-
THE PROBLEM The construction industry has been plagued with inefficiencies.
-
WHY IT MATTERS Broader development relies on built infrastructure.
-
THE SOLUTION Innovating how and with what we build.
July 31, 2025
When Tessa Lau, an expert in AI, machine learning, and robotics, undertook a home remodel in 2018, she was surprised to see how rudimentary the process was. Carpenters attended to each detail by hand. Reworking and mistakes proliferated, including the kitchen island almost being built in the wrong place. And, as is so often the case, the project ended up taking twice as long as planned and costing far more than originally quoted.
Anyone who has built anything, whether as a client or a contractor, knows this is how it goes. Building is an imperfect, iterative, and painstaking process. But Lau, who spent 11 years at IBM automating repetitive tasks, couldn’t accept that. “I knew there was a solution,” she says.
Lau started hanging out at construction sites and talking to contractors and workers there. “How does the process get held up?” she would ask. “Where do the mistakes happen?”
That’s when she learned about layouts. Before construction begins, contractors and tradespeople have to transpose the blueprints onto the building surface. Referring to the drawings, they meticulously check distances with a tape measure and, from their hands and knees, snap chalk lines. The process can take weeks and is rife with the potential for error, with any slight miscalculation triggering a cascade of problems down the line.
Lau’s solution is a Roomba-esque, AI-powered robot known as Dusty that—or should we say who—zooms around the worksite marking the layout. What can take the most skilled workers weeks to do, Dusty achieves in a couple of days with exact precision. “It takes about five seconds showing a demo to a grizzled old superintendent before they’re sold,” Lau says.


