Swinerton is a 130-year-old general contractor that also self-performs trades such as drywall, framing, carpentry, demolition and concrete. The company also runs a large renewable energy group for utility-scale photovoltaics. With so many projects happening at once, Swinerton wanted a way to predict whether or not a project would be successful – meaning one that is completed on schedule and on budget. Specifically, it wanted to use its historical datasets to determine which projects might be at risk, believing that there are indicators within the data that humans aren’t detecting. Swinerton believed that the answers might be found using machine learning and AI and turned to the experts at Avanade.
“The one thing that excites me is the possibility. With traditional technology dashboards and reports and analytics, it gives you a little bit of visibility, but it doesn’t predict. …What really excites me about this project is that predictability. Predicting an outcome from historical data.”
The company partnered with Avanade and Microsoft for a unique, co-innovation approach to building a proof-of-concept (PoC) – a data science competition. Multiple teams from Avanade used datasets provided by Swinerton to build a PoC tool that uses machine learning and AI to predict project outcomes. Teams presented their findings and insights to a panel of judges and a winning PoC was selected.
The data science competition proved that it was possible to predict project outcomes using data. With this valuable insight, Swinerton can identify at-risk projects and determine what changes can be made to fix them. The company is currently looking at how the PoC can be produced and deployed as a working solution.