Driving real value with AI – combine process and behavior data

There’s an unfortunate tendency in software today. It seems that every software provider is selling their solution plus a generous side dish of Artificial Intelligence. This situation leads those of us who consume software to view these claims with skepticism. Is AI core to one of the software functions? Does it deliver real value to someone in the value chain? Can that value be measured in affirmative terms (not just time savings). In other words, did someone slap a ChatGPT API into a text editor and call it AI?

Before I go much further, I want to affirm that I am a fan of machine learning, AI, and the possibilities that both hold for businesses. In construction, I’m seeing real AI solutions for pre-construction, modelling, course of construction, and handover/operations. There are demonstrated results beyond a sales pitch.

What’s the key to overcoming the hype and seeing the value? The data must make a connection. In my opinion, the most powerful connection is between what we have done, what we intend to do, and where we are going next. Let me use the example of safety. We know what our intentions, our practices are. They are well documented and are foundational to how building gets done. We have data on what we have done in the past – where we have seen near misses or incidents - and we have plans on what we are going to do next – which subcontractors or areas of work are coming next. What AI can do is mine both the historical and the prospective. It can provide analysis of similar situations with the same or similar conditions and workforces and extrapolate that for the near- and middle-terms.

The AI acts as a lantern, reflecting our previous behaviors on the path ahead. Showing us where we have tripped before so we can avoid those situations again. While we have certainly taken a similar approach before, AI tools can analyze data with a thoroughness and speed that will deliver much better results. This lantern is the real value of AI – reducing rework, eliminating time overruns, and improving safety and quality.