Real Time Project Success Prediction

Downer Group is an integrated services company active in Australia and New Zealand. Listed on the Australian Securities Exchange and New Zealand Stock Exchange as Downer EDI, Downer is an ASX 200 company.

Construction projects generate large volumes of data, but project status reporting often remains a subjective exercise. Making use of that data in a transparent and meaningful way requires integration with various 3rd party systems and correlation across data sets.

“Project Pulse” simulates the heart beat of the project, The software aggregates multiple sources, interrogates datasets and makes predictions about project success or issues allowing corrective action to be taken, Machine learning models run project simulations to predict likely outcomes and raise alerts for investigation, hyper-personalised for the user’s specific area of responsibility.

Proven results in weeks, not years

Project Pulse was deployed in January 2023 and is being used by a group of project managers and project executives to test the models in a production environment prior to wider deployment.

The data is aggregated into a Microsoft Azure based data lake and correlated across multiple data sources. Machine learning models run across the data nightly generating alerts, scoring project performance and predicting project outcomes.

The user interface utilises generative AI to enable real-time conversation with project data, enabling a user to ask questions such as “How is [Project x] tracking on Zero Harm?” or “Which project should I be most worried about?” Generative models created code in real-time to interrogate the database and retrieve relevant information for the user.