TL;DR: AI in commercial real estate fails when it runs on fragmented, vendor-controlled data. The operators capturing real value built a clean, owner-controlled data foundation first (admin credentials, export rights, governance), then layered AI on top. Without that foundation, AI produces confident outputs from incomplete inputs.
A CREtech survey found that AI access is now widespread across commercial real estate. Most firms have subscriptions. Most employees can use the tools. Most are using them for chat, transcription, and drafting basic documents. That pattern is not a failure of the AI platform. It is what happens when a capable tool meets a weak foundation.
Here is the progression visible across portfolios over the past two years. Firms invest in AI tools. They get incremental improvements in administrative tasks. They wonder why their building operations do not actually get smarter. Meanwhile, a smaller group of operators who built a clean, governed data foundation first are using AI to make operational decisions: maintenance prioritization based on system history, utility variance detection against normalized baselines, lease renewal probability scored from actual tenant behavior patterns. The difference between those two outcomes is not the model. It is the data underneath.
Why Widespread Access Is Not What It Sounds Like
When a survey reports that AI access is widespread in real estate, what it measures is subscription adoption. What it does not measure is whether those tools have access to clean, complete, trustworthy, governed data about what is actually happening in the buildings being operated. Those are entirely different conditions, and confusing them delays real operational improvement by years.
Consider what an AI system needs to help run a building better. It needs reliable sensor data from systems that are connected and actively monitored. It needs maintenance history that is documented, consistent, and queryable across time. It needs utility data in a normalized format across properties, not seven different formats from six different metering vendors. And it needs access controls and governance so it operates within owner permissions rather than on whatever data it can reach through default vendor configurations.
When that foundation does not exist, the AI trains on noise. The outputs sound confident. The inputs behind them are incomplete. McKinsey's recent work on agentic AI in CRE identifies four operational domains where AI can create real value: leasing, asset management, tenant experience, and facilities operations. What the framework assumes, and what most operators have not yet achieved, is a clean, integrated data layer those agents can trust and act on.
The Data Governance Gap
The common pattern across buildings is not a lack of technology. It is a lack of practical data ownership. Systems get installed. Data gets generated. That data lives inside vendor platforms, under vendor governance, on vendor terms. The owner has a dashboard and sees outputs. But the underlying data often cannot be extracted in a usable format, cannot be normalized across properties, and cannot be trusted enough to feed into a decision system that carries financial weight.
This is the invisible tax of fragmented data & digital infrastructure. It becomes most visible at exactly the moment an owner wants to do something ambitious: deploy AI that actually improves operations rather than just summarizes reports.
Rajiv Chandrasekaran, CEO of Quantum Data Technologies, described the core issue in a recent conversation: organizations are not lacking data, they are lacking clarity. Teams spend hours gathering data across systems, try to assemble a coherent picture, and by the time they reach a conclusion, the window to act has already passed. That describes most CRE operations today, and no AI subscription changes it without addressing the data layer underneath.
Forbes Tech Council recently argued that verifiable digital infrastructure, the ability to demonstrate that data meets regulatory and operational requirements, is the next competitive prerequisite. In building operations, it is simpler than that: without a governed data foundation, AI cannot distinguish signal from noise, and the owner has no way to verify which one it acted on.
What Actually Changes Things
The operators getting real value from AI share one characteristic: they built the data foundation before they bought the tools. They went through the work of mapping what data their buildings generate and where it actually lives. They established who owns what: admin credentials, export rights, portability terms. They normalized data across systems so it is consistent and queryable. They put governance in place so any tool operating on that data does so under owner rules.
In the PPP 5C™ framework, these are the Collect and Coordinate phases. They are the unglamorous prerequisites that make the Control phase (where AI and decision engines actually earn their cost) possible.
The Honest Test
Pick three systems in a building that generate operational data: HVAC, access control, energy monitoring. Ask three questions about each.
Can you export that data in a format you can use without going through the vendor? Is the data normalized and consistent enough to compare across two of your properties? Do you have admin-level access, or do you have tenant-level access presented as owner access?
If the answer to any of those is no, the building does not have a data foundation ready for AI. It has a collection of data subscriptions. That distinction matters more than which AI platform is under evaluation.
Where to Start
Start with a Clarify step. Map what you own. Identify the gaps. Understand what is vendor-controlled and what is yours. That one step changes every vendor conversation afterward, because you enter knowing exactly what you need (portable data, admin credentials, governance on your terms) instead of accepting the vendor's default.
If you are evaluating AI tools for buildings right now, pause on one question: what data will this AI run on, who controls that data, and what happens to the AI's value if that data relationship changes? The answer tells you whether you are building an intelligence layer or renting one.
Start with a Peak Property Performance® Review. It shows exactly where you stand and what it would take to build a foundation that makes the next technology investment pay off.
Own your data & digital infrastructure. Operate with strategic foresight. Build for the long game.
Bill Douglas is the CEO of OpticWise, where he leads the company's mission to help commercial real estate owners take control of their data and digital infrastructure. With over three decades of entrepreneurial experience and a track record of leading companies onto the Inc. 5000 list, Bill brings a systems-minded, owner-first approach to everything OpticWise delivers. He holds a mechanical engineering degree from Georgia Tech and is a graduate of MIT's Enterprise Forum Entrepreneurial Masters Program. Bill is the co-author of Peak Property Performance (Fast Company Press).
Your Next Step
Complimentary CRE Data & Digital Review Session
One building. Map who owns what, where data lives, and where operational burden stacks up vs your KPIs.
Topic clusters
This article is part of the following OpticWise topic clusters. Each pillar page summarises the topic and links to related Insights pieces:
