LEED v5 Requires Living Data from Buildings: Follow the Yellow Block Road
LEED (Leadership in Energy and Environmental Design) is the world’s most widely used green building certification system, developed by the U.S. Green Building Council.
At the recent Greenbuild 2025 LA conference, I spent a full day in AI workshops focused on the future of intelligence in the building industry. The work was thoughtful, technical, and genuinely impressive.
Then came a striking contradiction. A member of the AIA Large Firm Roundtable explained why their firm does not share detailed building data, citing concerns about AI misuse and IP theft, just after presentations showing how AI can extract structured data from 100-year-old scanned documents.
This forces an unavoidable question, and it starts with what I mean by flattening.
Flattening is what happens when rich, structured building intelligence, BIM, system relationships, asset data, and operational context are reduced to static documents: PDFs, drawings, disconnected models, and files optimized for viewing rather than use. Once flattened, that intelligence can no longer participate directly in operations, analytics, or verification. It has to be reconstructed.
So the real question:
Why do we flatten building intelligence into documents, only to spend enormous human, computational, and energy resources reconstructing what already existed?
The Hidden Cost of Flattening Data
The reality: not only are architectural drawings getting flattened, but so are mechanical and electrical data, control systems information, and everything that makes a building smart. These systems are the lifeblood of a building’s operations, and yet we’re still locking them away in formats that no one can use without wasting human and AI energy..
Consider the language we use:
- The industry’s core deliverables from design and construction are called “AIA Contract Documents.” Those documents are the default deliverables clients pay for and expect.
- And the file format used everywhere: PDF, literally stands for Portable Document Format.
Both terms tell the same story: what we treat as deliverables are representations of information, not living, queryable data. Formats optimized for viewing, printing, and exchange are fundamentally different from deliverables designed for machine use.
It’s worth acknowledging why PDF became so important to our industry. In the late 1990s and early 2000s, PDF solved a real problem: reliable printing, broad accessibility, and consistent distribution of technical documentation at a time when web standards were immature. For manuals, specifications, and human-readable reference material, PDF has proven incredibly durable and valuable.
The issue isn’t PDF itself. The issue is using document formats as a substitute for machine-readable building data, especially when that data already exists upstream in BIM and system models.
- The same mindset shows up in BIM and CAD workflows. We still talk about file exchanges, model handover, deliveries by phase, and files of record, as if intelligence lives inside files rather than in connected data. When BIM is treated as something to “export,” “submit,” or “archive,” it becomes just another document, regardless of how 3D it looks.
That language matters. It shapes behavior. And it quietly reinforces a file-centric workflow that flattens intelligence long before operations ever begin.
When PDFs, disconnected BIMs, and documents are fed into large models so that owners can recover information that has already been created and converted into data, it is not innovation. It is a waste.
LEED v5 Changes the Conversation
LEED v5 marks an important shift. LEED (Leadership in Energy and Environmental Design) is the world’s most widely used green building certification system, and its next iteration moves beyond documented intent toward verified performance and operational data. Sustainability is moving beyond documents and checklists toward measured, verifiable performance, which will require more than PDFs and promises of intent. It will need data, including operational data over time.
In simple terms, the building has to talk, not about what it was designed to do, but about what it is actually doing.
Contracts, tools, and roles have not caught up. But this year feels like a marker, the point where a document-centric approach no longer aligns with a performance-based future.
Reverse-Engineering BIM with AI is Not Sustainable
Using AI to extract data from PDFs and other documents for buildings that were already designed in BIM is an enormous waste of resources. It consumes data-center capacity, GPU cycles, electricity, and cooling to reconstruct information that already existed upstream in structured form.
In effect, we are spending energy to guess what was already known.
If AI has to “read” PDFs to understand a digitally modeled building, something is fundamentally broken.
AI Is Learning Anyway, So Why Fight It?
Here’s the irony: AI is learning anyway. It’s learning from the construction process itself, recognizing objects and relationships as buildings go up. Even if information is hidden in PDFs, AI can reverse-engineer it from the ground up. So why keep this data locked?
Contractors are reverse-engineering building plans.
AI, drones, LiDAR, and robots are capturing reality to optimize construction and train future workflows. The building is becoming the data source, but ownership of that intelligence remains undefined.
Deliver Living Data
Instead of flattening our intelligence into documents, let’s deliver living data. Whether it’s BIM, digital twins, or complete lifecycle information, owners want data that actually works. That is the necessary shift to seriously tackle sustainability and accurate building intelligence.
The World is Not Flat: How We’re Flattening Ourselves Out of the Conversation
The building industry is one of the largest sectors on Earth, yet we’re “flat” in more ways than one. We love to showcase what I call “Hollywood BIM”, beautiful 3D models and digital twins that look impressive on the surface but often lack the depth of data required for real operational use. In practice, many of these models are just as flat as a PDF when it comes to lifecycle intelligence.
What makes the PAE Living Building different is not that it exposed this gap, but that PAE chose to confront it. As engineer, owner, and occupant of their own building, PAE recognized that even a world-class, high-performance project could carry forward hidden data loss at handover. Rather than ignoring the issue (as most projects do), they treated it as an opportunity to learn, share, and lead.
By opening their building, their BIM, and their operational reality to the Coalition for Smarter Buildings and the Linux Foundation ecosystem, PAE made visible a challenge that many architects and engineers never encounter, because building data is typically collapsed into documents and handed off at project close, then quietly disappears from view. As engineer, owner, and occupant of their own building, PAE had a rare opportunity to see this gap firsthand and chose to address it rather than ignore it. In doing so, PAE didn’t expose a weakness; they demonstrated what leadership looks like when the industry decides to move beyond the flat world.
Viva Las Vegas 2026. The show goes on. Some stayed flat. Others followed the Yellow Block Road.
The Value of Information
This brings us to a bigger question: how do owners value information? Currently, they pay teams to collect existing data because the information was delivered in a flat document. That is how architects and engineers delivered. Sustainable goals are more than designing with eco-friendly materials. As stewards of the built environment, it is incumbent upon us to incorporate sustainable design processes that deliver value, data-rich information (the “I” in BIM).
Join Us at AHR: we’re showcasing how the Coalition for Smarter Buildings is delivering a model where information stays live and connected from the start. The world is not flat, and neither is the data that shapes our buildings. We’re making sure critical building information, mechanical, electrical, and everything beyond, doesn’t get flattened into oblivion. Take a look at our previous article, and join us in protecting the depth, context, and intelligence that will carry the built environment into the future. The world isn’t flat, and for those ready to move forward, the Yellow Block Road is already there.
Note:
Recent threads on the AIA Community Hub and LinkedIn about AI and Flat Files.
The Use of AI for Imagery in This Post
A fair question is whether using AI to illustrate this argument contradicts the point being made. It doesn’t. There is a fundamental difference between using AI once to communicate ideas and using AI continuously to reconstruct information that was unnecessarily flattened. The former replaces manual effort. The latter exists only because the structure was destroyed upstream.
