Did you miss a session from the Future of Work Summit? Head over to our Future of Work Summit on-demand library to stream.
Construction, particularly commercial construction, is expensive and time-consuming. In the U.S., the average commercial construction cost is around $490 per square foot, and most commercial buildings larger than 10,000 square feet take at least four months to build. As a 2020 survey from Levelset and Fieldwire shows, the process is rarely simple — fewer than 30% of contractors finish large projects on time and within budget. Errors are a significant contributor to overrun, with some sources estimating that between 4% and 6% of total project cost is rework-related.
Despite the fact that the construction industry has been historically slow to adopt new technologies, a growing crop of vendors are pitching software as the answer to these problems. Among them is SiteAware, a startup developing what it calls a “verification platform” for construction. Based in Houston, Texas, SiteAware leverages AI to keep a record of work performed and to alert customers of potential issues.
Injecting construction with AI
Founded by Guy Raz and Ori Aphek in 2015 as Dronomy, SiteAware uses fleets of drones to collect data from construction sites. Raz was formerly a senior researcher at GM’s R&D center in Israel before joining Pebbles Interfaces — which was later acquired by Meta (formerly Facebook) — to direct the electro-optics group. Aphek previously founded Optigo Systems, which develops electro-optical systems for defense applications.
Using SiteAware, customers get jobsite-level visibility into layout, workmanship, and omissions of things like concrete forms and steel and curtainwall embeds. The platform provides a breakdown of masonry installation and detailing, waterproofing and sealants, siding elements, and finishes and exterior elements verified by SiteAware’s “digital twin” technology.
Digital twins are essentially virtual representations of systems that span their lifecycles and use AI to help with decision-making. Platforms like London-based SenSat assist clients in construction, mining, energy, and other industries with creating models of locations for projects that they’re overseeing. Meanwhile, companies including GE offer products that allow companies to model digital twins of actual machines and closely track performance.
As for SiteAware, it autonomously records jobsite data and 3D design files entered into the platform for verification. Off-the-shelf drones and software supplied by the company create a virtual replica that can be accessed via a web browser. As the drones fly over a site, the 3D model is uploaded to remote servers and SiteAware’s photogrammetry tech superimposes models from consecutive scans to verify against building design documents.
SiteAware claims that it trained AI algorithms on over 100,000 databases of construction elements to compare each twin to the documents.
“We have unique datasets with hundreds of thousands of elements verified in the field and compared to plans as a ground truth. This rich dataset has enabled the training of AI models to reach a high level of accuracy,” CEO Zeev Braude told VentureBeat via email. “The rich data provided to managers and technical decision makers provides the certainty and confidence missing in construction field work today.”
With the AI in construction market anticipated to be worth $2.12 billion by 2023, a number of companies are developing products that apply computer vision, natural language processing, and other algorithmic techniques to different aspects of the construction process. Avvir’s AI system monitors construction projects for errors, while Swapp provides AI-powered software to automate construction planning. Gryps offers robotic process automation software focused on paperwork inefficiencies in construction. Meanwhile, StructionSite applies AI to construction project management.
SiteAware competes directly with HoloBuilder, Buildots, and OpenSpace, which similarly offer machine learning-powered construction site-reporting solutions. But the startup claims to be working with “some of the largest construction companies in the world,” including Hines and Ryan Construction. SiteAware’s customers had about 20 construction projects ongoing, totaling a combined $5 billion as of October 2020.
In a sign of investor confidence, SiteAware today announced that it raised $15 million, bringing its total capital raised to over $30 million. Vertex Ventures led the round with participation from existing investors Bosch, Axon, Horizon, and Israel-based lool Ventures.
“The company grew three times on its recurring revenue in 2021 [and we count] over 30 of the top general contractors and developers in the U.S. as customers. [The funds will be] used for expansion in the U.S. market [and creating] new features, including real-time AI-based verification and construction element-level analytics tools,” Braude added. “Digital technologies in the field that are impacting traditional construction execution need to be perfectly tailored to the field conditions and the cadence of construction teams to be implemented successfully.”
One barrier to 50-employee SiteAware’s future growth might be the construction industry’s reluctance to adopt technology platforms. While full-scale digitization could save $1.2 trillion in the design, engineering, and construction phases of projects alone, according to the World Economic Forum, the Levelset and Fieldwire survey found that only about half of construction businesses use some kind of digital application to manage tasks or coordinate field work. Twenty percent say they do not use any construction software at all.
As KPMG analysts write in a recent report, cost, scale, and a lack of existing use cases are a few of the factors keeping construction executives from investing in new technology. Most construction firms, the coauthors say, are waiting for competitors to take the first step toward the deployment of technologies to streamline workflows and improve data collection.”
“One example is the promise of small unmanned aerial vehicles, or drones, hovering over job sites for tasks such as aerial photography, site monitoring, and even material transportation,” the KPMG analysts noted. “A lack of clear regulations has kept many … firms from formally incorporating the equipment into their workflow.”
In the U.S., the Federal Aviation Administration (FAA) last year amended commercial drone regulations that replace the previous process to add a pilot certification requirement for night flights. Depending on the specifications of the drone, companies no longer need to apply for project-specific wavers to fly over people and moving vehicles — but autonomous drone operations still require FAA approval.
“We fly within the building’s footprint and at a height that doesn’t trigger any authorization from the FAA except in special circumstances,” Braude clarified. “When needed, we have been granted authorization in areas requiring it because of the previously mentioned implementation. We use third-party FAA licensed drone operators that do the data capture on behalf of the company.”
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More