Forty-seven seconds. That is how long Oak Ridge National Laboratory’s FLAT tool needs to tell you whether your foundation is level. It fires a 360-degree laser scan across the wet concrete, feeds 680,000 points per second into an unsupervised machine learning pipeline, and hands back a color-coded map of every high spot and low spot on the entire slab. Red means screed it down, blue means add mud. The finisher corrects the problem with a bull float while the concrete is still plastic, and that correction costs essentially zero dollars and maybe ninety seconds of labor.
Now consider what happens without the scanner.
Your inspector shows up the next day, after the concrete has cured for eight hours, carrying a 10-foot straightedge that dates conceptually to the Roman Empire. He sets it on the slab, slides a gap gauge underneath, and writes down whether the deviation exceeds the ASTM E1155 tolerance. He does this at five or six locations he picks himself, because the standard lets the technician choose where to test. Then he leaves.
Those five spots represent less than one percent of the slab’s surface area. A guess. That is what a $400,000 home sits on.
How FLAT Actually Works
FLAT stands for Flat and Level Analysis Tool. Nolan Hayes and his team at ORNL’s Building Technologies Research and Integration Center built it specifically for construction sites, which are messy. A 360-degree scan of a job site captures not just the slab but also rebar stacks, form boards, Porta-Johns, and a foreman’s truck parked too close to the pour. FLAT’s segmentation algorithm, trained using unsupervised machine learning on point cloud geometry, strips all of that away and isolates the concrete surface automatically.
Once segmented, FLAT runs automated flatness and levelness analysis per ASTM E1155 and returns FF/FL numbers for the entire surface. Not five spots. Not one line. Every square foot of the pour.
The scanner itself is a terrestrial laser scanner, something like a Leica BLK360, which captures a full 360-degree scan in under 30 seconds at 4mm accuracy from 10 meters. Set it up on a tripod in the middle of the slab. Press the button. Walk away and come back to a complete point cloud of the entire foundation surface.
What Fifteen Dollars a Square Foot Feels Like
Grinding hardened concrete runs about $15 per square foot according to iScano, a commercial scanning firm. Say the straightedge missed a 200-square-foot high spot. Three thousand dollars. Gone. And the framing crew is standing around billing you by the hour while the grinder works, which adds another $800 to $1,500 in delay costs depending on the size of your crew.
Compare that to fixing the same high spot during the pour. Zero. A finisher taps it down with a bull float, the concrete cooperates because it has not yet begun to set, and the total incremental cost for what would have been a $3,000 problem tomorrow is nothing more than ninety seconds of a finisher’s time.
| Scenario | Catch it wet | Catch it cured | Miss it entirely |
|---|---|---|---|
| Fix cost (200 sq ft area) | ~$0 | $3,000–$4,500 | $5,000–$12,000+ |
| Schedule delay | None | 1–2 days | 1–3 weeks |
| Who pays | Nobody | Concrete sub | Lawyers decide |
That last column is the ugly one. When an out-of-level slab causes cabinets that won’t sit flush, flooring that cracks along one edge, or a curtain wall anchor that runs out of its ±1/2-inch adjustment range despite the slab being within ACI 117’s ±3/4-inch tolerance, the dispute goes to mediation. iScano reports this slab-versus-glass conflict is a growing source of construction litigation in 2026.
Break-Even Math for a Production Builder
I ran the numbers for a builder pouring 20 slabs a year. ACI 302.1R acknowledges that most residential concrete is placed with minimal quality control, which means defect rates in the field run higher than commercial work. Using a conservative 12.5 percent rate of slabs exceeding tolerance (one in eight pours), and a $3,000 average rework cost per incident, the expected annual cost of not scanning is $7,500.
| Option | Annual cost | Break-even |
|---|---|---|
| Buy a Leica BLK360 (~$16,000) | $0 after purchase | ~2.1 years (43 homes) |
| Rent a scanner ($626/month) | $7,512 | Immediate (2.5 avoided events cover it) |
| Hire a surveyor ($500–$800/slab) | $10,000–$16,000 | Never (more expensive than rework) |
| Do nothing (current practice) | $7,500 expected rework | N/A |
Renting wins. Easily. At $626 a month, you need to avoid just 2.5 grinding events per year to break even, and at a 12.5 percent defect rate across 20 homes, you are statistically expected to hit exactly that many. Every defect the scanner catches while the concrete is still workable is pure profit.
Buying the scanner outright makes sense only if you are pouring more than 40 homes a year or if you can also use the scanner for as-built documentation, which most production builders should be doing anyway but aren’t.
Maybe You Just Need a Better Finisher
Strongest case against scanning: a skilled concrete crew with a good screed operator and a laser-leveled form system already achieves FF 25/FL 20 on residential slabs without any post-pour inspection at all. Most slab problems trace back to bad subgrade prep, poor form setup, or rushed pours, and no amount of scanning after the fact will fix soil that settles unevenly or forms that weren’t braced properly. If you spend $7,500 a year on scanning instead of spending that same $7,500 on hiring a better subcontractor who does not need a machine to tell him the slab is off, you have assembled an expensive digital archive of failures that a more skilled crew would have prevented in the first place, which is an argument that sounds devastating in a trade magazine comment section but ignores the central reality of residential construction labor markets in 2026.
Partially valid. But the data cuts in both directions. Even the best crew pours in variable conditions: afternoon heat accelerates cure time, rain threatens open pours, and the Ready-Mix truck is 45 minutes late so the first batch is setting up while the second batch arrives hot. FLAT’s value is not that it replaces craftsmanship; it catches the one in eight pours where conditions conspired against a crew that was doing everything right.
What FLAT Does Not Do Yet
FLAT has been demonstrated at two housing developments. Two. The peer-reviewed ISARC 2024 paper by Hayes, Maldonado, Tang, and Hun validated the algorithms on a single slab-on-grade point cloud. No published data exists on how many defects FLAT caught that the straightedge method missed at those demonstrations, or how the tool performs on post-tension slabs, pier foundations, or suspended decks. ORNL says future work includes all three, but right now FLAT is a slab-on-grade tool validated in a small number of field trials.
Licensing terms for FLAT remain unclear, which matters if you want the specific ORNL software rather than the general laser-scanning approach. It is a DOE-developed tool, which typically means technology transfer through a licensing agreement with ORNL, but no commercial product exists yet. You can replicate the workflow today using any terrestrial laser scanner and point cloud processing software like Autodesk ReCap, but you lose the automated segmentation and instant FF/FL reporting that make FLAT fast enough to use during a pour.
My 12.5 percent defect rate is an estimate based on ACI commentary about minimal QC in residential concrete, not a measured failure rate from a controlled study. If the actual rate is lower, say 5 percent, the scanner rental barely breaks even at 20 homes per year, but if it runs higher, the economics become a clear win. Nobody has published residential slab defect rates at scale, which is itself a data point worth noting: the industry does not measure the thing that costs it money.
If You Pour Concrete for a Living
Rent a scanner for one month, scan every pour you do in that window, and compare the FLAT-style analysis to what your straightedge check would have found. If you catch even one defect the stick would have missed, the rental paid for itself three times over, and you have your answer on whether to keep going.
If you are a homeowner in new construction, ask your builder one question: “How do you verify the foundation is level?” If the answer involves a straightedge and five spots, you now know what they are not checking.