Every day, across the world, across countless projects, AEC professionals in the built environment open up BIM tools like Revit, Archicad, or Tekla to model buildings — placing walls, tagging doors, generating schedules, and exporting drawings. These are familiar actions. They’re the lifeblood of digital delivery as it occurs today, but for the most part, the end deliverables are still quite traditional – models, drawings, schedules – things we in AEC have understood for centuries.
But beneath these activities lies a deeper shift — one that’s quietly reshaping how we understand, exchange, and act upon information in the digital age. And that is the exchange of digital data.
That shift is toward meaning — meaning that can be understood by both humans and machines.
Not just data. Not just models. But structured, shareable meaning — the kind machines can read, humans can trust, and whole systems can reason with and rely on.
To work effectively in a world of digital twins, AI agents, and smart buildings, we need to get better at making our information not only machine-readable and human-understandable, but able to seamlessly exchange across the web, between multiple systems and agents. And to do that, we need to become fluent in three key concepts:
– Semantics
– Taxonomy
– Ontology
🧩 Semantics: What Does It Mean?
Let’s start with semantics, which is simply the study of meaning.
When you label something “Wall-001” in a BIM model, what do you really mean by “wall”? Is it load-bearing? Is it a partition? What’s it made of? Where is it used?
Humans can often intuitively guess the context. But machines don’t.
Semantics ensures that the terms we use are clear, consistent, and interpretable — not just by us, but by every other system and stakeholder who touches the data. Whether that’s a cost estimator, a facility manager, or an AI tool parsing your data for another purpose.
Without semantics, we may all be using the same words — but meaning different things. Like people speaking a different language to each other – a lot of noise, but not much actionable understanding.
🗂️ Taxonomy: Where Does It Belong?
If semantics is about meaning, taxonomy is about classification — a structured way of organizing complex things into simpler hierarchical structures, so we can sort, search, and retrieve them effectively.
In construction, that might look like:
“Wall” → “Vertical Element” → “Building Element” → “Constructed Asset”
Taxonomies help software systems — and people — make sense of complex datasets. They’re like the shelving system in a library or the folder structure on your computer.
Classification systems like Uniclass, Omniclass, and Masterformat provide these taxonomies for our industry. But many BIM users only engage with them superficially — or sometimes not at all — relying on default tags or generic object types that lack clarity and consistency.
🧭 Ontology: How Is It Connected?
Then there’s ontology — the most powerful and perhaps least understood of the three.
Ontologies define how things relate to each other. They don’t just say what something is or where it fits — they show how it connects to everything else – like an interconnected web, or graph.
For example:
A thermostat controls the temperature in a room.
A room is part of a floor, which belongs to a building.
A duct serves a zone, and a sensor monitors CO₂ levels.
Ontology is what transforms a list of data points into a web of meaningful relationships — a semantic graph that both humans and machines can follow, query, and reason with.
🏗️ BIM in Practice: You’re Already Doing This (Almost)
Here’s the good news: if you use BIM tools, you’re already engaging with semantics, taxonomies, and ontologies — even if you’re not aware or calling them that. They are built into the systems.
The object types in your software? That’s taxonomy.
The names and parameters you assign? That’s semantics.
The spatial hierarchy or system connections? That’s ontology.
And when you export to IFC, you’re attempting to package all of that — structured data, geometry, metadata, and relationships — into a model that can travel between tools and stakeholders.
The challenge? Most of us are doing this implicitly, not deliberately. We simply rely on the export function of the tools we use, without really understanding what is going on under the hood. And as a result, much of our data lacks clarity, consistency, and interoperability once it leaves our immediate environment. It is like we have exported communication in our own bespoke language and dialect, and we expect everyone else to understand it.
⚔️ The Battle of Ontologies — and Why AI Might Be the Peace Treaty
Over the past decades, many in the built environment have tried to solve interoperability by creating the perfect structure — a universal ontology everyone should adopt.
We’ve seen major efforts like:
IFC (open BIM and geometry-based models)
Project Haystack (tagging and controls)
Brick Schema (semantic graph for building systems)
RealEstateCore, ASHRAE 223P, and others
Each of these brings real value to their specific domain. But together, they’ve created a landscape of many overlapping languages — similar in spirit, but structurally divergent. The result? Confusion, duplication, and high integration costs.
This has become what some call the “Esperanto problem”:
Just like Esperanto tried to be a universal human language but never gained mass adoption, many of these ontologies tried to standardize everything — and struggled to get everyone to agree.
🤖 From Esperanto to Google Translate
But now something new is changing the game: AI.
Rather than forcing everyone to use the same ontology, AI models (like GPT and others) can now interpret, translate, and reconcile differences in terminology and structure — much like Google Translate does for human languages.
This changes the conversation:
We no longer need to force everyone into one system.
We just need to make sure our systems are clear, consistent, and understandable. (translatable)
AI can do the heavy lifting of connecting the dots between them.
This doesn’t eliminate the need for standards, semantics, ontologies, or taxonomies — but it allows for less rigid and brittle adherence.
In this new paradigm:
Semantics are still essential — because even AI needs to know what we mean.
Taxonomies remain useful — they help systems sort and organize information.
Ontologies continue to matter — but we no longer have to choose just one. We can map, align, and relate them dynamically.
This shift allows us to move from rigid standardization to more flexible interoperability — from “everyone must conform” to “everyone must be coherent.” (it is the shift from herd mentality to swarm intelligence).
🤖 Why This Matters More Than Ever
We’re entering a world of smart buildings, digital twins, autonomous controls, and AI-driven decision-making. In that world:
Machines need to understand the meaning and structure of data.
Owners and operators need lifecycle information that survives past handover — and across multiple assets.
Cross-industry integration (BIM + BMS + GIS + ESG) is becoming the norm, not the exception.
If we’re still handing over “Wall-001” in a PDF drawing, with a PDF schedule — we’re falling behind, even if we have used sophisticated BIM software to deliver that PDF.
🛠️ What Can We Do Now
If you’re involved in BIM, digital construction, or facilities information, here’s how you can start leaning into this future:
Name things with intention — don’t just tag for compliance, tag for clarity.
Classify your data using available standards — even a basic taxonomy is better than none.
Understand your relationships — spatial, systemic, or temporal — and model them when possible.
Be open to hybrid models — BIM is becoming part of a larger semantic ecosystem. You don’t have to pick sides; you just need to be understood.
🧠 Final Thought: From Drawings to Dialogue
Remember, with BIM, you’re not just drawing. You’re not just modelling.
You’re building a language of the built environment — one that must speak across time, teams, tools, and technologies. Exchange of Information is the essence of communication.
We’ve done the hard work of going digital. Now it’s time to go semantic.
Because in the end, the most valuable model isn’t the one that just looks pretty —
It’s the one that makes sense, to both humans and machines.
Reference: This article originally appeared on BIM Heroes Community – a global platform for AEC professionals interested in digital transformation. See Here.
👉 You can Join BIM Heroes Community Here
🔗 For those interested in how these ideas are being advanced across the industry, explore the work of the Semantic Tiger Team under the Linux Foundation. Read more here from our partners at www.AutomatedBuildings.com https://www.automatedbuildings.com/2025/04/c4sb-hosts-the-semantic-tiger-team-kickoff/
