Ep 38: 3 Ways AI Can TRANSFORM Your Construction Business Today

Join our mailing list & gain more
construction business insights from our experts.

Episode Transcript:

So much is being made of AI in construction. We’re getting a lot of calls from our clients and prospective clients about how can I best incorporate AI into my business. I feel like I’m behind. We want to get into this. We want to be on the front end of it.

There’s a lot to be talked about there, but I thought I wanted to spend just a quick minute talking about this in simple terms to help give us a frame to talk about AI and how you’re going to experience it in the marketplace and how it may plug into your business.

So I’m looking at this in three different ways, and this is kind of in order.

Level one would be AI-enhanced software. Think of this like you have legacy systems, your ERP perhaps, maybe it’s Procore for example, or some other project management solution, where AI is layered into that legacy system in some way.

Think Procore Helix. Think Sage Copilot. Think Copilot inside any one of your ERP systems. Most everyone is rolling that out or has rolled it out in the past six months to a year. They’re taking an existing system and adding it to an existing legacy system.

So what is that good at? It’s good at summarizing things. It’s good at taking the data that’s in your legacy system and flagging things or giving you the ability to assist with workflows inside that, or giving you the ability to create a workflow that wasn’t possible in that legacy system. So it can supercharge a legacy system in a way that wasn’t possible before without just blowing the code up and starting all over.

Think about that as a good starting point. If you’re already using Procore, you have Sage Copilot, you have Procore Helix, you have Autodesk Construction Cloud. These are ways you could get started using AI in a way that you’re probably already paying for. These are quick productivity gains. These are good ways for you to get your feet wet.

You have to remember this is limited to the platform you’re on. It’s not going to go out and do other things. It is only going to use the data inside that legacy platform. But it can be a really good extra boost.

The second way is an integration or an orchestration layer. So that’s a platform that’s going to merge or normalize and then analyze data from two different systems. So legacy systems like your ERP, and you have a project management system over here that either nominally integrates with your ERP now or it doesn’t at all. Perhaps it doesn’t at all. You put a layer over top of those two systems and you put data up to that integration layer to help you look at things in a way that those two systems couldn’t before.

Again, we’re using legacy systems that maybe integrate, maybe they don’t, maybe they integrate but not where you need them to. There are a couple examples out there. Power BI being one that’s sort of AI and has some components. Brick is another app. HH2. Rivet, however you want to pronounce it. All examples of that.

Basically the purpose of that is it’s going to eliminate those silos that you’ve created. You’ve got two different systems. This gives you the ability to bring those two systems together and eliminate those data silos. In the opposite way, if you have an AI-enhanced system and you introduce a third piece of software, you can think about it like middleware. Some people may term it middleware. When you introduce middleware, you are syncing data across multiple systems.

So implementation is much more complex. It will expose poor data discipline. So if you’ve got two systems maybe tracking the same kind of stuff, you better be thinking about that because it introduces more complexity.

The third way, AI-native, has the core engine as artificial intelligence, large language model, some probabilistic outcome, an analytic engine. These are tools that some of which live outside as an integration or orchestration layer, some of them are more cohesive, but they’re going to detect safety violations from pictures, for example. They could analyze the picture and say there are tools out there now where you can wear glasses that are filming as you’re walking and it’s detecting potential safety violations that your human brain did not detect in that moment because you were just turning your head to go over there and in that swipe your eyes were already looking somewhere else so you didn’t trip over something, and the camera picks it up.

That is science fiction to me, and I think that is so cool.

Maybe it compares your BIM model to site pictures. That would be another example. Predicting schedule delays, etc.

These are your ChatGPTs, your Claudes. There are a couple of others out there. Those are the names that people know. These are systems that are very good at pattern recognition. That’s what they’re looking for is pattern recognition, predictive analytics. They’re very automated and also very dependent on the data quality.

So again, I’m going to go back to something I’ve said over and over and over again: if you don’t have good processes and you don’t have good data, any technology is going to be a problem for you.

I hope this helps, just to break this down into a little bit of a chunk so when you call me for a consultation, we’ll have a frame of reference to talk about this.

Best of luck. Thanks.