AI Risk Demands Owner Data Control

AI risk is no longer a technology issue. It is an ownership issue.

For years, commercial real estate owners treated software choices as vendor-side decisions. Pricing tools lived with revenue teams. Listing platforms lived with marketing. Accounts payable automation lived with accounting. AI pilots lived wherever someone had the budget and appetite to try something new.

That model is breaking.

The RealPage fallout, the CoStar and Zillow listing data fight, and the continued limits of AP automation all point to the same conclusion: legal exposure, decision quality, and asset value follow the data. Not the demo. Not the dashboard. Not the vendor’s sales deck.

The owner who controls the data controls the strategy. The owner who does not control the data inherits the risk.

If you don't own your data & digital infrastructure, your vendors do.

That line used to sound like a warning about vendor lock-in. Today it is bigger than that. It is a warning about capital risk, compliance risk, valuation risk, and the future of AI in real estate.

The AI Question Owners Are Asking Too Narrowly

Most owners are still asking the wrong AI question.

They ask: Which tool should we buy?

The better question is: Which data will this tool use, who controls it, how is it governed, and can we move it when the business requires it?

That distinction matters because AI does not create judgment from nothing. It works from data, context, permissions, history, and rules. If those inputs sit inside vendor-controlled systems, the owner may get outputs without gaining control.

That is not a durable capability. That is dependency with a better interface.

The broader business world is waking up to the execution gap. In an AI transformation discussion, Laurent Cochet’s interview with Keith Carter frames the challenge plainly: many leaders are not struggling with access to AI tools, they are struggling with execution and adoption. I see the same issue in CRE.

The tool is rarely the real bottleneck. The bottleneck is whether the owner has a governed operating foundation that makes the tool useful, safe, and portable.

Without that foundation, AI becomes another layer of abstraction between the owner and the asset. You see an answer, but you cannot always see the assumptions. You get a recommendation, but you cannot always test the inputs. You automate a process, but you may not know whether the workflow reflects the owner’s risk appetite.

That is not acceptable when the decision affects rents, expenses, tenant relationships, compliance, debt service, and exit value.

Legal Exposure Follows the Data Trail

The RealPage controversy changed the tone of the conversation. Whether an owner uses pricing software directly or relies on an operating partner that does, the strategic lesson is hard to miss: data-sharing, algorithmic recommendations, and market behavior can create exposure that sits much closer to ownership than many expected.

A recent Propmodo signal framed it in owner terms: legal liability follows the data, not the vendor. That is the point owners cannot delegate away.

A vendor can provide a tool. A vendor can process information. A vendor can recommend an action. But the owner still needs to understand what data is being shared, what decisions are being influenced, what records can be produced, and what governance exists around the process.

That is not a legal lecture. It is an asset management issue.

If a regulator, lender, buyer, investor, or board asks how a decision was made, “the vendor handled it” is not a strategy. It is a weak position.

The same applies to AI-generated workflows inside leasing, operations, finance, procurement, and capital planning. If the owner cannot explain the source data, permissions, decision rules, and record trail, then the owner does not have operating intelligence. The owner has exposure.

I have sat in the owner’s chair. I know how this shows up. It rarely announces itself as a technology problem. It shows up in diligence. It shows up in investor questions. It shows up when a vendor contract ends and the data does not move cleanly. It shows up when a new platform promises better answers but cannot read the building’s operating history.

By then, the price has already moved.

Listing Data Is Not Just Marketing Data

The listing data fight between CoStar and Zillow is not just a portal dispute. It is a signal about control over one of the most valuable data categories in real estate: market visibility.

Propmodo described the Zillow case as one that could decide who controls real estate’s listing data. Owners should pay attention because listing data does not stop at marketing. It shapes pricing, availability, demand signals, broker behavior, leasing velocity, and valuation narratives.

For an asset manager, that means listing data affects the story you tell in the board deck. It affects assumptions in the refinance package. It affects how you interpret rent growth, downtime, concessions, and competitive position.

When that data is controlled by outside platforms, owners risk building strategy on someone else’s terms.

I am not arguing that owners should stop using listing platforms. That would be unrealistic and unhelpful. These platforms have real market function. The issue is whether the owner has a parallel owner-controlled record of its own assets, availability, pricing history, tenant interactions, performance context, and decision history.

That owner-controlled record becomes the difference between renting access and building memory.

Without it, every new platform becomes another partial view. Every market cycle forces the owner to reassemble the same facts. Every portfolio review depends on exports, manual reconciliations, and assumptions that may not hold up under pressure.

With it, the owner has a portable data foundation that can support pricing analysis, leasing strategy, AI-driven recommendations, lender reporting, and buyer diligence without surrendering control.

That is where asset value starts to separate.

AP Automation Shows Why Judgment Still Matters

Accounts payable automation is a useful example because it looks simple from the outside.

Read the invoice. Extract the vendor, amount, date, property, and GL code. Route it for approval. Done.

But CRE does not work that cleanly.

Propmodo recently highlighted a $4,200 landscaping invoice as an example of what AP automation keeps getting wrong. The issue was not extraction. The issue was judgment. Coding decisions require context.

Was the work part of routine maintenance or a capital improvement? Was it recoverable under a lease? Was it tied to storm damage, tenant request, seasonal contract scope, or a one-time exception? Should it hit one property, multiple properties, a tenant account, an insurance file, or a CapEx project?

AI can assist with those decisions, but only if the owner has the data model, workflow history, lease context, vendor history, approval rules, and governance structure to support the answer.

Otherwise, automation makes the wrong decision faster.

For an asset manager, that is not a back-office inconvenience. It is OpEx variance. It is recoverability leakage. It is noisy financial reporting. It is weaker DSCR visibility. It is less confidence in the numbers that drive valuation.

This is why I push owners to stop thinking about AI as a set of apps and start thinking about AI readiness as an ownership discipline. The model matters. The workflow matters more. The owner-controlled data foundation matters most.

The Owner Moat Is Governed Portable Data

In CRE, the owner moat is not a proprietary chatbot. It is not a nicer dashboard. It is not another vendor promise that this platform will become the one place everyone works.

The owner moat is governed, portable, owner-controlled data.

Governed means the owner knows what the data is, where it came from, who can use it, what it can be used for, how long it is retained, and how decisions are recorded.

Portable means the owner can move data across vendors, decision engines, properties, and portfolio strategies without rebuilding from scratch every time.

Owner-controlled means the data compounds inside the owner’s business, not only inside the vendor’s platform.

That is the strategic shift. AI raises the value of data, but it also raises the cost of losing control of it.

Commercial Observer’s coverage of large owners and construction technology firms leaning into AI shows how quickly major CRE players are moving from experimentation toward adoption, with owners and major platforms putting AI into core workflows. That direction is not surprising. AI will become normal operating machinery across the industry.

But normal does not mean safe. Normal does not mean owner-controlled. Normal does not mean investable.

The owners who win will not be the ones with the most disconnected AI pilots. They will be the ones with the clearest data governance, the most portable operating memory, and the strongest ability to connect AI outputs to capital decisions.

That is how you protect NOI. That is how you improve refi readiness. That is how you reduce counterparty risk. That is how you make each property smarter without making the portfolio more fragile.

The OpticWise Read

At OpticWise, we start with the asset owner’s problem: you are accountable for performance, but too much of the data that explains performance sits outside your control.

That is why our work is built around Peak Property Performance® and the PPP 5C™ plan: Clarify, Connect, Collect, Coordinate, Control.

Clarify means you start with a practical review of the current state. What data exists? Who owns it? Which systems are trusted? Where is decision context missing? Which vendor contracts limit portability? Where does AI introduce risk instead of confidence?

Connect means building the owned network layer through managed data & digital infrastructure. This is where SIC® creates the operating standard for Security, data & digital infrastructure, and Connectivity. It includes the BoT® or Building of Things® layer, ElasticISP® for owner-controlled managed connectivity, and the 5S® user experience promise: Seamless Mobility, Security, Stability, Speed, and Service.

Collect means aggregating the operational data that matters into an owner-controlled data foundation. Not every data point deserves the same priority. The asset manager should care about the data that supports NOI, risk, compliance, tenant experience, capital planning, and valuation.

Coordinate means governing how systems, vendors, workflows, and permissions interact. This is where AI becomes safer because the owner sets rules of use instead of accepting whatever each vendor allows.

Control means delivering owner-controlled intelligence through Property Brain™ at the asset level and Portfolio Brain™ across the portfolio. The goal is not another dashboard. The goal is a governed intelligence layer that can support decisions, swap tools when needed, and preserve institutional knowledge.

PPP Audit™ is a named product in the Peak Property Performance® canon, but the first move for an owner does not need to be complicated. Start with a focused review of the places where AI and vendor platforms already touch your data.

Ask five questions.

What data are we sharing?

Who can use it?

Can we retrieve it in usable form?

Can we explain decisions made from it?

Can we move to another tool without losing operating memory?

If the answer to those questions is unclear, you do not have an AI strategy. You have a vendor dependency strategy.

A Call to Owners

AI is not waiting for CRE to get organized.

Pricing tools will keep advancing. Listing platforms will keep fighting for data control. AP automation will keep pushing deeper into judgment-heavy workflows. Major owners will keep testing AI inside leasing, operations, construction, finance, and portfolio strategy.

The question is not whether AI enters your portfolio. It already has or soon will.

The question is whether it enters on your terms.

For asset managers, this is the moment to move the conversation out of the technology budget and into the investment committee. AI risk belongs next to NOI, DSCR, LTV, expense trajectory, compliance, and exit value.

Do not wait for a vendor dispute, diligence request, or regulatory question to discover that your data is not portable, your decision trail is incomplete, or your operating memory lives outside the business.

Start with control. Build the data & digital infrastructure foundation. Govern the data. Then choose the AI tools that serve the strategy, not the other way around.

Own your data & digital infrastructure. Operate with strategic foresight. Build for the long game.

References Cited

• Propmodo — "Legal liability follows the data, not the vendor" — https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/aff7e527-5e93-4a5c-86e0-af98320e92c8/propmodo-logo-blue-decoding-background.png?t=1736963030

• Laurent Cochet Substack — "How to Embrace AI-Driven Transformation: Insights from Keith Carter" — https://laurentcochet.substack.com/p/how-to-embrace-ai-driven-transformation

• Commercial Observer — "Proptech: Contech Giant Procore, Owner BGO Lean Into AI in Big Ways" — https://commercialobserver.com/2026/06/thiago-da-costa-procore-ai/?utm_source=sailthru&utm_medium=email&utm_campaign=proptech&utm_term=06/16/2026

• Propmodo — "A $4,200 landscaping invoice shows what AP automation keeps getting wrong" — https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/aff7e527-5e93-4a5c-86e0-af98320e92c8/propmodo-logo-blue-decoding-background.png?t=1736963030

• Propmodo — "The Zillow case that could decide who controls real estate's listing data" — https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/aff7e527-5e93-4a5c-86e0-af98320e92c8/propmodo-logo-blue-decoding-background.png?t=1736963030