Somewhere around 1995, a surveyor stopped writing numbers in a field book and started pushing a button. Somewhere around 2005, a dozer started steering itself. Somewhere around 2015, a robot laid brick on a real wall for money.
And somewhere around right now, a project engineer is retyping a number out of a PDF into a spreadsheet cell, by hand, for the four hundredth time this month.
That gap is what I think the real story of construction automation is. It worked really well in one direction and barely worked at all in the other. The closer a task is to dirt, steel and concrete, the more it got automated. The closer it is to paper, the less it did.
The 1990s: the drawing leaves the board
The decade started with drafting tables still in use and ended with a lot fewer of them. CAD had been around since the seventies, but the nineties are when it went from a mainframe curiosity to the thing every architect and engineer actually drew on.
The field got its own version. Total stations replaced the transit and the chain. Layout that used to take two people and a day started taking one person and an afternoon. Laser screeds showed up and concrete flatness went from a craft to a tolerance you could hit over and over.
Scheduling went digital too, and this is where I see the first warning sign. Software made the schedule easier to draw, but it didn't make anyone better at building to it. A CPM schedule with four thousand activities can be a very confident drawing of a plan nobody builds to.
The 2000s: the model, and the machine that drives itself
Two things happened in this decade that still shape the industry.
The first was the model. Revit shipped in 2000 and Autodesk bought it in 2002, and over the next few years "BIM" went from a vendor word to a contract requirement. Clash detection turned a lot of coordination meetings from an argument into a screenshot. That's a real win you can measure, and I want to say so plainly because most of what got promised in this decade didn't pan out.
The second was GPS machine control. A grader that knows where the finished surface is doesn't need stakes, and it doesn't need a grade checker walking in front of it. Earthwork got faster and more accurate, and it got less dependent on the one guy who could read the plan. I'd call it the most complete automation success the industry has ever had, and almost nobody outside heavy civil talks about it.
Meanwhile the paperwork moved onto the internet and stayed just as much work. Document management moved onto the web. Procore was founded in 2002. Bluebeam gave everyone PDF markup. The RFI log went from a binder to a spreadsheet attached to an email, which changed where it lived and didn't change how anybody did it.
The 2010s: the jobsite gets a camera and a robot
This was the decade of the pilot project.
Drones went from a novelty to standard practice for site survey and progress photos, and that one stuck. Laser scanning got cheap enough to use on a renovation instead of only on a refinery. Reality capture companies figured out that a 360 camera on a hard hat does a better job than a written report.
The robots got more attention and proved less. A semi automated mason laid brick faster than a person on long straight runs and wasn't any use on anything else. Autonomous excavation companies raised real money and got real machines onto real dirt, in narrow applications. Concrete 3D printing produced houses that got a lot of press and didn't change how houses get built.
Layout robots deserve a mention because they were the quiet win of the decade. A machine that prints the wall lines on the slab from the model takes out a whole category of human error right where errors get expensive.
The decade also gave us a cautionary tale. Katerra raised well over a billion dollars to industrialize construction end to end and went bankrupt in 2021. A lot of people took that to mean construction resists automation. I think the more useful lesson is that it resists automation brought in from the outside by people who haven't done the work.
Autodesk bought PlanGrid for about $875 million in 2018. Drawings on an iPad, with punch lists and field reports alongside them, were worth nearly a billion dollars.
The 2020s: the machines can read
Everything before this could see, move or measure. The current wave can read, and that's a different capability pointed at a different problem.
Computer vision started tracking progress against the model without anyone filling in a percentage. Language models started doing something more interesting: they can read a document and answer a question about it. They don't do it perfectly and they need review, but a machine that can take a forty page spec section and tell you what's in it didn't exist in 2019.
That matters because it's the first automation wave pointed at the paperwork.
What never got automated
Look at what an automated jobsite in 2026 actually looks like. The dozer knows the grade. The layout is printed on the slab. The drone flew the site this morning and the model updated itself. The concrete came from a batch plant that adjusted the mix on the way over.
Inside the trailer it's a different picture.
Somebody is retyping numbers out of a PDF into a spreadsheet. Somebody is lining up two versions of a schedule by eye. Somebody is reading back through an email thread to figure out what got decided three weeks ago, because the decision lives in that thread and nowhere else. The estimator is checking a takeoff against a drawing by looking at both and counting.
I've done my share of that work. On a school renovation where we were doing an addition and a gymnasium, I sat there with a stack of two giant spec books, read both of them front to back and made the submittal log by hand in Excel. Then I had to wrangle all the subcontractors and suppliers to get the documents together. It felt like it was never going to end.
None of that is hard the way grade control is hard. Grade control needed satellites, a control box and a machine that could act on a centimeter. The trailer work needed somebody to care about it, and for three decades nobody did. I think there are two reasons for that.
The first is that automation follows the demo. A machine that steers itself makes a thirty second video that sells a machine. A robot laying brick makes the news. A person spending Thursday afternoon retyping a table doesn't make any video at all, and it's never been in anybody's keynote.
The second is that it's nobody's territory. Trailer work sits in the gap between the design team's software and the field team's software, and anything that lives between two systems doesn't belong to either one. Both vendors can point at it and say it's the other one's job, and they're both right, so the work stays manual for another decade.
Now think back to PlanGrid. Autodesk paid roughly $875 million for a product built mostly around viewing and marking up documents in the field, rather than writing them, reconciling two of them or answering a question out of one. To me that number shows how much of this job is documents, and how little of that work anyone had bothered to automate.
That's why I find the current wave more interesting than the robot videos. Reading is the first capability aimed right at the half of the job that never got a machine. Whether it holds up on a real project is a separate question, and it's the one worth asking every time somebody demos something.
That's what this publication is going to cover: the tools, who they actually help, and whether what they claim holds up on a real project. Some of them are very good, and some of them don't do much more than a spreadsheet does.