As more firms store their project data in the cloud, it offers the chance to dig through that data for new insights. Oracle has already offered some limited business intelligence dashboards for its cloud service, but the technology giant is now bringing machine-learning advice to company's scheduling and project data. 

Oracle's Construction Intelligence Cloud Service is a new product from the company, which can examine past project data and provide users with advice on their current and future projects. Intended as an AI "advisor," the service will provide bespoke advice to companies based on their past scheduling and project data, says Karthik Venkatasubramanian, vice president of data science and analytics at Oracle Construction and Engineering. "The industry standard is P6 data, which we're now using to do analytics and make predictions," he says. 

The system will be linked into a company's P6 and other project data, and will provide advice on current and future projects. "The insights and intelligence it will offer will prompt change," says Venkatasubramanian. "[Customers] will start to think about delays ahead of time and take proactive steps to deal with risk."

While basic business intelligence dashboards have been able to provide general warnings on when certain values are out of whack, Venkatasubramanian says that the Oracle Cloud advisor goes much further, watching dozens of metrics in order to generate its advice. "The number of factors now is 70-plus, and that will keep growing," he says. "The better the data, the wider it gets, we can consider the quality of a schedule: Is it up to par? Here are the areas to improve schedule quality." Issues with scheduling and planning, including allocation of project resources, is one area the advisor is designed to catch early in the process.

Venkatasubramanian adds that presentation of the AI advice is designed to be easily understood. Rather than present a dense report, it can instead show basic indicators on the likelihood of a project getting derailed or otherwise delayed. "The focus is on accessible machine learning, that [customers] feel comfortable using."

Oracle's AI advisor is intended for all stages of a project, including bidding and preplanning, explains Venkatasubramanian. "At the start of the project, you run this to help plan better, estimate better and bid better. When you create the first schedule, it starts to show you risk of delay." And once the project is underway, it can generate fresh advice each time the P6 schedule is updated, with the data flowing in automatically in Oracle's cloud service.

P6 integration is a key feature, adds Venkasubramanian. “You can go back to the P6 schedule it's advising you about. It will take you to the project and to the right screen. But it’s up to the individual manager for the project what do with it. We will not make the change for them [in the schedule].”

And it's that quality of an AI advisor, and not an AI decision-maker, that Venkatasubramanian says is key. "Are you worried about the risk of delay on your project, is that something you want to fix? These things rely on human intuition," he explains.

The cloud AI advisor is now available, and Venkatasubramanian says the company plans to further expand its capabilities over the next year.