Control Is a Symptom. Trust Is the Cure.
Hello. My name is Kimon. And I am a control freak.
As an architect, I care about design intent. As a technologist, I care about how information is structured. As a standards advocate, I care about whether systems can actually talk to each other. I hold on to all of it, because if I don’t, who will?
And here is what I have watched the industry do for decades: the same thing, one handover at a time. We reduce risk by containing responsibility. We protect our domain by not crossing into the next one. We don’t share what we know because sharing means owning what happens next. So we let go of the model, of the data, of the context, and we move on to the next project, telling ourselves it’s someone else’s problem now.
It usually becomes nobody’s problem. Which means it becomes everyone’s problem.
Now AI has arrived, and the industry is nervous. We say we don’t trust AI. But I think we’re projecting. We don’t trust what we’ve handed AI to work with: decades of disconnected data, abandoned models, and systems that never learned to talk to each other. AI didn’t create that. It just showed up and asked the question we’ve been avoiding:
Where did all the intelligence go?
Most of us have been operating on the same unspoken assumption:
If everyone would just use our approach, interoperability would happen.
Trust me.
Welcome to Control Freaks Anonymous.
Now it is your turn to say it.
I’m not alone in this room.
“Hi. My Name Is…”
Engineers care about precision and performance. The risk: they optimize inside their domain while the system next door remains a black box.
Builders care about cost, schedule, and risk. The risk: they rebuild information at every handover because they don’t trust what was passed to them, and honestly, often for good reason.
Building automation and controls vendors even have the word “control” in their title, and proudly display it on their products. The risk: connectivity becomes a competitive advantage to protect rather than a foundation to share.
GIS and mapping professionals know exactly where everything is. The risk: location without context, who owns it, what it serves, what happens when it fails, is just a dot on a map.
Facility managers and operations teams inherit the building after everyone else has moved on, and are often still rebuilding the asset list from scratch.
Product suppliers and manufacturers control specifications, product data, and warranties, knowledge that too often disappears the moment it leaves the factory.
Standards organizations and there are many across this industry. Are control freaks by definition. The risk: each one believes its standard should become the center of the ecosystem.
AI companies may be the newest control freaks, promising one platform, one repository, one source of truth. The risk: repeating every mistake we’ve already made, only faster.
And owners? Owners are the ultimate control freaks, and rightly so. It is their money, their buildings, their risk, their future. They are the ones who pay for all of the above when it doesn’t work.
Every one of us has sat in a meeting and thought: if only they understood what we were doing, everything would connect.
History suggests otherwise.
The built environment’s approach to interoperability:
2. It’s all about me and what is mine, mine, mine…
When Nobody Shares, Nobody Trusts
The industry has structured itself around handovers: architect to engineer to builder to commissioning agent to owner to operator. At every transition, trust decreases. The building survives the handover. The intelligence often doesn’t. And so AI hallucinates the answer.
This isn’t accidental. Contracts, professional liability, errors and omissions insurance, all of it reduces risk by containing responsibility. It’s rational. It just happens to destroy the continuity of building intelligence at every handover. We built silos into the legal structure of the industry, and then wondered why the data didn’t flow.
The PAE Living Building in Portland is one of the most advanced buildings in operation, and it had a control problem. (I’ve written about PAE extensively) Eight systems. Disconnected dashboards. No shared identity. The lesson is what PAE did next: they admitted the problem, opened their building and their data to an open coalition, and chose trust over control. Within a year, a disconnected building became AI-ready. That is the model we’re asking everyone to follow.
Owners aren’t asking for more dashboards. They’re asking for trustworthy answers. Trust is not only a cybersecurity problem, but it’s a context problem.
“Mine, mine, mine” is not a trust model.
3. AI Doesn’t Create This Problem. It Demands the Solution.
We don’t distrust buildings. We distrust AI operating inside them. That pressure, owners asking “how do I know this recommendation is right?”, is forcing a conversation the industry has avoided for decades.
At SCaLE 23x in Pasadena, I met Drummond Reed, who was presenting on decentralized trust and digital identity. He wasn’t talking about buildings. But everything he described, context, relationships, verifiable identity, governance, I kept recognizing as building problems. The vocabulary was different. The challenge was identical.
Drummond joins Hart Montgomery of the Linux Foundation, Tristan de Frondeveille of SkyCentrics, and me this week at the USGBC session Ensuring Trust in AI for Building Operations (June 17) to explore exactly that overlap. Two domains, different languages, same problem. That is what this moment feels like everywhere I look.
4. From Mine to Ours
For decades the design world and the operations world talked past each other, each assuming the other should adapt. The BIM side starts from spatial context. The controls side starts from what the building is doing right now. Classic control freaks on both sides, each with their own fish.
Then something shifted. At the buildingSMART International Summit in Porto this year, the smart building working group said it plainly: a detailed design model is not the operational endgame. Owners need identity, relationships, and context that survive beyond handover, exactly what the operations world has been saying from the other direction.
Neither side had to win. They found each other in the middle. No single center. Just connection.
That is what the diagram above shows. Not a hub and spoke. A network. Every domain a node. Every connection a choice to trust.
5. We Are All Architects of Intelligence
Letting go of “mine” is not the same as losing it.
When silos connect, each domain becomes more valuable. Your intelligence reaches the systems that depend on it. Your data means something outside your own walls. You stay in your domain, you just open a door.
We are Architects of Intelligence, not because we all design buildings, but because we all shape the intelligence buildings depend on. In the AI era, that intelligence needs to flow across every domain we’ve been protecting.
Owners have the leverage to move this. Every contract and procurement is a chance to ask the question control freaks rarely ask each other: Can your system connect to theirs?
The Internet didn’t succeed because one protocol won. It succeeded because each part stayed open enough to connect. Buildings need the same model.
Nobody has to lose. Nobody has to win. Everybody has to connect.
That is the peace treaty. Come sign it.
This is not a position paper. The Coalition for Smarter Buildings is doing this work right now, under the Linux Foundation, with real buildings and real data. PAE is the test case. cloudBIM.org and C4SB.org are where the work lives.
Who’s in?
