Maureen Waters is CEO of Measurabl, the world’s most widely adopted sustainability data platform for commercial real estate, supporting more than 180,000 properties across 97 countries, representing over 28 billion square feet and more than $6 trillion in real estate assets. In her first three months as CEO, Waters met with more than 80 commercial real estate organizations across 23 cities and eight countries to understand how customers are preparing for the next generation of AI, trusted data, and intelligent automation. In August 2026, Measurabl announced a partnership with Inveniam to make commercial real estate sustainability data traceable to its source and independently verifiable.
Commercial real estate is at an inflection point. Over the past decade, the industry has done an incredible job of digitizing sustainability and operational data. Today, the challenge is no longer collecting data, it’s making that data trusted, connected, and actionable.
Sustainability and operational data are no longer used just for reporting. They’re increasingly influencing financing, valuations, insurance, due diligence, investment decisions, and regulatory compliance. As a result, expectations around data quality, traceability, and auditability have fundamentally changed.
During my first three months as CEO, I met with more than 80 customer organizations across 23 cities. Regardless of geography or customer type, I consistently heard four themes: better data quality, more automation, clearer insights, and AI they could trust. All four depend on one thing the industry has yet to solve, the data underneath them.
Priced In, but Not Yet Trusted
The market has already made its decision. Efficient, high-performing buildings command measurable rent and valuation premiums, while assets exposed to rising energy costs, regulatory penalties, and stranding risk increasingly trade at a discount. Sustainability metrics are no longer peripheral, they’re becoming financial metrics.
But pricing something into an asset isn’t the same as trusting the number behind it. If these figures now influence valuations, financing terms, and investment decisions, how confident are we in the data itself? For many organizations, the honest answer is simple: not enough.
The Data Challenge
The challenge isn’t AI. It’s the data feeding AI.
Much of the information driving financial decisions lives inside the least standardized and unstructured documents in commercial real estate: utility bills, leases, appraisals, engineering reports, compliance filings, and certifications. Utility data alone spans more than 10,000 utility providers worldwide, each with its own format, language and convention, and no shared standard between them. That’s the data feeding GRESB scores, financing terms and net-zero commitments.
In our own work, we pull data from 3,800 utilities, across 40 countries, and run quality checks on roughly 65,000 utility bills a month. Even at that scale, the hard part isn’t collecting the numbers, it’s proving where they came from once they’ve left the source document.
Point a capable model at real estate data today, and it will do exactly what it’s built to do: reason with complete confidence over numbers that may be incomplete, inconsistent or simply wrong. That’s the uncomfortable truth about generative AI, it doesn’t know what it doesn’t know. A confident output built on an unreliable foundation is still unreliable, no matter how fluent it sounds.
The foundation problem, and therefore the solution, is the data that underlies the model. You can’t reach the high-value outcomes — automation, agentic workflows, faster and better decisions — without first fixing what sits underneath: data that is trusted, structured and connected. McKinsey’s research on enterprise AI found that only 7% of companies have fully scaled AI across their organization, and more than two-thirds of high performers say data, not model capability, is the primary obstacle.
It’s a costly problem to ignore. Every time a team points a general-purpose model at a pile of unstructured documents, it pays again to have the model reorganize the same information it structured last time. That’s a recurring tax, and structuring data once at the source removes it.
Trust Is the Ceiling, Not the Technology
The trade-off most firms are stuck on is this: to get value from AI, you have to feed it data, but the most valuable real estate data is proprietary and confidential, and you can’t hand it to a model or third party without risk. The upside is real, but the loss of control is too high a price. Across hundreds of practitioners, trust and accuracy rank as the single biggest barrier to AI adoption — ahead of security, compliance and cost. This is a trust ceiling, not a technology ceiling.
The data shows how widespread this is. Today, 96% of CRE professionals are using AI, but 51% say trust is a barrier to greater adoption. It isn’t that people distrust AI in the abstract; it’s that they don’t trust what would happen if they connected it to their real, unstructured, unverified data.
Frequent use and low trust aren’t a contradiction — they’re exactly what you’d expect when the tool is good enough to try, and the data isn’t yet good enough to rely on. Most adoption stays at the surface, at analyst-level tasks; only a small minority have moved into genuine workflow automation. The gap between what firms believe about AI and what they’ve actually deployed is a trust gap.
Real Estate’s Fintech Moment
I’ve spent much of my career at the intersection of real estate and technology, and this moment feels familiar. Financial services went through the same evolution a decade ago: first you digitize information, then you standardize it, then you establish trust and provenance, and only after that can you automate decisions and realize the value of AI.
Fintech taught us that digitizing information wasn’t enough. Banks and investors ultimately needed trusted infrastructure before they could automate lending, improve risk management, and confidently adopt AI.
From Fragmented to Verifiable Data
Trust can’t be bolted on at the moment of decision. It has to be established where the data originates. Information that is structured, validated, and traceable back to its source carries significantly greater value than information assembled later with uncertain lineage.
One approach emerging across the industry is to credential information at its source before it enters AI workflows: structuring information from complex real estate documents, validating it against domain expertise, and preserving its provenance through an auditable chain of custody.
When individual data points can be independently verified and traced back to their original source documents, organizations gain something far more valuable than efficiency, they gain confidence. That confidence enables permission-based collaboration, faster diligence, more reliable reporting, stronger investment decisions, and ultimately AI systems organizations are willing to trust.
Done well, this resolves the trade-off between value and control rather than trading one risk for another: independent auditability without exposing the underlying documents, permission-based sharing across owners, lenders and investors, and a record of exactly how each figure was derived and who touched it. What it eventually means is speed made safe — diligence, valuation and underwriting timelines that historically ran months can compress to days, not because anyone is cutting corners, but because no one has to re-check the foundation before building on it.
The Firms That Win Will Trust Their Own Data
The scale of what’s on the table is easy to underestimate. Morgan Stanley recently surveyed 935 executives across the five sectors it judges most exposed to AI. Real estate management and development came last on productivity gains realized so far. The industry’s greatest opportunity is currently constrained by its weakest foundation.
Every organization I met during my listening tour wrestled with the same question: How do we unlock the value of AI without compromising control of our most important data? Fortunately, that isn’t an either-or decision. The competitive advantage of the next decade won’t come from access to better AI models. Those models will become increasingly available to everyone. The real differentiator will be trusted data.
Everyone will have access to good models — that part of the race is already over, and it was never going to be what separates the winners. What will separate the organizations that compound their advantage from those treating AI as another line item is whether they can trust and safely use their own data. That’s not a model problem. It’s an infrastructure problem — and unlike most of what this industry has been asked to fix over the past decade, it’s one we can solve now.
