Building A Home For AI In Construction

Building A Home For AI In Construction
Getty

Technology moves rapidly, and over the past 20 years, it’s been developing at a lightning-fast pace. Only recently has it crossed into the construction industry, where new software and tools continuously emerge and evolve. However, many of these breakthroughs face pushback when they’re born out of an aimless development process, where software developers found a hammer and are looking for nails rather than finding a nail and grabbing a hammer.

That isn’t to say all construction software development is helmed by excitable college kids just flexing their programming muscles. My company, for example, came into being after a then-project-manager grew fed up with document collection during closeout and working with half-baked submittal logs during precon. He came to me with these problems, and we spent the next five years developing our automated solutions to them. And we’re not alone, either — several key players in the industry have taken significant strides toward developing technologies using artificial intelligence to prevent seemingly inevitable delays in their project schedules.

Take Bechtel, the EPC firm headquartered in Reston, Virginia, which has over 25,000 projects under its belt, including the English Channel Tunnel, the Bay Area Rapid Transit (BART) system and the Hoover Dam. Bechtel’s team has built itself a big data and analytics center of excellence (BDAC), which can process a data lake of a whopping 5 petabytes (5 million gigabytes) to shape its AI-based tools. Chief among these is its photo recognition technology, which labels worksite photos for clients and has saved the company $2 million.

Bechtel also relies heavily on its natural language processing (NLP) technology to read and parse contracts, claims, RFPs and other documentation, reducing the time it takes to create estimates from weeks to mere hours. David Wilson, Bechtel’s chief innovation officer, also announced the expansion of its AI development to include a host of human resources tasks, including schedule creation around material availability, labor shortages and even the weather.

And Bechtel’s not alone. AI development has become one of the fastest-growing sectors in the tech industry, with Tractica predicting the market for AI to grow to $11.1 billion by 2024 (up from $202.5 million in 2015) and Marketwatch predicting the market for AI in construction to grow to over $2.11 billion by 2023. With that much growth, a wealth of data is needed to train these systems’ pattern-recognition abilities for their intended uses. Just as construction professionals learn through experience, AI learns through analyzing past data and looking for patterns that human analysts may miss.

As a result, the data collection and analytics industry has seen a boom mirroring AI’s boom, leading major construction firms like Kiewit Corporation to invest in data. Kiewit, having acquired the software company InEight to develop solutions for its projects, has been investing in SAP’s ERP technologies to track and analyze its data for the past seven years. This data set has formed the basis of information InEight uses to develop its technology for Kiewit. Without it, Kiewit would be left grasping at straws.

Amidst the rapid innovation tech and AI are seeing, it’s no surprise that everyone is rushing to be the Steve Jobs of AI. However, instead of developing a tool with a specific end use in mind, many developers are throwing AI at the wall and seeing what sticks.

Such is the idea behind the expansion of IBM’s Watson, an illustrative AI example outside of the construction industry. Watson is an advanced supercomputer originally developed to play against humans in Jeopardy. After its success, IBM decided to develop the question-answering AI for a wide variety of commercial uses, including oncology. However, retrofitting Watson into these roles turned the computer into a jack-of-all-trades and master of none, having prescribed unsafe treatments for cancer patients.

Even when AI tools are developed perfectly and are masters of their assigned tasks, the industry still faces pushback. After all, if a robot does what a human does but better and longer than a human can, what happens to the human?

David Wilson, the chief innovation officer of Bechtel, told BIM+ that AI has a long way to go. “AI and the pending robotic apocalypse are very exaggerated at the moment,” he said. He added that while AI may be able to automate certain repetitive tasks, “I certainly don’t see a robotic army replacing humans anytime soon.”

He went on to state that he and developers like him consider the experience on-site to be the ultimate measure of a technology’s efficacy. “The goal is to make sure we are using technology innovation to get the right resources to the right person, at the right place, at the right time.” 

Technology and AI are unique in that they mimic abilities that thus far have been the sole trades of humans — critical thinking, pattern recognition and reading — but can integrate with machines and tools that have physical abilities beyond humans. Major players across all industries have begun to see the benefits of utilizing this new toolset, saving them time, money and materials. The construction industry is starting to embrace this wave of innovation, too, realizing that when these tools are focused and built with a problem in mind, they give PMs and engineers tools with the same level of specialization to work with as the crews have on site.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives.
Do I qualify?

“>

Technology moves rapidly, and over the past 20 years, it’s been developing at a lightning-fast pace. Only recently has it crossed into the construction industry, where new software and tools continuously emerge and evolve. However, many of these breakthroughs face pushback when they’re born out of an aimless development process, where software developers found a hammer and are looking for nails rather than finding a nail and grabbing a hammer.

That isn’t to say all construction software development is helmed by excitable college kids just flexing their programming muscles. My company, for example, came into being after a then-project-manager grew fed up with document collection during closeout and working with half-baked submittal logs during precon. He came to me with these problems, and we spent the next five years developing our automated solutions to them. And we’re not alone, either — several key players in the industry have taken significant strides toward developing technologies using artificial intelligence to prevent seemingly inevitable delays in their project schedules.

Take Bechtel, the EPC firm headquartered in Reston, Virginia, which has over 25,000 projects under its belt, including the English Channel Tunnel, the Bay Area Rapid Transit (BART) system and the Hoover Dam. Bechtel’s team has built itself a big data and analytics center of excellence (BDAC), which can process a data lake of a whopping 5 petabytes (5 million gigabytes) to shape its AI-based tools. Chief among these is its photo recognition technology, which labels worksite photos for clients and has saved the company $2 million.

Bechtel also relies heavily on its natural language processing (NLP) technology to read and parse contracts, claims, RFPs and other documentation, reducing the time it takes to create estimates from weeks to mere hours. David Wilson, Bechtel’s chief innovation officer, also announced the expansion of its AI development to include a host of human resources tasks, including schedule creation around material availability, labor shortages and even the weather.

And Bechtel’s not alone. AI development has become one of the fastest-growing sectors in the tech industry, with Tractica predicting the market for AI to grow to $11.1 billion by 2024 (up from $202.5 million in 2015) and Marketwatch predicting the market for AI in construction to grow to over $2.11 billion by 2023. With that much growth, a wealth of data is needed to train these systems’ pattern-recognition abilities for their intended uses. Just as construction professionals learn through experience, AI learns through analyzing past data and looking for patterns that human analysts may miss.

As a result, the data collection and analytics industry has seen a boom mirroring AI’s boom, leading major construction firms like Kiewit Corporation to invest in data. Kiewit, having acquired the software company InEight to develop solutions for its projects, has been investing in SAP’s ERP technologies to track and analyze its data for the past seven years. This data set has formed the basis of information InEight uses to develop its technology for Kiewit. Without it, Kiewit would be left grasping at straws.

Amidst the rapid innovation tech and AI are seeing, it’s no surprise that everyone is rushing to be the Steve Jobs of AI. However, instead of developing a tool with a specific end use in mind, many developers are throwing AI at the wall and seeing what sticks.

Such is the idea behind the expansion of IBM’s Watson, an illustrative AI example outside of the construction industry. Watson is an advanced supercomputer originally developed to play against humans in Jeopardy. After its success, IBM decided to develop the question-answering AI for a wide variety of commercial uses, including oncology. However, retrofitting Watson into these roles turned the computer into a jack-of-all-trades and master of none, having prescribed unsafe treatments for cancer patients.

Even when AI tools are developed perfectly and are masters of their assigned tasks, the industry still faces pushback. After all, if a robot does what a human does but better and longer than a human can, what happens to the human?

David Wilson, the chief innovation officer of Bechtel, told BIM+ that AI has a long way to go. “AI and the pending robotic apocalypse are very exaggerated at the moment,” he said. He added that while AI may be able to automate certain repetitive tasks, “I certainly don’t see a robotic army replacing humans anytime soon.”

He went on to state that he and developers like him consider the experience on-site to be the ultimate measure of a technology’s efficacy. “The goal is to make sure we are using technology innovation to get the right resources to the right person, at the right place, at the right time.” 

Technology and AI are unique in that they mimic abilities that thus far have been the sole trades of humans — critical thinking, pattern recognition and reading — but can integrate with machines and tools that have physical abilities beyond humans. Major players across all industries have begun to see the benefits of utilizing this new toolset, saving them time, money and materials. The construction industry is starting to embrace this wave of innovation, too, realizing that when these tools are focused and built with a problem in mind, they give PMs and engineers tools with the same level of specialization to work with as the crews have on site.

Read More