Rishabh Aggarwal is CTO of Raise Robotics, a company that builds robots for facade installation and other high-risk construction-site work. Before Raise, he worked on satellites with NASA during his graduate studies at CU Boulder. He also helped grow Tortuga AgTech from one agricultural robot to 150. In this episode of the AEC Business podcast, we talk about why robots are still rare on construction sites, why pilots can mislead, and where robots pay for themselves.
Start where people should not have to go
Rishabh’s view is that robots should go first to places where people should not have to work. Warehouses are now designed around robots, so automating them is technically easier. Construction sites are unstructured, hazardous, and often cold, and the industry’s productivity has stagnated for decades.
Raise started with facade brackets after the team watched a worker lie on his belly at the slab edge, tightening bolts while holding on to lanyards so they wouldn’t fall. A robot can do the same work from behind the safety rail while the operator stays inside on the floor with a tablet.
Why robots remain rare on sites
Robots spread first to factories and then to warehouses, where tasks repeat, and conditions stay predictable. Construction sites differ from project to project and change week by week, with different lighting, materials, and people moving around.
A robot trained on one site meets features on the next one that it has never seen. Rishabh calls this the generalizability problem and expects robots to become common on sites within the next five to ten years.
From pilot to scale
“Pilots are rigged for success,” Rishabh says. A pilot runs a narrow task in a controlled setting with the best engineers on hand, so most variables are removed. When many robots work across many sites, those variables come back: high winds close the site, brackets arrive late, or the customer changes the design.
Raise handles this by fitting into the contractor’s existing way of working. The robot arrives on a truck, gets lifted onto the floor, and takes over a defined scope of work, and the contractor doesn’t have to change its sequence.
A robot that paid for itself
On one facade project, the contractor had no experienced layout markers available, and the whole schedule was on hold. Raise deployed its robots for layout marking, and the customer estimated the work saved about $1 million in rework.
Inaccurate layout can leave glass panels that fail to meet in the middle, which leads to removal, reinstallation, and disputes over who was at fault. The project showed the team that customers value accuracy and repeatability as much as speed.
Wheels, legs and cheaper hardware
Rishabh is skeptical of humanoid robots in construction. He suggests choosing which tasks to automate first, then the robot’s physical form. Construction materials often weigh hundreds of pounds, and site surfaces are mostly flat and already built for wheeled equipment, so legs add complexity without a clear benefit.
Hardware has also matured. A LiDAR sensor that cost $ 75,000 to $ 80,000 a few years ago now has a capable equivalent for about $ 2,000, and the main work has shifted to choosing off-the-shelf components and integrating them into a reliable system.
What comes next
Raise recently announced a partnership with Genie, the maker of scissor lifts and boom lifts. The companies are mounting Raise’s general-purpose robotic platform on lifts to autonomously coat and paint at heights of 20 to 30 meters.
Workers who now spray-paint large structures from ladders can operate the machine from a safer position. You can learn more at raiserobotics.ai or contact Rishabh on LinkedIn.
Watch on YouTube
Watch the interview on YouTube.
Listen to the episode and subscribe to AEC Business wherever you get your podcasts
Podcast: Play in new window | Download
Subscribe: RSS
