There’s a lot of chatter these days about companies overstating their reliance on AI—and rightfully so. The initial hype over AI capabilities indeed led to labeling any machine learning or computer vision process as AI, where ‘automated’ or ‘automated with AI techniques such as machine learning or computer learning’ may have been more accurate.
Sometimes, it seems that the overstatement of AI in data-driven construction services is intended to make customers feel more confident in the data, when, in reality, incremental AI adoption is what truly serves customers.
After all, AI can’t do everything better than a human can—at least not reliably, and not yet. For all of the anticipation and speculation AI has garnered over the years, it has been limited in its autonomy and reliable application to date.
What customers should be looking for when searching for are solutions and services that emphasize incremental AI adoption in their workflows. These “balanced” solutions leverage the power and speed of AI when it’s best indicated and then rely on real human experts and proven technologies to handle the rest.
In the construction industry, this balance isn’t just a philosophy—it’s a practical reality. The stakes are high, the workflows are complex, and the margin for error is razor-thin. That’s why the best processes are built around a human-in-the-loop model, where AI accelerates and enhances workflows, yields more accurate data, and saves time and money, but never completely replaces expert human judgment.
Here are five ways that Hexagon balances lightning-fast AI technologies and the careful, informed brain power of real, experienced humans.
1. Progress reporting: AI speeds it up, humans keep it honest
Every construction company organizes schedules differently. Our onboarding process maps each customer’s bespoke schema to a universal standard, enabling consistent reporting. AI helps accelerate this by processing vast amounts of data quickly, but it doesn’t replace human oversight.
Progress reporting is a KPI for every project. And while AI can help track and predict, it’s not yet capable of delivering 100% accuracy. That’s why our process starts with human data capture and ends with human validation, ensuring confidence in every report.
On a recent hospital build, for example, AI flagged a delay due to weather. But the site supervisor, drawing on years of experience, knew the site’s drainage would prevent any impact. The pour went ahead as planned, saving time and avoiding unnecessary costs.
2. Firestopping: When safety is on the line, humans close the loop
Firestopping is a life-safety issue. AI helps generate reports from captured data, but the final responsibility lies with humans—whether it’s the general contractor during construction or the facilities team post-occupancy.
During construction, code inspectors often delay progress if firestopping isn’t up to spec. Our AI tools allow teams to identify and correct issues before inspections, avoiding costly delays and rework. But it’s still a human who signs off on the fix, every single time.
In hospitals, firestopping scans are even used before The Joint Commission audits. Failing those audits can mean losing Medicare or Medicaid funding. AI helps teams prepare, without fail—and for good reason—humans are accountable for ensuring safety and compliance.
3. Photo documentation: AI as internal spellcheck
Multivista, part of Hexagon, captures millions of images monthly on behalf of its customers. AI performs pixel-level QC to flag blur or poor quality, prompting our specialists to retake photos if needed. It’s like spellcheck for construction photography, ensuring the final archive is useful for years to come.
In one case, a blurry 360 photo flagged by AI was retaken before drywall went up. Ten years from now, that image will still be valuable to the owner. Without AI, that flaw might have gone unnoticed, and if information about what was under that wall was required down the line, costly demolition and disruptions would have ensued.
Clients don’t want to sift through hours of raw data—and they don’t want to pay us to do that, either. They want confidence that what’s captured is clean and usable. But at the end of the day, it’s still a human who’ll return to the job site and recapture what they missed.
4. Trajectory mapping: AI connects the dots, humans confirm
With GoCapture, users define a start and end point on a floor plan. AI automatically maps the trajectory between those points. On the backend, a human validates the output to ensure correctness.
This process is binary—either it’s right or it’s not. AI gets us 99% of the way there, but human QC ensures we don’t miss that final 1%. Incremental AI adoption is all about being honest about the balancing act. Knowing what AI can do better and where a human must step in is key to ensuring customers are getting the best data and the most value, every step of the way.
5. Earthwork and point cloud classification: AI gets you to the dirt
When drones or scanners capture point clouds, AI classifies millions of data points, identifying vegetation, vehicles, and terrain. This allows teams to strip away irrelevant data and focus on what matters, like raw dirt for grading.
This is an excellent example of AI doing the heavy lifting. Humans step in only to adjust or reclassify if needed. It’s fast, efficient, and accurate, but still guided by human priorities. And, once again, it’s evidence of incremental AI adoption: AI can do it better, and so it does.
Ai in construction is most powerful when paired with human expertise. Hexagon helps you capture, validate, and deliver construction data with confidence, saving time, money, and reducing risk. Contact us to get started.