How AI is Shaping Our Office's BIM Adventures

Bulding Information Modeling (BIM) has revolutionized the construction industry, enabling improved project design, management, and collaboration. At our office, we wanted to see if Artificial Intelligence (AI) could do more than just complement BIM—we wanted to see if it could challenge us, surprise us, maybe even outshine us. This article pulls back the curtain on what we discovered: the good, the weird, and the moments that felt like plot points in a Kafka story.

 

Experimenting with AI-Driven Decision Making

Our journey with AI started with a question: How can we use AI to enhance and transform our day-to-day workflows? The answer, as it turns out, involved a lot of demos, a healthy dose of skepticism, and the occasional high-five when things actually worked.

We fed project datasets into GPT Agents and watched as AI offered predictive insights—sometimes eerily accurate ones. There was that one time AI pointed out an error in a product data submittal, something we wouldn't have considered as a reason to reject the product we were reviewing. Or when it pointed out a code issue that we had a different interpretation of. Imagine the look on our faces—“Did it just... solve a problem before we even knew we had one?” Yes. Yes, it did.



Exploring Skema.ai and Hypar for Generative Unit Layouts and Text-to-BIM Building Generation

Recently, we've been diving into Skema.ai and Hypar to see how generative design could revolutionize our workflow. Skema.ai allows us to experiment with generative unit layouts, which we could quickly compare, refine, and (the best part), bake into fully developed Revit models. It wasn't just about design iterations; it was about giving us a new toolkit to explore possibilities that might have taken us weeks to achieve otherwise.

Then there’s Hypar, which lets us explore text-to-BIM building generation. It's still pretty primitive, but there's something undeniably fun about typing out a few requirements and watching it spin up a building concept. Sure, it’s not quite ready for prime time yet, but it feels like a glimpse into the future of architectural ideation. Imagine being able to sketch out ideas in words and get a digital model to play around with in minutes. That's where we think things are headed. Even if it’s just the start, it’s a pretty exciting beginning.

These tools are not just automating what we already do; they're reshaping how we think about design itself. They invite us to play, experiment, and push beyond the conventional limits—whether it's refining unit layouts with Skema.ai or having some fun with early-stage concepts using Hypar.