts at about 31% of total compensation, and the fully burdened rate runs roughly $86 an hour for a construction manager and $60 for a super.

Fourteen hours a week at those rates is $1,200 a week, per manager, in labor cost that isn’t producing anything. Multiply that by the number of PMs and supers on your active jobs, and the number stops being an industry statistic — it becomes a line on your income statement that nobody has labeled yet.

“$1,200 a week, per construction manager, in labor cost that isn’t producing anything.”

Rework compounds it further. The Construction Industry Institute puts direct rework at about 5% of total project cost, and nearly half of that traces back to poor communication and bad project data rather than bad workmanship. On a $30 million job that’s $1.5 million, and on a portfolio of $30 million jobs, the compounding is the kind of number that tends to get a CFO’s attention.

RFIs stack on top of that. A Navigant Construction Forum analysis of more than 1.1 million RFIs across 1,362 projects found the average large project generates 796 of them, each costing about $1,080 to process — roughly $860,000 per project in RFI overhead before anyone swings a hammer. The median response time was 9.7 days, and about 22% of RFIs never got a response at all, a rate Navigant flagged as a leading predictor of disputes and claims.

What the current workflow costs, by the numbers

Cost driver	Benchmark	Source
Time lost to non-optimal activities	More than 14 hours per person, per week	FMI and PlanGrid
Labor cost of that lost time, U.S. workforce	$177.5 billion a year	FMI and PlanGrid
Fully burdened hourly rate, construction manager	About $86	BLS wage and benefits data
Fully burdened hourly rate, first-line supervisor	About $60	BLS wage and benefits data
Weekly cost of 14 lost hours, per construction manager	About $1,200	Calculated from the rates above
Direct rework	About 5% of total project cost	Construction Industry Institute
RFIs on an average large project	796	Navigant Construction Forum
Cost to process one RFI	About $1,080	Navigant Construction Forum
RFI overhead per project	About $860,000	Navigant Construction Forum
Median RFI response time	9.7 days	Navigant Construction Forum
RFIs that never get a response	About 22%	Navigant Construction Forum

You don’t need all of those numbers. You need one that maps to something your CFO is already accountable for, with a firm-specific dollar figure attached to it. That specificity is what separates a business case from a grievance.

How do you translate a document management problem into language a CFO responds to?

CFOs don’t evaluate technology — they evaluate exposure. The field and the finance office are looking at the same failure through different instruments: one sees a lost morning, the other sees nothing, because no report captures it. A request earns budget when it makes that invisible cost show up where finance already looks.

CFOs think in three categories — margin, risk and labor cost — and the document management problem has a real story in all three. Your job is to tell it in their language rather than yours.

On margin, the numbers are more uncomfortable than most people realize. The CFMA 2025 Financial Benchmarker, which analyzes data from more than 1,500 construction companies, puts pretax net income for commercial GCs and CMs at 4.4% — not the 5–6% that gets thrown around in general industry discussions. At 4.4%, a rework event running 5% of contract value doesn’t just eat into profit; it erases the job’s margin entirely and then keeps going. That’s the reframe worth landing with a CFO: document-driven rework isn’t an operational nuisance; it’s a margin event that can flip a job from profitable to a loss.

On risk, the exposure is quieter but just as real. Every unanswered RFI and every version conflict is a potential claim waiting for a trigger, and the audit trail that doesn’t exist today is the one you’ll need in six months when a sub says they built off what they were given. Courts and arbitrators don’t care that the right drawing existed somewhere in a shared drive; they care whether you can prove who had what version and when. The cost of not being able to prove it often dwarfs the cost of the original mistake.

On labor, the math is straightforward and uncomfortable in equal measure. Construction wages rose 4.2% year over year as of August 2025, the most recent year for which data are available, and the industry needed to attract 349,000 net new workers in 2026 on top of normal turnover, per ABC. Meanwhile, 92% of construction hiring firms report difficulty filling positions, and 45% name worker shortages as the leading cause of project delays, according to AGC’s 2025 Workforce Survey. Firms are paying more for experienced people than they ever have at exactly the moment those people are hardest to find, which makes sending them into the week with 14 hours of information-hunting baked into their schedule not just inefficient, but the most expensive possible use of the industry’s scarcest resource.

What’s the most effective way to structure a construction technology business case?

A business case gets funded in order: the problem, its price, proof that a peer solved it, then the ask. The order is the argument. A CFO who hears the solution before feeling the cost has nothing to weigh it against, and a request with nothing to weigh against it gets deferred.

Every CFO in this industry has the same four questions before they approve anything, and the goal isn’t to surprise them with the problem — it’s to make it impossible for them to set it aside.

Start with something specific rather than a category. Not “our document management isn’t working,” but the version conflict on the Riverside Avenue project that cost two weeks of rebar schedule, or the RFI that sat 11 days and triggered a change order nobody budgeted for, or the closeout package that took six weeks to assemble because nothing had been documented in one place throughout the job. A real thing that happened, with a real cost attached to it. Specificity does what statistics can’t — it makes the problem real to someone who wasn’t there.

Then put a number on it. Use the industry benchmarks as a floor — 5% rework, $860,000 in RFI overhead per project — and build toward your own. Firm-specific data always leads, but if you don’t have it, the benchmarks still establish what the category typically costs, and your firm’s experience is unlikely to be better than average.

Then show them a firm that fixed it, because this is where the room usually turns.

DPR Construction was staring down a 31-day schedule delay on the Alta Bates Summit Medical Center project in Oakland — a $500 million hospital expansion where every day of delay carries a very specific dollar figure. Rather than continuing to pass PDFs back and forth between the engineer of record and the rebar detailer, the team ran real-time collaborative reviews using Bluebeam Studio Sessions, which cut the submittal review cycle by more than 33% and compressed a 25- to 30-workday process into 10 to 12 days. Up to 10 days of schedule recovery, per review cycle.
Balfour Beatty put a dollar figure on the same principle at its DFW Airport renovation: $5 million in project savings from centralizing documents and giving field teams real-time access to current information. The savings didn’t come from better technology in any abstract sense; they came from fixing the moment where the wrong information reached the wrong person.
Premier Mechanical saved $70,000 digitizing project delivery, and Turner Construction cut paper costs 73% making the same shift. Rudolph and Sletten, a commercial GC running projects north of $500 million, cut preconstruction review times 30% by standardizing on Bluebeam Studio.

Different firm types, different project sizes, same pattern: the savings don’t come from the tool itself; rather, they come from the workflow the tool makes possible.

Close with a specific ask rather than a category. Not “better document management,” but a defined scope: standardized drawing workflows across active projects, with version control and a single shared set accessible to field teams in real time. That’s something that can get a line item. A category cannot.

What objections should you expect when making the case for construction document management software?

The four objections that stall most construction technology proposals — cost, adoption, integration and rollout complexity — all have the same underlying structure: risk aversion. Each one is most effectively answered by reframing the risk. The status quo isn’t the safe option; it’s just the one whose costs don’t get tracked.

If you’ve walked in without answers to all four, you haven’t finished preparing.

“What does it cost?” Less than one rework event on a mid-size job, and that’s before you count the schedule impact or the claim exposure. The more useful frame is total cost of ownership vs. total cost of the status quo, and the Bluebeam ROI calculator can help you build that comparison before you walk in. Come with the math already done.

“Will the field actually use it?” This is the right question, and it deserves a real answer rather than a reassurance. AGC’s 2025 Workforce Survey found 43% of construction firms name lack of time to implement and train as a primary obstacle to tech adoption, because the most powerful platform in the category is worthless if the foreman won’t open it on his phone in a trailer with bad signal. Bluebeam earns 4.7 out of 5 on Capterra across nearly 1,000 verified reviews, and on G2’s head-to-head with Procore, it leads on “Meets Requirements” (9.2 vs. 8.8) and “Ease of Setup” (8.5 vs. 8.2). It was built around the drawing rather than added to a broader platform as a module, and that focus shows up in how quickly field crews ultimately adopt it. The document management software evaluation guide is direct about the principle: ease of use for field crews is the criterion that determines whether any of this pays off.

“How does it fit with what we already have?” Bluebeam integrates with the PM platforms and design tools most construction firms already run, and many teams use it for drawing and markup work inside a Procore or Autodesk Construction Cloud-managed project — the two are complementary rather than competing. The more productive question isn’t whether Bluebeam fits in the stack; it’s whether the tools you already have are talking to each other the way they should.

“What does rollout look like?” Shorter than the question implies. A focused document and markup tool isn’t an enterprise ERP implementation — run a two-week pilot on a live project with a small crew, and you’ll know more than any demo or feature matrix could tell you. The risk of a two-week pilot that doesn’t work out is two weeks. The risk of another year of the current workflow is whatever you calculated it to be in step one.

What to Bring Into the Room

You don’t need a slide deck for this conversation. You need three things, and they should ideally fit on a single page you can hand across the table.

One number that represents your firm’s own estimate of what the current workflow is costing — in rework percentage, RFI overhead, or PM hours spent on non-building work. Use the industry benchmarks to anchor it, round it to something defensible, and don’t get too precious about the precision. The point is that you did the math and the number isn’t small.

One story from a firm that already fixed it. DPR, Balfour Beatty, Premier Mechanical, Turner and Rudolph and Sletten are all in the public record, so pick the one that maps most closely to your firm’s project type and scale. A story from a contractor doing the same kind of work you do lands differently than a case study from a world the CFO doesn’t recognize.

One specific ask with a defined scope — not a category like “better document management,” but a workflow: standardized drawing sets across active projects, version control, field access in real time, markups tracked from issue through closeout. That’s something that can get funded. The category cannot.

“The firms winning in this market haven’t found ways to absorb the cost of document chaos. They’ve found ways to stop incurring it.”

The Conversation You’re Having

Here’s what’s worth remembering before you walk into that room: you’re not there to explain technology. You’re there to explain money, and specifically the money the current workflow is costing the firm every time a project runs.

The document management problem is a margin problem, a risk problem and a problem about what your most expensive and hardest-to-replace people are doing with their time. CFOs have those conversations every day — they’re fluent in all of it — and the only thing separating the people who win those conversations from the people who get the sympathetic nod and a deferral is whether they showed up in the CFO’s language or their own.

The firms that have already won this argument aren’t running better technology in any abstract sense. They’re running better information: one set of drawings that everyone trusts, markups that are tracked and auditable, RFIs that move in hours rather than days, and closeout packages that get built throughout a project rather than assembled in a panic at the end. The technology is almost beside the point — what changed was who had access to what and when.

You know your firm’s version of the problem, you have the numbers to show what it’s costing and you have proof from firms that already fixed it. Now you have the structure for the conversation. Walk in and close it.

Bluebeam’s Document Management for Construction 2026 Guide lays out the full framework, and the document management software evaluation guide walks through how to evaluate the options and match a platform to your team’s actual workflow. Both are worth having open before the meeting.

Frequently Asked Questions

How much does poor document management cost a construction firm?

Construction professionals lose more than 14 hours a week to non-optimal activities such as hunting for project data, according to FMI and PlanGrid. At fully burdened BLS wage rates, that is about $1,200 a week per construction manager. Direct rework adds about 5% of total project cost, per the Construction Industry Institute, and RFI processing adds roughly $860,000 on an average large project, per Navigant.

What is a typical profit margin for a commercial general contractor?

The CFMA 2025 Financial Benchmarker, drawn from more than 1,500 construction companies, puts pretax net income for commercial general contractors and construction managers at 4.4%. At that margin, a rework event equal to 5% of contract value erases the profit on the job.

How long does it take to pilot construction document management software?

A focused document and markup tool can be tested in a two-week pilot on a live project with a small crew. That is long enough to see whether field teams will use it, and the downside of a pilot that doesn’t work out is two weeks.

Does Bluebeam work with Procore and Autodesk Construction Cloud?

Yes. Bluebeam integrates with the project management platforms and design tools most construction firms already run, and many teams use it for drawing and markup work inside a project managed in Procore or Autodesk Construction Cloud. The tools are complementary rather than competing.

Put one live project on a single trusted drawing set.

Start Free Trial

Your Daily Reports Are About to Matter More Than Ever

Construction & The Economy, Technology & Innovation
Parth Tikiwala

About $36.8 billion a year in federal transportation money disappears Oct. 1. When funding gets tight, the questions about what got built get harder, and the projects with a clean record, from the model to the pour, will be the ones with answers.

If you build roads, bridges or transit, the news from Washington in early September sounded like a win. Congress kept federal highway and transit programs running through Dec. 11. No lapse, no frozen contracts, no bid openings pushed into next season while everyone waits on paperwork.

The bill passed the Senate 90-6 in August, cleared the House 370-48 on Sept. 1 and was signed the next day. The statements went out within hours. Most of them used some version of the same word. Averted.

The relief is warranted as far as it goes. When federal highway authority lapses, states can’t commit new federal dollars to projects. Lettings stall, awards slip and work gets shoved into a construction season with no room for it. Anyone who has lived through one is right to want it avoided.

The trouble is that the extension keeps the programs open. It doesn’t cover the money.

One deadline moved. The other didn’t.

The 2021 infrastructure law did two things at once. It set the rules and funding levels for federal highway and transit programs through this September, and it added a big pot of extra money on top of the formula funding states get every year. The extension renews the first. It lets the second run out Sept. 30. That’s roughly $36.8 billion a year in federal transportation funding gone starting Oct. 1, the yearly share of the $184 billion the law set aside up front. Transit and passenger rail take the steepest cuts in percentage terms. The American Public Transportation Association puts them at 20% for transit and 81% for passenger rail. Highway and bridge programs lose about $9.5 billion a year.

The extension buys 10 weeks, not a plan. The House Transportation and Infrastructure Committee approved a five-year, $580 billion highway and transit bill May 22, and the full House hasn’t voted on it. It also leaves out the extra up-front money entirely. The Senate hasn’t released a bill of its own. House Speaker Mike Johnson has said he wants to finish the job this year. With members heading home to campaign, that means the short session after the Nov. 3 midterms, with a Dec. 11 deadline.

So if you’re planning next year’s work, plan around three things: a short runway on the programs, a real cut to the extra money starting Oct. 1 and a fight over the next bill that happens with less money on the table than last time.

What tightening looks like on a project

Funding pressure doesn’t arrive as a headline. Rather, it arrives as a letting that slips a quarter, or a grant award that never gets announced. It arrives as a corridor project split into two phases, so the first one can go now and the second can wait for a bill that may or may not come.

Formula money, the dollars states get automatically every year, is the last to get touched. Competitive grants get squeezed first, because they’re the easiest thing to pause. For a design firm that turns into a decision about which pursuits are worth chasing. For a contractor it’s a bid schedule with holes in it. For a DOT or transit agency it’s a program that has to justify itself in more detail than anyone asked for in 2022.

There’s a second-order effect that concerns me more, and it lands directly on digital delivery, the shift from paper plan sets to designing, building and inspecting projects from 3D models and connected field data.

Owners in the middle of that shift are carrying a real cost. New deliverable standards, retraining, model-based workflows that take longer the first several times through. The payoff comes later. Budget pressure makes putting it off look like prudence. Push the pilot to next year, get through the cycle, pick it back up when the money settles.

That trade looks reasonable on a spreadsheet, and it’s the wrong one, for a reason that rarely gets said out loud.

The documentation is the argument

The pattern was consistent enough to plan around. When money gets tight, scrutiny goes up, and the scrutiny is about paperwork. It shows up as a request for detail the record was never organized to produce, landing on staff who have two weeks to reconstruct four years of work. Programs that can answer keep their standing. Programs that hand over a summary when someone asked for the underlying data lose an argument they didn’t know they were in.

None of that appears as a line item. It appears as a revised figure, a missed window, an application that went in thinner than it should have.

The next highway bill will be written by people asking what the infrastructure law bought. Every agency will have an answer. Some of those answers will be documented and some will be assembled from memory, and the difference will be visible.

Where the record breaks

Owners digitize the design side first, which is the rational order. Plan review is contained, it happens indoors, and the people doing it already work in software. The improvement shows up within a couple of quarters.

Then the project mobilizes and the record hits the shoulder of the road.

Field capture is the harder half by a wide margin. The inspector or superintendent is outside, often without signal, working a day that doesn’t accommodate careful data entry. What gets written down is written under conditions nobody would choose. That’s also where federal accountability lands, because the concrete pours, the materials tests and the daily reports are the evidence. Everything upstream is intent.

A project with a spotless plan review history and a reconstructed field record hasn’t closed the gap. It has documented one side of it well.

Texas is worth watching because it’s being honest

In September 2025, a 0.311-mile stretch of rural highway in Guadalupe County finished construction. FM 1977 runs from FM 621 to the Caldwell County line, a full-depth reconstruction with widened shoulders. It was the first Texas Department of Transportation project designed, bid and built through digital delivery, and the first where the 3D model, not a plan set, was the contract document contractors bid on. It was let at about $1 million, well under the engineer’s $1.9 million estimate. That estimate priced in risk for a first pilot, and the bids came in lower anyway.

TxDOT has since directed each of its 25 districts to nominate a project of their own that goes to letting by the end of 2027. All 25 have. The second pilot to go out for bid, a bridge replacement on FM 44 in the Paris District, was scheduled to let this month.

FM 1977 is instructive less for what worked than for what TxDOT published about what didn’t. The pilot page notes that new model standards, including pay-item data attached to model elements and cleaned-up 3D surfaces for machine control, introduced a learning curve and added design effort, with the hope of recouping the time on later projects. The design review had its own stumble. Only one reviewer fully engaged on the project’s primary review platform, so the team switched late to a Bluebeam session to keep things moving. Full disclosure, since that’s my company: TxDOT’s lesson wasn’t about which software. It was, instead, that reviewers need earlier training and clearer expectations. The agency’s wrap-up is just as plain about the field. During construction, not every planned digital workflow got used, and a shared online workspace with the contractor fell through over IT security requirements. That’s the half of the record where accountability lands. A TxDOT engineer described the program in the agency’s own newsletter as stair-stepping, and put the first-generation pilots at roughly a 25% change from traditional practice rather than a wholesale one.

That is an unusually candid thing for an agency to put in writing, and it’s worth more to the rest of the country than any success metric in the file. It tells the other 49 states how big the first step actually is, which is the number that decides whether anyone attempts it. It also makes the case against waiting. A 25% change absorbed across several projects is manageable. The same change attempted in a hurry, in 2028, because a funding application needed records nobody kept, is not.

The part that’s yours

Congress will decide what replaces the infrastructure law, on what schedule and whether the extra money comes back. That belongs to the committees, and I have no business prescribing a mechanism from where I sit. The industry is right to press for a multiyear bill and right to say so loudly, because the math behind the Highway Trust Fund, the federal gas tax account that pays for most of this, hasn’t worked since 2008. Pretending otherwise helps nobody.

Still, none of that timeline is yours to set.

What happens on your own projects is. Whether the field record gets captured once or rebuilt later. Whether the documentation behind the infrastructure law’s five years can be produced on request or has to be assembled on deadline. Whether the districts that have never run a digital project volunteer for one or wait to be told.

That goes for everyone on the job, not just the owner. The contractor’s daily reports, the inspector’s notes, the designer’s markups and the photos from the pour are the record. Captured once, cleanly, they’re ready when someone asks.

None of that will make an editorial page. No floor speech is going to mention a crew that captured its inspection record once instead of twice, and no reporter is going to call about it.

Those decisions are sitting in front of you anyway, and they have a deadline of their own.

Parth Tikiwala is head of government affairs at Bluebeam, a construction technology company and member of the Nemetschek Group, and head of global public sector for the Nemetschek Group. He previously served as acting director of technology modernization and data, and acting director of science and infrastructure policy, at the U.S. Office of Management and Budget, Executive Office of the President.

Capture the field record once, not twice.

Try Bluebeam Free

Six Shortcuts Are Speeding Up Data Centers. All Six Land on Electrical

Construction & The Economy, Technology & Innovation
Frank Kalman

The Wall Street Journal catalogued the shortcuts driving the data center boom. Here is what they mean if you are the one wiring the building.

A lot of money is going into building data centers faster. The Wall Street Journal on Sept. 1 catalogued six of the techniques doing it, with reported time savings running from a few hours to six months.

Bloomberg, using data from the market research firm Sightline Climate, reported that of roughly 12 gigawatts of U.S. data center capacity slated for 2026, only about a third was really under construction. The reasons cited: shortages of transformers, switchgear and batteries.

Plenty of people have strong opinions about the buildout. Set those aside. The work is being bid and built right now, and it asks something different of the trades than the work before it.

What are the six techniques speeding up data center construction?

Six unrelated products from six unrelated companies. Read them as a group and one strategy shows up. Speed is not coming from anyone working faster on site. It comes from decisions moved upstream, out of the field and into a shop or a spec.

Three land directly in electrical scope.

The drilling robot. DEWALT and August Robotics launched it this year reporting 99.97% accuracy across more than 90,000 holes. It drills anchors for server rack stops and the structural legs carrying overhead MEP. That robot sets the geometry your hangers land on.

The connectors. 3M’s expanded beam technology cuts plug-up from roughly three minutes per connector to as little as 30 seconds, across hundreds of thousands of them. Connection speed stops mattering if you connected the wrong thing quickly.

The factory-built rooms. The big one. Rooms get assembled in a plant, trucked in and set. Your coordination package is no longer a guide for a crew. It is an input to a production line.

Every shortcut there buys schedule certainty by giving up flexibility, and flexibility is what the field has always used to make imperfect drawings work. None of that hurts when the information is right. All of it hurts when the information is wrong, because the correction that used to happen on a Tuesday afternoon now happens on a truck.

How much of a data center’s cost is electrical work?

Data centers are not big office buildings with more outlets. On this work the electrical scope has moved from a supporting line on the cost sheet toward the center of it, which changes who carries the risk when something goes wrong and how much leverage the trade has at the table.

On an AI data center, electrical is the biggest line on the cost sheet. Turner & Townsend’s 2025 data center cost index puts electrical systems at 48% to 54% of construction cost, ahead of mechanical, and several times the shell. The inputs are getting more expensive, too: copper wire is up 17.9% from a year ago.

The International Brotherhood of Electrical Workers estimates 45% to 70% of the construction budget goes to the electrical subcontractor. That covers the sub’s full scope rather than one line on a cost sheet, and the union publishes it without showing its math. Still, a cost consultant and the electricians’ union, measuring different things, land in the same place: roughly half the job.

Electrical and power share of construction cost

Project type	Electrical/power share	Source
Typical commercial building	15% – 20%	Industry baseline
AI data center (2018)	28% – 32%	CBRE, JLL cost-stack analyses
AI data center (current)	35% – 45%	CBRE, JLL cost-stack analyses
AI data center, full electrical sub scope	45% – 70%	IBEW estimate

The IBEW figure measures the electrical subcontractor’s entire scope rather than a single line on a cost sheet, which is why it sits well above the CBRE and JLL numbers.

Why can’t electrical mistakes get fixed in the field on these jobs?

On most work, rework is a cost line. Somebody eats it and the job moves. On this work that tolerance is mostly gone, and the reason has less to do with craftsmanship than with what the building does once it is energized.

Chris Doyle has been in the trade since 1988 and works as a detail manager at Redwood Electric on an active data center project in Reno, Nev. The work has not only grown in volume, he said. It has raised the technical floor, “from single lines to ladder logic and the understanding of how these systems are supposed to work.” Not only for the people running jobs. “That’s the people that are building it.”

“At the end of the day, if there’s any mistakes in a UPS system, you’re talking about millions of dollars in downtime. The bar is very, very high, not just because we as an institution here want it to be, but because that’s what the industry demands.”

Most rework was never a field execution problem to begin with.

A Construction Industry Institute study of industrial projects found design changes, errors and omissions drove 79% of the cost of quality deviations. Construction deviations drove 17%.
A 2018 FMI and PlanGrid survey traced 48% of U.S. rework to poor project data and miscommunication, about $31.3 billion that year.

The field has spent decades absorbing mistakes made upstream. Prefab takes that away. The errors don’t stop. The shock absorber does.

The standards are still catching up. IEEE authorized Project P3710 in June 2025 to write design guidance for modular data centers, and it is still in development. The NEC and UL frameworks everybody works under were written for equipment installed in place, not built in a plant and shipped.

We have written before about what a 40-minute drawing hunt costs and why projects run late for reasons that have nothing to do with weather.

What’s really slowing down data center construction?

The six techniques all attack labor hours on site. Neither of the two things ultimately gating these projects lives there. Equipment and qualified people are procurement problems, not productivity problems, and no amount of speed in the field moves either one. Both get decided long before a crew mobilizes.

Gear. Wood Mackenzie’s Q2 2025 survey put power transformers at 128 weeks and switchgear around 44. Its 2026 read, reported by Data Center Knowledge, has substation transformers past 160 weeks, up from roughly 140 in 2023.

Electrical equipment lead times

Equipment	Lead time	As of	Source
Power transformers	128 weeks	Q2 2025	Wood Mackenzie
Switchgear	~44 weeks	Q2 2025	Wood Mackenzie
Substation transformers	~140 weeks	2023	Wood Mackenzie via Data Center Knowledge
Substation transformers	160+ weeks	2026	Wood Mackenzie via Data Center Knowledge

Wood Mackenzie’s most recent per-equipment survey is from Q2 2025. Its 2026 substation transformer figure was reported by Data Center Knowledge.

Lead times are cyclical and will come down. Distribution transformers already have, falling from more than 100 weeks in 2023 to about 30. Schneider Electric announced in March 2025 it will invest more than $700 million in U.S. operations through 2027, including switchgear and medium-voltage production. Shorter lead times still do nothing to restore the ability to adjust a prefabricated room after it comes off the truck. The shortage is a cycle. Front-loaded risk is a change in how the work gets built.

People. Google’s May 2025 white paper “Powering a New Era of American Innovation” reports that nearly 10,000 American electricians retire or change careers each year against about 7,000 new entrants.

The Bureau of Labor Statistics counted 818,700 electricians in the U.S. in 2024, projects 9% employment growth through 2034 and expects about 81,000 openings a year.

Watch what the companies with the most riding on these things are doing about it. Google.org funded the electrical training ALLIANCE, the joint IBEW-NECA apprenticeship organization, targeting 100,000 electrical workers and 30,000 apprentices. Siemens set a goal of bringing 200,000 electricians and manufacturing experts into the workforce by 2030.

One company doing that is a press release; two is a signal about which input is hardest to source. The money is going to apprentices, not another robot.

What should electrical contractors change to win this work?

The answer to industrialization is not working faster. Rather, it is refusing to rebuild the same thinking on every job while everything around you gets standardized. The firms that will hold margin here are the ones treating their own process as an asset with a version number, not something reassembled from memory each time.

Doyle argues data centers are more alike than most projects. “There’s an execution that can be about 80% the same,” he said. “So we shouldn’t have to start at zero like most projects tend to.”

The industry is moving his way. The Open Compute Project is building standardized reference designs for AI data centers, and Google contributed the first one, a facility design built around its TPU systems. But OCP’s own leadership describes the first wave of AI data centers as siloed, built with little coordination between vendors. The repeatable share runs high inside one customer’s program and lower across customers.

That tells you where to aim. The roughly 20% that does not repeat is where the job is won or lost, and the only way to have capacity left for it is to stop re-solving the 80%.

For an electrical sub, four habits:

Build tool sets once and reuse them, so takeoffs start from your standard, not from scratch.
Verify before release, not on arrival. Anything headed to a fab shop gets checked while changing it is free.
Treat submittals and RFIs as schedule work, not paperwork handled between other things.
Catch the conflict in design review while it is still a markup, because the field can no longer catch it for you.

The part that can’t be prefabricated

The pouring, the coating, the drilling, the connecting, all of it is getting faster and further from the hands that used to do it. The judgment that has to happen before any of it starts has not moved.

Somebody still has to know the drawing is wrong: that the anchor pattern doesn’t match the rack layout; that the routing won’t clear the structure; that the gear on the submittal isn’t the gear on the plan.

That work is getting more valuable, not less. The trades who do it consistently and can prove it get the calls.

Frequently asked questions

What percentage of data center construction cost is electrical work?

Turner & Townsend’s 2025 data center cost index puts electrical systems at 48% to 54% of construction cost, the largest single category. The IBEW estimates 45% to 70% measured as the electrical subcontractor’s full scope. Both figures exclude servers and IT equipment, which make electrical a much smaller share of total project cost.

How long are transformer and switchgear lead times in 2026?

Substation transformer lead times passed 160 weeks in 2026, up from roughly 140 weeks in 2023, according to Wood Mackenzie as reported by Data Center Knowledge. Wood Mackenzie’s most recent per-equipment survey, from Q2 2025, put power transformers at 128 weeks and switchgear at about 44.

Why is prefabrication riskier for electrical contractors on data center projects?

Factory-built assemblies lock decisions in before the crew arrives, so a drawing error the field once fixed with a Tuesday afternoon adjustment has to be caught before the assembly ships. That matters because most rework starts upstream: research from the Construction Industry Institute and from FMI traces it largely to design errors, changes and poor project data rather than field execution.

Is there a shortage of electricians for data center construction?

The Bureau of Labor Statistics counted 818,700 electricians in 2024 and projects about 81,000 openings a year through 2034, many from retirements and career changes. Google.org is funding an IBEW-NECA effort to train 100,000 electrical workers and 30,000 apprentices, and Siemens set a goal of adding 200,000 electricians and manufacturing experts by 2030.

Bluebeam builds software for specialty contractors, MEP trades and energy and infrastructure teams. See how AI-assisted drawing review changes what gets caught before the truck shows up.

Catch the conflict while it’s still a markup.

Start Free Trial

The Same Seven Things Keep Slipping Through Design Review

People & Project Spotlights, Technology & Innovation
Frank Kalman

Common design errors in construction drawings, and why reviewers keep missing them.

Sometime around 2018, Jean-Pierre Trou sat down at his dining room table in Austin, Texas, with 250 pages of drawings and a red pen. Three stories, 90,000 square feet, architectural plans stacked against structural and MEP. The review took him a week. We wrote about that table.

Anybody who’s done that work knows how it goes. You’re sharp on sheet 12. You’re a different reviewer by sheet 240.

The errors that get through are almost never the hard ones.

Why do the same drawing errors keep showing up?

Reviewer attention is a budget, not a switch. The industry treats a missed dimension as a discipline problem and answers it with more checklists and sets of eyes. The research on how people search points somewhere else, at a failure that shows up in any job built on hunting for something rare.

Ask why a missing door tag reached the field, and you’ll get some version of somebody should have caught that. Yet there’s a better answer: The errors get missed because they’re rare, and rare is where trained eyes fail.

Start with the airport. In a visual search study published in Nature, people missed 7% of targets when the target turned up in half the images. Drop it to 1% of images and misses jumped to 30%. Same people, same task. The only change was how often there was something to find.

That’s the problem with a clean set. Two hundred good sheets teach you the next one is good too, so you stop looking a little sooner each time. You’re not scanning worse, just quitting earlier.

Then there’s the study that should bother anyone who reviews drawings for a living. Researchers asked 24 radiologists to hunt for lung nodules in CT scans and dropped an image of a gorilla into the last case, 48 times bigger than the nodules they were hunting. Eighty-three percent didn’t see it. Eye tracking showed most of them looked right at it.

That one gets thrown around as proof that experts are no better than anybody else. The authors drew a narrower conclusion. Even expert searchers, working in their own field, are vulnerable to this. Their attention was tuned to nodules, which is why a gorilla didn’t register. As one of the researchers later put it, these weren’t bad radiologists, just human ones.

Nobody has run this test on a drawing set. The mechanism carries. The proof doesn’t.

Stack that on real volume. At workplace design firm Oktra, a drawing pack runs about 60 sheets and passes 100 with subcontractor drawings in, through roughly eight rounds of design review, with comments coming from design, coordination, cost, the client and suppliers.

What are the most common errors found in construction design review?

Most design errors aren’t exotic. They fall into a handful of shapes that repeat from job to job, which is what makes them worth naming out loud. A reviewer working a short list of known shapes is doing something different from one reading every sheet with the same attention and the same fatigue.

The research sorts these by cause. The UK’s Get It Right Initiative traces a large share of construction error back to design that is uncoordinated, incomplete, late or miscommunicated, and its design guide is built around fixing that upstream. Useful if you run a firm.

On the sheet, at 9 p.m., it looks like this.

The error	What you’re looking at	What it turns into
Dimension conflicts	The same element dimensioned two ways on two sheets	Whichever sheet the field reads first becomes the building
Missing or wrong tags	Doors, fixtures or equipment on the plan but not the schedule	A submittal for something nobody can identify
Gridline mismatches	Structural and MEP working off grids that don’t agree	A clash nobody modeled
Missing sheets	A detail callout pointing to a sheet that isn’t in the set	An RFI on day one of the trade’s work
Spec-to-drawing conflicts	The drawing says one product, the spec says another	Somebody prices the wrong one
Scope gaps	Work sitting in the seam between two packages	Both trades assume it belongs to the other
Revision drift	A change that landed on one discipline’s set and not the rest	The set is current and wrong at the same time

These stick around. A 2019 study of BIM coordination on two public projects found 67% of design discrepancy issues, and half of repeated clashes, still open when construction finished. They don’t just happen. They survive.

The last two are process failures wearing a drawing problem’s clothes. Comments scatter across email and meeting minutes, versions blur and markups get repeated. Oktra’s answer was one authoritative markup set for every downstream use. “If we’re going to review something, we do it properly once, not three different times for three different audiences,” Creative Director Jenny Edwards told us. Keeping the set straight is a document management problem, and it feeds everything downstream.

Which of these needs a human?

The useful split isn’t hard vs. easy. Rather, it’s whether a question has one right answer or takes somebody who knows how the building goes together. That line decides what belongs to software and what stays with a person, and it’s where most review time gets spent in the wrong place.

Four of the seven have exactly one right answer.

A tag matches the schedule or it doesn’t.
A grid aligns or it doesn’t.
A referenced sheet exists or it doesn’t.
A revision landed on every set or it didn’t.

None of that takes field experience, and every one is rare on any given sheet, which is where the search research says people fall down.

The other three need somebody who’s built something. A spec-to-drawing conflict is often an ambiguity, not a mismatch. A scope gap depends on how the work gets sequenced. Those are judgment calls, and judgment is what a reviewer should spend the good hours on.

Why this took so long to automate comes down to what a drawing is. A square might be a wall. It might be a table. You know which because you know what you’re looking at. Software didn’t. Tech entrepreneur Shir Abecasis made the longer case in AI Can Read Your Specs. But Can It Read Your Drawings?.

That’s where the tools have landed. In Bluebeam Max, Smart Review scans the set for the tag, sheet and gridline family. Smart Overlay flags changes across disciplines and drawing scales, which is revision drift, handled directly.

What changes when the first pass isn’t manual?

The value of catching an error early is easy to assert and hard to prove, which is why the numbers around rework are so unreliable. What changes when the scanning gets handed off has less to do with hours saved than with what a reviewer can put their name on.

Start with a number that cuts against the pitch. Peter Love has spent two decades studying rework. This year he went into a contractor’s actual cost records and found rework before completion averaged 0.38% of contract value, or 0.76% counting post-completion fixes. He notes the rework costs reported in the literature vary widely, depending on who is defining rework and how they are counting it.

The gap is the story. Nobody knows what this costs, including the people paying for it.

Downstream is easier to see. Navigant’s study of 1,362 projects and roughly a million RFIs found 9.9 RFIs per $1 million of construction cost, with a median 9.7 days to get an answer. Every one of those is a question somebody had to stop and ask.

Two caveats. That data is from 2013 and skews toward Australia and New Zealand, and Navigant’s own warning holds. A high RFI count doesn’t by itself prove the design was bad.

What teams report has less to do with hours. Rudolph and Sletten completes preconstruction reviews 30% faster running them in Studio. CannonDesign managed design review on a $2 billion hospital, which is the volume problem in its purest form.

Steve McElwee, director of construction at Foresight Custom Homes, put it in bid terms. AI reviewing the drawings alongside your own process lets you bid more accurately, knowing nothing slipped through.

The part nobody says out loud

Nobody in this business expects a perfect set. Ask any architect whether one exists and watch the reaction.

That’s not a complaint so much as the premise. If the set was never going to be perfect, catching things is the whole job, and the boring half of that shouldn’t depend on somebody being as sharp on sheet 240 as on sheet 12.

The table is still where quality control happens. The complete guide to construction design review covers the rest of the workflow, and the design review and QA/QC workflow page covers how it runs in Bluebeam.

Frequently asked questions

What are the most common errors found in construction design review?

Seven categories account for most of them: dimension conflicts between sheets, missing or inconsistent tags, gridline mismatches across disciplines, missing sheets and dead references, spec-to-drawing conflicts, scope gaps between trades, and revision drift where a change reaches one discipline’s set but not the others.

Why do experienced reviewers miss obvious drawing errors?

Because the errors are rare. Visual search research shows miss rates climb sharply as targets become less frequent, and a study of 24 radiologists found 83% failed to notice a gorilla image inserted into a CT scan while hunting for lung nodules. The authors concluded that even expert searchers, working in their own field, stay vulnerable to missing what they are not looking for.

Can AI catch design errors before construction starts?

It can catch the deterministic ones. Tag mismatches, gridline inconsistencies, missing sheets and unpropagated revisions are rule-checkable, and software doesn’t get tired at sheet 240. Conflicts that require reading design intent, like an ambiguous spec-to-drawing discrepancy or a scope gap between trades, still need a person. Bluebeam Max runs the first pass with Smart Review and Smart Overlay.

Does a high RFI count mean the design was bad?

Not on its own. Navigant’s research found roughly 9.9 RFIs per $1 million of construction cost across 1,362 projects, and cautioned that a high RFI count does not by itself indicate defective plans. Treat RFI volume as a signal worth investigating rather than a verdict on the design.

Let software take the first pass at the boring half.

Start Free Trial

Joey Brown Runs Commercial Construction Projects. Then He Plays Everyone Who Ruins Them.

People & Project Spotlights
Victoria Lucia

A commercial project manager by day, a construction satirist by night. Joey Brown thinks the two jobs are the same.

Halfway through his interview for this story, Joey Brown’s video froze.

The audio held. Someone suggested he cut the camera and keep talking. “Do you guys hear me now?” he asked, and then picked his answer back up where he’d left it.

That’s Joey Brown in one scene: two jobs running at once, and neither one waiting on the other.

Two Jobs, Same Guy

By day, Brown manages schedules, submittals and subs as a commercial construction project manager, most recently with Taft Construction in Culpeper, Virginia. By night (and lunch break, and red light) he’s @__joeybrown__, 31,200 followers deep into a hashtag he’ll happily admit is corny: #MakeConstructionFunAgain.

“It’s a little on the nose,” he says. He keeps using it anyway.

The two jobs shouldn’t fit together as easily as they do. They do, because both come down to the same thing: getting people who don’t usually listen to each other to hear one another. That’s the whole job on a jobsite. It turns outit’s also the whole job on a phone screen.

A ‘Total Accident’

Ask how the second job started, and Brown doesn’t dress it up.

“It was a total accident,” he says. A college friend who’d interned alongside him at a Hampton Roads contractor started posting on TikTok and talked him into trying it. Brown figured it might be a side hustle, nothing more. Then the messages started.

“People actually care about what I have to rant about,” he says. “Maybe I should do this.”

By then, he’d already put in the years on the first job: an internship at 20, a mechanical engineering technology degree from Old Dominion University, a climb to project manager.

The career came first. The camera just found someone who already knew how to explain one department to another, then pointed that instinct at a bigger room.

The Bit, the Instruction

Scroll his social media feed and you’ll find two different things happening.

One half is a rotating cast of characters, all of them blaming each other for the same jobsite headache: GC project manager, GC superintendent, architect, engineer.

The other half is straight instruction. Sprinkler lines vs. ductwork. How to read coordinated drawings. What’s going on with the bottom half of a fire hydrant. Hard hat brim back, wraparound sunglasses, usually filmed from a truck cab, whichever job he’s doing that day.

Both halves work for the same reason. Brown has sat in every one of those chairs, so the jokes land like recognition instead of mockery.

There’s a bill attached to the joke, too. A 2018 survey of nearly 600 construction professionals conducted by FMI Corp. found that poor communication accounted for 26% of all rework in the U.S., a potential cost of roughly $17 billion a year. Bad project information accounted for another 22%.

“I’ll have a civil engineer who will message me and say, ‘Hey, I really never thought about that,’” he says. Even the general contractor bit took some getting used to. “I just would like to think that I’m one of the better ones,” he says. “So, it was difficult for me to make fun of myself, which I have practiced.”

Making It Look Like a Good Job

The construction workforce is older than the country’s. The median age in the industry is 42, a year above the national labor force, according to the National Association of Home Builders’ analysis of the Census Bureau’s American Community Survey. Baby boomers still made up 14.2% of it as of 2023, and most of them are heading toward the door.

Brown’s read on it is less a strategy than a bet.

“I have a suspicion that it’s going to solve itself with supply and demand,” he says. “The more social media covers financial wellness, and then, of course, people are willing to pay tradespeople more and more, I think that more and more young men and women will join the trades.”

He doesn’t claim to be part of the fix. The research suggests people like him already are. In Thumbtack’s 2024 “Future of Skilled Trades” report, 55% of Generation Z respondents said they were considering a trades career, up 12 points in a year, and two out of three said social media had increased their interest. Among those with a college degree, it was three out of four.

The bet may already be paying off. Gen Z’s share of the construction labor force more than doubled between 2019 and 2023, from 6.4% to 14.1%.

Brown would probably wave that off. He’s just a guy in a hard hat, on camera, making the work look like a good time and a good living.

Fair to Both Sides

Brown doesn’t go easy on either half of his own audience.

On the industry he works in: “The white-collar side of the industry can be so serious and so polished,” he says, “and it is such a huge mess in comparison to the people in the field actually performing the work.” He doesn’t stop there. “There is just so much animosity at all times,” he says. “It doesn’t need to be that way.”

He’s just as hard on the people who’ve never set foot on a site. Walk one onto a big commercial project and the noise gets them first. Steel hanging overhead. Dump trucks backing up. Off-road equipment working the dirt. Voices carrying across an open slab. “People yelling,” as Brown puts it, “not because they’re angry, but because that’s how you have to communicate sometimes.” Most spend their first couple of days simply overwhelmed by it.

Then something else lands. “This is actually genuinely dangerous, loud and intimidating,” he says.

None of which comes through from the outside. “You have no real perspective whatsoever of what it is like to pour concrete when it is 100 degrees outside for an entire summer,” he says. His comparison: someone mulching their yard on a Saturday, calling it hot.

“That’s nothing.”

Boxes of Paper

The same gap between old and new shows up in his tools. Brown remembers 60-plus-page submittals, five to eight copies hauled into a single meeting, boxes of paper riding in the back of a truck to get there.

Now the work itself lives inside the software, and Brown’s reaction the first time he showed it to someone outside construction has stuck with him. “How is this gate kept?” he says of Bluebeam. “Is it just construction?”

Every Building Is a First Attempt

Ask Brown to explain why the industry moves the way it does and he gives you the most honest answer available.

“Construction is behind schedule because it is immensely complicated, and every single building is the first attempt at that structure,” he says. The crews aren’t new to the work; they’re new to this building. “It’s their first time building that building.”

No assembly line, no template, no muscle memory to fall back on. He’s not wrong, though the industry has more say in it than that. McKinsey’s construction productivity research found that only half the firms it surveyed kept a standard design library, and it names repeatable design elements as one of the sector’s biggest missed opportunities. Some of the one-off-ness is the work. Some of it is habit.

“The immense amount of complexity within construction is lost,” he says, “and no one really understands it until they’ve lived it for a while.”

That’s the industry Brown works in every day, and the one he spends his nights explaining on social media. He just happens to keep standing inside two versions of it at once.

“Everyone should look at the industry and try and put themselves in other people’s shoes,” he says, “because all of it is difficult in its own way.”

Including his own.