TMIConstruction productivity has barely moved in 50 years. Not because the labor is bad. Not because the materials are worse. Because the information flow on a job site is broken in ways that compound every day the job runs.
An RFI sits in an email thread for six days. The subcontractor who needed the answer made a judgment call on day three. That judgment call may or may not match what the engineer eventually specified. Nobody knows until the inspector shows up.
A change order gets scoped and priced in the trailer, approved verbally by the owner's rep, and written up four days later when someone has time. The crew has already done the work. The pricing is now a memory reconstruction.
The foreman knows the east wing masonry pour is scheduled for Thursday but doesn't know the inspection record from the last comparable pour on this project, which flagged a recurring slump issue with that supplier's batch. That information is in a report somewhere. Nobody put it in front of the person who needed it.
The information gap, not the labor gap
Every one of these failure modes is an information problem. The data exists. The RFI response was eventually sent. The change order was eventually documented. The inspection report was filed. The problem is that the information existed in one place at one time, and the person who needed it was somewhere else.
This is what drives the productivity problem in construction. Not that people are working slowly, but that they're making decisions without the information they need, re-doing work that didn't need to be re-done, and spending hours of coordination time to answer questions that should have been answerable in seconds.
"The information is always somewhere. The problem is it's never where I am. I'm walking the site, not sitting in front of a terminal."
What changes when information follows the work
The shift is from systems that store information to systems that surface it. The difference sounds subtle. It isn't.
A superintendent walking a section of work shouldn't have to call the PM to ask when that section was last inspected. That information should be on their phone before they get there. A foreman reviewing submittals for a structural connection shouldn't need to search through a project folder. The relevant documents should appear in context of the work they're reviewing.
Change orders and scope as a data problem
Change orders deserve special attention because they are the single largest source of margin erosion on construction projects. Not because owners are difficult, though some are, but because the documentation and pricing discipline required to capture every scope change in real time is extraordinarily hard to maintain under field conditions.
An AI system that captures change orders when they happen, prices them from the project's established labor and material rates, routes them for approval before work continues, and logs them in a way that makes the scope trail clear at closeout, that system doesn't replace the project manager. It makes the project manager's job something other than chasing paper.
The labor problem is real. There are not enough skilled tradespeople and there won't be for a generation. But the productivity problem in construction is not primarily about headcount. It's about how much useful output each person on the project can deliver when the information they need is reliably in front of them.