We’ll be blunt. Construction has a huge productivity problem. In fact, it’s bigger than you probably imagine, because construction isn’t just failing to keep up with the gains of the overall U.S. labor force, which have averaged about 2.1% annually since 1994. No, construction’s productivity has actually declined from 1970 to 2020 by 40%, according to the National Bureau for Economic Research. And this productivity problem isn’t just an issue for construction — it’s a problem for the entire global economy. Construction directly and indirectly accounts for 13% of the world’s GDP.
Compounding the productivity problem, construction is also suffering from a massive — and likely long-term — labor shortage. To keep up with projected demand, the industry will need to attract an additional 546,000 workers on top of its usual pace of hiring, according to research from the Associated Builders & Contractors. Unfortunately, unless conditions undergo a rapid turnaround, that’s unlikely to happen. Last year, construction averaged 390,000 job openings each month, which is the highest level since stats have been recorded for the industry.
Other industries such as manufacturing have been able to increase productivity, despite sizable challenges of their own. But in construction, we are only in the early stages of what will be a massive transformation.
Suffolk Tech sees AI and automation as the ultimate solution to the industry’s productivity challenges. With our new fund, we invest in solutions and leverage our platform to help rapidly scale the right solutions across the industry. We firmly believe that these technologies can help the industry reverse course to start seeing the kind of productivity growth other industries have experienced.
But, of course, our position begs the question: why is construction still relying on manual processes and technologies that are decades old? Is construction made up of a bunch of technology-hating Luddites?
A Low-Margin, High Risk Industry
The answer to that last question is easy: a strong and emphatic, “No.” Though the explanation why is a bit more complex…
First, it’s important to understand that construction is a high-risk, low-margin business. The typical net profit margin for engineering and construction is less than 3%. That’s comparable to commoditized industries such as grocery stores (2.1%), electronic components (4.1%) and auto parts (2.6%).
As for risk, construction has a ton. In manufacturing, for instance, the environment is largely closed, and conditions inside a facility are controllable. Processes are also highly repeatable, because manufacturers make the same product over and over, gaining efficiency. Construction, on the other hand, is largely outdoors, subject to unpredictable weather and ground conditions. Each building is different, so processes must change from project to project. So, many of the factors that determine the success or failure of a project are completely out of a single company’s control.
Additionally, in construction, time is extremely precious. Late projects incur penalties that can easily wipe away profits, and given how much leverage is involved, owners need to start seeing revenue from their investments as soon as possible. A failed or late project can tie up all the stakeholders in years of costly lawsuits and red ink.
The fact is, in a high-risk environment with low margins and aggressive deadlines, the appetite for changing processes and adopting new technologies historically has been understandably low. Why take a chance on a new technology that could fail and put the entire project at risk? Construction managers see far less risk remaining with the status quo. Nevertheless, the pressures to deliver more buildings, faster, cheaper, and more sustainably are turning many stakeholders towards seeking better technological solutions to previously manual and slow processes.
Near-Term Opportunities for Automation and AI in Construction
To succeed, technology entrepreneurs must first understand the unique challenges of the industry. But that’s just the start. Their solutions must also be:
Able to deliver rapid ROI: If a construction organization is going to take the risk of incorporating new technology into their workflows, the benefit needs to be significant and immediate.
Easy to use: In construction, people are extremely busy and working to aggressive schedules. They simply do not have time to learn how to use a complicated UI.
Mature: While there are construction organizations like Suffolk Construction that are eager to help entrepreneurs develop and improve their solutions, the market at large expects technologies to be ready for prime time.
Designed with construction’s unique challenges specifically in mind: If horizontally-oriented, general-purpose technology solutions would work for construction, they would have adopted them.
Now that we’ve described the traits a technology solution needs to have for construction, let’s dive a bit deeper into automation and AI, specifically. Generally speaking, there are two spheres where it’s applicable. The first is physical automation, which includes equipment, labor and building management. The other is digital automation for use cases such as data analytics, model-driven design, project management, and others.
There are a multitude of potential use cases under each sphere, but here are a handful of the ones we believe have big near-term potential.
AI-Assisted Design: Currently, the processes for building models, plans, and designs are mostly manual and, probably surprising to many, even paper-based. It’s slow with a high risk of introducing human error. Contractors, for instance, spend weeks creating detailed models prior to build out. It’s a reactive process with very slow iterations, and decisions are often based on gut feelings. A good example of a company that is transforming how trades can leverage design to improve their process is Augmenta.
Augmenta uses AI to create electrical (and eventually mechanical and plumbing) designs within minutes vs months. Their designs are also tied to the bill of materials, which makes estimating fast and accurate. Considering it can take up to two months to create these designs manually with frequent changes required, employing this tool presents a game-changing advantage. Currently, Augmenta can create designs for fully detailed, code compliant, and constructible electrical raceway systems. In the near future, the company will add capabilities for creating plumbing, mechanical, and structural designs.
Data Analytics and RPA (robotic process automation): Capturing and analyzing data about a jobsite is a huge challenge. Manual data entry is, of course, slow and error prone, and manually captured images are often not comprehensive enough and can’t be analyzed at scale by humans. OpenSpace uses machine vision and AI to provide a complete, up-to-date digital view of the jobsite from 360 degree images captured from a camera mounted on a worker’s hard hat or, if available, an autonomous robot.
OpenSpace’s AI goes well beyond creating a digital jobsite view. It can show progress over time, and compare the actual site to a model to identify discrepancies early, when they are less expensive and easier to address. Contractors can also use the solution to simplify communications about scope of work and to document changes.
Robotics: If robots can perform tasks – especially dangerous ones – that are ordinarily completed by humans, they can help alleviate the labor crunch while improving safety, accuracy, and productivity. But the construction environment is super challenging for robots. In manufacturing, robots are ubiquitous, but factory floors are very different from construction jobsites, which are unstructured and always changing. Robots can’t work in isolation; they have to work alongside human workers safely and effectively. They must also be rugged enough to function in bitter cold and scorching heat, braving dust, standing water, uneven floors, and other hazards.
In the last decade, however, robotics has seen tremendous advancements, especially in their ability to move autonomously and collaborate with both humans and other robots. As a result, we’re starting to see young companies roll out robotics solutions designed specifically for construction.
Rugged Robotics, for example, replaces the slow and manual process of field layout with an autonomous robot that marks architectural and engineering (A/E) designs directly onto unfinished floors. It’s faster, more accurate and less expensive than marking floors manually.
Another company, Canvas, makes a worker-controlled robot that uses AI and machine vision to finish drywall 1.5x faster than a completely manual process. Incredibly, nearly every component for the robot is off-the-shelf, including the initial AI and machine vision algorithms, which goes to show just how much robotics and AI have advanced. With Canvas, one worker and a robot can do the work of an entire team faster and more consistently, which is a huge plus for an industry in the midst of a huge labor shortage. Plus, the Canvas robot creates very little dust, creating a healthier work environment with little to no cleanup afterwards.
Construction can see the same kind of productivity gains other industries have experienced through automation and AI, but these technology solutions must specifically address the unique challenges of the construction industry.
There’s still a lot of work to be done, even in getting things like baseline data to enable true AI and automation use cases (which we are also focused on solving with our focus on connected industry). But, one thing is clear, construction needs to change and new AI and automation tools can increase productivity across the industry. The idea of automated construction is not new (as can be seen from Villemard’s depiction of future of building in his 1910 image), but we see great potential on the horizon.
If you’re building a startup that addresses these or other pressing issues in the built environment, please reach out to our team.