AI Promises 20-30% Savings in Your Building. That's Only True if You Own Your Data.

TL;DR: AI is delivering 20–30% operating expense reductions in commercial buildings — but only for owners who control their data. Vendor-locked data limits AI to the vendor's tools and pricing. The savings flow to whoever owns the data layer, not whoever buys the platform.

There's a number circulating at every major CRE conference right now: 20 to 30 percent. That's the operating expense reduction that AI-powered predictive maintenance and real-time energy optimization are delivering in buildings where the right foundation exists. CREtech — which hosts the largest real estate technology conference in the US — is building its entire 2026 program around exactly that claim. Senior principals from Blackstone, Hines, RXR, and PGIM are taking it seriously.

Here's what nobody is saying clearly: that upside only flows to the owner if the owner controls the data layer.

Who Actually Captures the AI Upside

AI applied to building operations needs data — sensor readings, HVAC performance metrics, energy consumption, maintenance records, occupancy patterns. That data lives somewhere. It lives in systems. Those systems are owned and operated by someone.

In most commercial buildings today, that someone is a vendor. Your building management system vendor. Your property management software provider. Your connectivity provider. Your IoT platform.

When an AI algorithm identifies that replacing a chiller component three weeks earlier would have saved $40,000 in emergency repairs — who gets that insight? If the vendor controls the data, the vendor controls the insight. They can package it as a premium feature and charge you for access to your own building's information. They can — and often do.

That's the governance problem that stays invisible until it isn't.

The Newmark Report Is Asking the Right Question

A Newmark report released this week cuts through what the firm calls technology panic to provide grounded insights on how AI actually affects office space and operations. Their finding: AI doesn't replace the infrastructure problem — it amplifies it. Buildings with cleaner data foundations capture more upside. Buildings with fragmented, vendor-siloed data find AI gives them one more tool they cannot fully control.

The winning office buildings right now — the ones driving the market's strongest quarterly leasing performance since 2018 — are separating themselves through operational intelligence, not just aesthetics. Square footage and coffee bars don't close leases anymore. Reliability, predictable operating costs, and demonstrable performance data do.

Owners who can show a prospective tenant the trend line on operating costs and the improvement trajectory have a measurable advantage in lease-up and in refinancing conversations. That evidence only exists if the owner has access to the data and can actually use it.

The Foundation Question

Before any building owner invests in AI tools — before they sign the contract, add the sensors, or integrate the new platform — one foundational question has to be answered: do you own your data and digital infrastructure?

Not “do you have access to a dashboard.” That's not ownership.

Ownership means admin credentials in your hands. Data export rights written into contracts. Vendors operating under your rules, not the other way around. When you switch vendors, your data comes with you — it doesn't stay behind in a platform you can no longer access.

Most owners, when they really audit this, find they don't own what they thought they owned. A property management company installed the BMS software. A connectivity vendor owns the network. An IoT company has proprietary sensors that only communicate with their cloud. Three different systems. Three different vendors. Three sets of data that don't speak to each other and aren't portable.

Ask any of those vendors for a full data export and watch the conversation get complicated.

That's not a foundation you can build AI on. That's a liability you have been paying for monthly.

The 5C Path to AI Readiness

The first step toward capturing the operational upside AI can deliver is not buying AI. It's doing a Clarify audit — mapping what you own, what your vendors control, where data lives, and where it leaks.

The five steps are Clarify, Connect, Collect, Coordinate, and Control. You cannot skip ahead. Buying an AI platform before you have Clarified who owns your infrastructure is like installing a state-of-the-art home theater in a house with faulty wiring. The investment is real. The returns are not.

After Clarify comes Connect — establishing a secure, owner-controlled connectivity layer repeatable property to property. Then Collect — capturing high-fidelity, usable data in a consistent model. Then Coordinate — governing identity, access, privacy, and rules of use. Then Control — enabling any decision engine, any AI tool, any analytics platform to operate under owner permissions.

That last step is where the 20-30% actually lives. But you cannot get there without the foundation.

What to Do This Week

If you have been watching AI discussions in CRE and wondering when to make a move, the answer is: start with ownership, not tools.

Audit your current vendor agreements. Do you have admin credentials to your building systems? Do your contracts include data export rights? If a vendor walked tomorrow, could you operate the building without them — or would the data go dark?

If you have not done a serious audit of your building's data and digital infrastructure in the last 18 months, that's the first conversation to start. Not an AI demo. An ownership audit.

The savings are real. But they belong to the owner — and only if the owner has built the foundation to claim them. If you want to start with a Clarify audit for your portfolio, reach out at opticwise.com to begin a PPP Review. That's where the 20-30% begins.