It should be no surprise that prominent research has found that individuals’ decision-making is shaped by cognitive biases. When we think about decision-making on a construction project – where margins are slim and both schedules and budgets can slip – is ‘good enough’ really good enough? Or can we adopt artificial intelligence (AI) and machine learning (ML) to help us make better decisions and remove bias as much as possible?
Decisions are made every day on a construction project. There are multiple activities taking place at any one time, and strategic decisions can impact any of these. But often when confronted by a decision, we delve into our experiences to identify the best course of action, backed by some historic projects data.
But what if we could use data from several sources as both a guide to decision-making and also to make recommendations using AI/ML? This ensures we are considering several options and evaluating the best course of action, not just the one that we are most comfortable with.
How AI assists our decision-making
The opportunity to leverage all data to make a tangible difference to project delivery is huge, particularly given the disparate nature of construction ecosystems. But as an industry, we need to move towards accepting predictive analytics and the concept of ‘probability’ as opposed to ‘certainty’ in our approach to using data for decision-making. You can be certain of what has happened but not what might happen.
We already trust probability in other aspects of life by using video streaming recommendations or weather predictions, but the idea of using predictive analytics within construction and engineering can be game-changing. In some situations, we even discount what the data tells us because as humans, we allow emotion into our decision-making.
We need to accept that technology is smart, unbiased and doesn’t add emotion to the result it provides. It’s the difference between a decision made on all data points available versus a decision made on fewer data points and with emotion and/or bias thrown in.
Technology has matured to the point where the industry needs to take predictive analytics and probability seriously. Concepts like generative designs are getting progressively embraced, but the notion of using AI and ML for project and construction management is in its nascent phase.
At Oracle, we are looking to address these challenges by using data to tackle questions like: how can we predict project delays as soon as a plan is created? How can we mitigate unseen risks before they reach the risk register? How do we reduce rework and defects? How can we prevent health and safety incidents?
The idea here is that you can truly transform the business by complementing lag indicators (looking at historical project information) with lead indicators (looking ahead). Doing this will help the industry deliver safer, faster and better projects that can significantly improve productivity. The transformation might not be overnight, but the oversight gained can be transformational.