In markets across the country, owners are becoming increasingly sophisticated in demanding alternative project delivery to reduce costs and time while also improving quality and outcomes.

More companies are exploring technology as a route to those efficiencies. One such emerging technology is the use of artificial intelligence (AI) to reduce human error and optimize tasks.


For example, the creation of submittal registers is a time-intensive, detailed task that is prime for optimization. Earlier this year, McCarthy Building Cos. partnered with Pype AutoSpecs to solve this challenge. Pype is a software platform that filters project specs through a machine-learning algorithm to identify submittal register items the company provides to its partners on each project.

While McCarthy had systems for processing and housing specifications, engineers could search and retrieve information only manually using Bluebeam. This process required project engineers to spend considerable time validating search results and building the submittal register in Excel.

Submitting this process to further review revealed frustration from project executives over the amount of time it took an engineer to complete the task. For example, on the $250-million, 19-mgd Tomahawk Creek Wastewater Treatment Facility Expansion Project in Leawood, Kan., where the Pype technology was first tested, a project engineer spent three full weeks producing the submittal log.

In addition, McCarthy’s innovation and field applications team found that human error meant diminished quality. Finally, the team discovered difficulties with the version-review process. The team was required to sift through multiple iterations of the specs and could not identify or clearly explain the changes in each version.

Using Pype to complete the same process took only three hours. While the software wasn’t perfect, it did identify more than 99% of the individual submittal items in the timeframe. The company still dedicates an estimated 20% of the time it would take for manual review to complete quality-control review by an engineer.

In all, McCarthy tested Pype on four projects before implementing it company-wide. Currently, the company is using the software on more than 110 projects, 20 of those located in the Southwest.

The Pype software also integrates with Procore’s construction management software, and the company reports efficiency gains of between 200% and 500%.

Flying High

Machine learning isn’t the only new technology at the company. McCarthy also is using unmanned aerial vehicles on many jobsites.

Drones have been used on most jobsites since the company found that using them to take photos and videos eases communication among project teams, clients and trade partners. The technology allows teams to show stakeholders visuals in real time. Photos can then be matched with CAD and other documents to show exactly how work lines up with site plans.

Using drones equipped with cameras, the company performs inspections for things like windows and steel instead of putting employees in a manlift or exposing them to other fall hazards.

Looking to the future, McCarthy anticipates two major trends in the industry. They include integration of artificial intelligence into tools offered by construction technology platforms and more ties between reality capture solutions and artificial intelligence.