Is BIM dead? Will AI save the construction sector?

I’ve been in this industry long enough to see hype cycles come and go. I remember the days when BIM was pitched as the silver bullet for every problem in construction. The promise was huge, the excitement was electric.

Now, years later, the initial shine has worn off. I often hear that less BIM model files are shared in construction projects. We’ve all been in those projects where BIM didn't deliver on its grand promises,, and the sector is trying to find a more realistic, practical place for it.

And right on cue, a new hero has entered the ring: AI. The buzz feels incredibly familiar. I hear the same chatter in meetings and see it online: "AI will finally fix things! It can even analyse our 2D drawings, so maybe we don't need that complex BIM stuff after all."

Frankly, I think that's a dangerous and misguided way to look at it.

Let's be honest: Tech can't fix trust

Here's a lesson I've learned the hard way: our biggest problems aren't about technology, they're about people. The real challenge in construction is collaboration. It's getting everyone to share information openly and on time.

I'm going to repeat again something said thousands of times: BIM is more of a process than a piece of software. It’s a framework for collaboration, a way to build a single source of truth. And that process absolutely includes the 2D drawings we all rely on.

Trying to "save" the industry by applying AI to a pile of disconnected 2D drawings is like putting a brand-new engine in a car with no wheels. It looks powerful, but it's not going anywhere. If the underlying process is broken, if stakeholders aren't willing to share information, no algorithm on earth can fill in those gaps.

Let me put it in another way: AI can't invent data that your partners are hiding. It cannot magically create a collaborative culture.

The future I'm excited about

This isn't an "either/or" situation. For me, the conversation shouldn't be about AI versus BIM. The real magic will happen when we use both together.

This is the future that gets me excited, especially for complex tasks like quantity takeoff. Think about what becomes possible:

• Make data richer for everyone: AI can make it seamless for every stakeholder to add their own data to the model. This means our BIM models become exponentially more customised and useful for specific use cases. The designers cannot anticipate the needs of all project stakeholders who sometimes arrive later in the project.
• Ask your model anything: Imagine interrogating a model in a flexible, natural way. Instead of clicking through endless menus, you could apply complex rules of measurements in plain text: "Formwork to vertical sides of cast-in-place concrete walls, measured as the net surface area in contact with the concrete. No deduction is made for openings with an area of 1.00 m or less. For openings greater than 1.00 m, the area of the opening shall be deducted, and the formwork to the resulting reveals shall be measured." That’s a level of intuitive control we just don't have now.
• Automate the tedious work: We can finally automate highly repetitive tasks without relying on hard-coded rules that fail at the first opportunity. A perfect example is generating a bill of quantities, where AI can apply complex descriptions and rules of measurement multiple times, freeing up skilled professionals for more strategic work such as cost analysis and quality assurance.

To do any of this, AI needs one thing: good, structured, reliable data. And that is exactly what a well-executed BIM process provides. BIM is the fuel, and AI is the engine. You can't have one without the other.

So, is BIM dead? Not a chance. But that's just my take.