How Big Data & AI Are Transforming Pre-construction

Prescient has taken a new approach to the construction and engineering of large buildings. Working alongside our engineers, our clients design their dormitory and apartment buildings in cyberspace. Their designs are then sent to our factories, where assembly lines of automated roll formers and robotic laser cutters manufacture each building’s steel core.

This approach has been made possible by Prescient’s Unified Truss Component System (UTCS), which has effectively standardized the construction of multi-unit housing. The UTCS allows these projects to be completed in record time. (As our president, Satyen Patel, recently said, “Buildings go up like erector sets.”) After a building’s components are forged in our factories, they are shipped to the construction site, where workers assemble them rapidly.

To enable subcontractors to benefit from the UTCS, Prescient is also developing a collaborative technology platform, the Digital Thread. This platform consolidates all of a project’s design, communication, and coordination in one place. The result is a notable decrease in disagreements, miscommunications, and errors.

The success of this approach has allowed Prescient to expand significantly. Our engineers are now standardizing the construction of balconies, stairs, decking, and sheathing. Our developers are digitizing the designs that mechanics, electricians, and plumbers must use. The twin forces of standardization and digitization are even transforming my own department, preconstruction.

Stuck in the 20th Century

In a traditional design-and-build company, the preconstruction department oversees all of the activities that must occur before a project can begin. This includes the management of planning, procurement, coordination, and, most importantly, estimating. In its assessment of a project, preconstruction has to establish a balance between the needs of the client and the costs of construction. Above all, its duty is to ensure profitability.

Most design-and-build companies lack the trades necessary to complete a large project on their own. When this is the case, preconstruction divides the project into packages, which are sent out as requests for proposal (RFPs). The department then compares the bids it receives. In its evaluation, preconstruction must identify the most cost-effective set of proposals, selecting the subcontractors that will get the job done without running over budget.

What I have just described is a straightforward process that represents the industry standard for operating a preconstruction department. It is also the reason why most preconstruction departments are stuck in the 20th century. More often than not, the job of cost management is split between a number of departments and subcontractors, each of which may use distinct methods for recording cost data. Without consistent sources of such data, most preconstruction departments struggle to accurately predict and control a project’s budget.

21st Century Tools

Prescient is not a traditional design-and-build company. In fact, we view ourselves as a technology company first, design-and-build second. Our core technologies, the UTCS and the Digital Thread, not only enhance our ability to predict costs, but also allow us to implement cost-controlling innovations.

By standardizing the construction of a building’s core, the UTCS effectively sets the budget for that project. From manufacturing to labor, every aspect of the budget can be estimated precisely. By gathering data on the progress of the project, the Digital Thread enables real-time cost management. This ensures that we adhere to our original estimate.

As director of preconstruction at Prescient, I’m working to further improve our ability to predict and control costs. My first step is to make my department more data-centric. During the construction of a student housing or apartment building, the Digital Thread generates a tremendous amount of data. To an extent, this data is already structured, meaning it has been recorded using a consistent method, and gathered at every point in the building process.

In a traditional preconstruction department, truly understanding the costs of construction requires thousands of engineering hours. In a data-centric preconstruction department, that understanding can be achieved in far less time. Imagine a scenario in which data scientists store the data streaming from the Digital Thread in a database, and then use that data to build advanced analytics. Those analytics can then be applied at a project’s start, allowing preconstruction to identify issues quickly and implement cost-saving measures in the early stages of the project life cycle.

As information gathers in the database, it becomes a valuable asset known as big data. Big data consists of immense data sets that can be analyzed by various forms of artificial intelligence. One such form, machine learning, can be employed by data scientists to progressively improve the accuracy of their analytics. Another such form, deep learning, can compare the data sets produced by different projects and identify patterns that would have otherwise remained hidden.

Using these 21st-century tools, the preconstruction department will generate insights that benefit the entire value chain. A better understanding of the drivers of costs will enable the creation of key performance indicators (KPIs) for every department and subcontractor involved in a project. The ability to monitor those KPIs in real-time will not only keep a project on budget, but also uncover opportunities to increase its profit margin. As our analytics grow more precise, RFPs will become more refined, and subcontractors will be selected based on their predicted performance. This is the future of pre-construction. It is the department we are building.

 

Written by: Bart van Praag