At Realcomm IBcon this year, a case study session walked through something the industry has been promising for a decade: AI moving beyond isolated workflows toward autonomous building operations, with agentic workflows coordinating tasks across a portfolio in real time. Tenant requests, facilities work, and tenant services, all talking to each other. It is a genuinely impressive shift, and it is finally leaving the demo stage.
So here is the question an asset manager should ask before getting excited: coordinated by what, running on whose data, and under whose permissions?
That question is not skepticism for its own sake. It is the difference between an autonomous portfolio you control and an autonomous portfolio that quietly hardens into a set of vendor dependencies you cannot unwind. The technology is arriving faster than the architecture discipline to hold it. Let me demystify what is actually happening under the hood, and why the part that matters most is the part nobody is demoing.
What "autonomous" actually requires
An autonomous building operation is not one clever application. It is a coordination problem. To decide anything on its own, an AI agent has to read the state of the building across systems that were never designed to talk to each other: the access control platform, the HVAC and building automation system, the metering and submetering feeds, the work order and maintenance system, the space utilization sensors, the tenant request app, and the lease and occupancy records.
Each of those systems was sold separately, installed at a different time, and governed by a different vendor contract. Each one holds its slice of the truth in its own format, behind its own login, in its own cloud. For a single workflow, that is annoying but survivable. For autonomous coordination across the whole property, it is the entire game. The agent is only as capable as its ability to read and act across all of it under a consistent set of rules.
This is why the market is now shifting from workflow tools to integration architecture. The value is no longer in the individual app. It is in the layer that lets decision engines coordinate across systems. And that layer is exactly where ownership gets decided, usually by accident.
The integration bill nobody prices in
When every vendor brings its own data model, its own permission scheme, its own automation logic, and its own cloud dependency, autonomy does not eliminate complexity. It relocates it. You are no longer managing five dashboards. You are managing five systems that now need to make coordinated decisions, and the coordination logic lives wherever the strongest vendor decides to put it.
Watch what is happening at the platform level and the pattern becomes obvious. Siemens acquired Brightly Software to serve as the foundation for asset health inside its Building X platform, consolidating maintenance and asset data into a single vendor's stack. In the same window, Dormakaba acquired the access control hardware platform Azure Access Technology, and MaintainX kept expanding its maintenance footprint. None of this is bad technology. It is a market consolidating the coordination layer into vendor-owned platforms, one acquisition at a time.
The risk for an owner is structural, and it does not show up on the P&L until it matters most. When your building's coordination logic lives inside a vendor platform, three things happen. Your operating history accumulates in a format you do not control. Your automation rules become a proprietary asset the vendor owns. And your ability to change vendors gets priced against the cost of rebuilding the entire coordination layer from scratch.
That is the silent version of vendor lock-in. It does not announce itself. It compounds quietly, one integration at a time, until a diligence team finds it during a refinancing or a sale and the price moves before you can respond.
The AI part is converging, so the moat moved
Here is the piece that should reframe the whole conversation for an asset manager. The intelligence itself is becoming a commodity.
The performance gap between open models and frontier models has collapsed, and the cost to reach a given benchmark is falling several times over year after year. Gartner's predictions for the back half of the decade describe physical AI, agent swarms, and automation spreading into ordinary operations. When the models converge, choosing a model becomes procurement, not strategy. You will swap decision engines the way you swap a phone carrier.
The firms building these tools understand this. Green Street now offers a server that connects its intelligence directly into Claude, ChatGPT, Gemini, and other AI tools a team already uses, no new platform and no new login. The message is explicit: the model layer is interchangeable, so meet the customer inside whatever tool they already run.
Apply that same logic to your building. If the decision engine is replaceable, then the durable advantage is not which AI you picked this year. It is the governed foundation the AI reads from and acts upon. Four things determine whether one portfolio outperforms another, and only the owner can build them: proprietary operating data, coded operating workflows, an orchestration layer, and institutional knowledge encoded into systems. The model is the commodity. The layer above it is the moat.
This is the reframing every owner should internalize before signing the next autonomous operations contract. If you don't own your data & digital infrastructure, your vendors do. And in an autonomous building, owning the intelligence layer is the only thing that keeps the intelligence working for you.
Building the owner-controlled orchestration layer
The answer is not to avoid autonomous operations. The answer is to build the foundation that lets you adopt them without turning every property into a custom integration project you can never leave. At OpticWise, that is the whole point of the two-layer model, and the path runs through the PPP 5C™ plan from Peak Property Performance®.
Clarify. Before connecting anything, a PPP Audit™ review maps who currently owns each data stream, where it lives, and what is actually portable. In most properties, this is the first time anyone has documented which vendor controls which slice of the operating record. You cannot orchestrate what you cannot see, and you cannot negotiate a position of strength you have never measured.
Connect. Establish secure, owner-controlled connectivity that repeats property to property, built on a single segmented foundation through BoT® (Building of Things®). One governed network, not a stack of vendor-specific tunnels, each with its own cloud dependency.
Collect. Capture and normalize the building's operating data into a consistent model the owner holds. When your metering, access, maintenance, and occupancy data land in a format you own and reuse, your operating history stops being a vendor asset and becomes yours.
Coordinate. This is the orchestration layer. Property Brain™ governs identity, access, privacy, data lineage, retention, and the rules of use across every connected system. It is the trust plane that decides what any agent is allowed to read and do, under owner permissions, regardless of which vendor supplied the system.
Control. Now the decision engines plug in. Any vendor platform, any internal analytics, any large language model can act on the governed foundation, under your rules. Swap a vendor without losing your history. Swap a decision engine without rewiring the building. Standardize it once and Property Brain™ becomes Portfolio Brain™, so the intelligence compounds across the portfolio instead of restarting at every address.
That is the architecture difference. In the vendor-owned pattern, autonomy makes you more dependent with every system you add. In the owner-controlled pattern, autonomy makes you more capable, because the foundation is yours and the decision engines are interchangeable parts sitting on top of it.
What this protects, in asset-manager terms
Run the math the way the investment committee will. The orchestration layer is not an IT expense. It is digital CapEx with a documented return, because it protects the two things that set valuation: NOI and the credibility of your operating record.
On the NOI side, coordinated operations reduce OpEx variance and surface the causal drivers behind energy, maintenance, and utilization decisions. Every dollar of recoverable NOI capitalizes into roughly fifteen to twenty-five dollars of asset value at typical cap rates, so operating discipline is not a rounding error. It is the value-creation lever.
On the diligence side, owner-controlled operating data is the difference between a clean refinancing package and a discovery that moves the price. When a lender or buyer asks for the operating history and you can produce a governed, portable record instead of a scramble across five vendor logins, you are negotiating from strength. Vendor-controlled data is the silent NOI tax that stays invisible until diligence finds it. By then, the terms have already moved against you.
Autonomous building operations are coming, and they will deliver real value. The only question is who owns the layer that makes them work. Build that layer yourself, on data and a foundation you control, and every future decision engine becomes an upgrade instead of a dependency.
Start with a PPP Audit™ review of one property. Map what you own, prove the coordination layer is portable, then scale the standard across the portfolio.
Own your data & digital infrastructure. Build for the long game.
