Design and build with AI-based knowledge management

a 2016 Survey A McKinsey Global Institute study published in the Harvard Business Review ranked the construction industry as the second least digitally advanced out of 22 industries surveyed, behind agriculture/hunting. However, McKinsey’s follow-up articleRising infrastructure demands, a skilled labor shortage, and stakeholder demands for modernization are forcing the $12 trillion architecture, engineering, and construction industry to participate in a digital transformation. While breakthrough advancements such as autonomous construction robots and drones tend to get the attention, AI-powered knowledge management and data analytics represent a less visible yet highly impactful opportunity for the industry.

“AI isn’t just about increasing efficiency. It’s about enabling humans to work with AI to be more ambitious and creative, and ultimately achieve better outcomes,” explained Tristan Randall, executive of strategic projects at Autodesk, the industry-leading software developer, at a recent Society of Military Engineers workshop on AI applications in the AEC industry.

Pooja Jain, senior vice president of strategic innovation at WSP, a $3 billion global design firm, agrees. WSP’s building optimization tool, called DAISY, balances three key aspects of sustainability — daylight, operational carbon, and embodied carbon — to maximize sustainability overall. “DAISY allows us to prioritize sustainability from the beginning of the design, rather than implementing it later,” Jain says.

Optioneering during the construction phase and engineering is also made possible by ALICE Technologies. Founded in 2015, ALICE uses AI to “simulate and optimize construction execution across millions of scenarios to minimize risk, shorten schedules, accelerate planning, and recover from schedule delays,” says founder Dr. René Morkos.

AI-powered data extraction and knowledge management is expected to reduce the time designers spend searching for information by up to 30%. AECOM, an $8 billion global design firm, is addressing this issue with a proprietary in-house system that searches a variety of historical documents, including proposals, drawings, white papers, and cost estimates, while staying within the company’s boundaries. “By organizing our knowledge and providing references to source materials and employee names, we’re helping our employees be more efficient and avoid the ‘horror of the blank page,'” explains Carl Jensen, executive vice president of national government operations.

Knowledge management and AI also benefit the construction industry through inventory management and compliance during construction. Revaka is an AI-powered information retrieval tool for the construction industry. “Information search is time-consuming because there are too many apps to search,” says co-founder Fauzan Reza Maulana. Revaka’s ACE (Assistant for Construction and Engineering) uses machine learning to collate data from specifications, drawings, scope of work, and construction applications. “We save designers time by searching all these sources to answer specific questions, with citations, like, ‘What is the range of concrete costs used on previous projects over $2 million in Illinois?’” says Maulana. WSP uses computer vision to automate tedious data entry tasks, like vendor invoices and concrete mix certificates, significantly improving efficiency and quality control.

Kaushal Diwan is a portfolio manager at WND Ventures and executive sponsor of DPR’s Innovation and Research and Development group. DPR Construction is ranked the sixth largest contractor by Engineering News Record with $9 billion in annual revenue. DPR’s approach to an AI-assisted data strategy operationalizes data by incorporating analytics into decision-making. “We’re leveraging historical data for analytics during data entry to move from lagging analytics to leading and predictive analytics to make better, more data-driven decisions,” Diwan said. One interesting outcome of the pilot project is DPR’s use of generative design for drywall. DPR’s pilot, partnered with startup Hypar, significantly reduced material waste and improved safety by optimizing the layout of drywall, cutting it offsite, and shipping it to the job site along with stacking plans for installation.

Unlike other technologies, “the barrier to entry is low and you don’t have to be a large company to benefit from AI,” said Sal Nodjomian, CEO of Matrix Design Group. “Small businesses can easily leverage AI technology. The private sector moves faster than government, so industry must continue to proactively encourage its government partners to follow suit.” Workshop leader Lt. Commander Tim Dahms, U.S. Navy Civil Engineer Corps, agreed. “We need the A/E/C industry and those working for the federal government to accelerate the pace of learning to adopt new AI applications to reduce costs and improve schedule performance on the billions of dollars of design and construction contracts signed each year,” he said.

Autodesk is at the forefront of the industry’s digital transformation, investing about $1.4 billion, or about 25% of revenue, in research and development in 2023. That research has led to features like predictive analytics to help with decision-making in the Autodesk Construction Cloud, the Autodesk Validation Tool to check models against requirements, and AutoCAD Macro Advisor. “Advisor is one of the ways we make suggestions to designers to improve their use of our software, directly within the tool itself,” Randall says.

Jain cited key principles for WSP’s AI adoption: anchoring accountability, including governance and AI policies; talent prioritization, including change management, upskilling and training; and data strategy and governance, including data orchestration, safeguarding AI validity and avoiding risk. “Without the right guardrails, leaders sometimes put up gates,” Jain said. “Companies can say no to everything, but that’s not going to move the industry forward.”

“We tend to fear what we don’t understand,” Nojomian concluded. “An effective way to reduce fear and increase the implementation of AI is through education and thoughtful discussion, which SAME provides.”

Disclosure: I moderated one of the panels at this workshop. LinkedIn Or check out our other columns here.