Your Board Just Made AI a Strategy Question. You Need a Strategy Answer.

A note hit my inbox this week that crystallized what every CRE owner I know is feeling.

"Your board just changed the AI question." For two years AI lived in the IT budget. Last week it moved into the boardroom. Most executive teams will walk into the next board meeting with four different vendor pitches sitting in their inbox and no actual answer to a question the board is now going to ask out loud: what is our AI strategy as an asset owner. That same week, Realcomm announced the CEOs of Yardi, MRI Software, VTS, and Altus Group are sitting on the same main stage to talk AI and platform strategy. A new book — Artificial Intelligence in Commercial Real Estate — landed on the same Monday. Fortune ran a long piece on how the AI data center boom is rewriting CBRE, the biggest CRE services firm in the world. Everyone is talking about AI. Almost nobody is talking about what the AI is actually going to run on inside an owner's portfolio.

That gap is the whole story.

The board is not asking the right question

Most board AI conversations sound like this. The CIO walks through three or four vendor pilots, names a couple of LLMs, talks about a co-pilot rollout, and asks for budget. The directors nod. Somebody says "we should be doing more with AI." Someone else asks "are we behind?" The meeting ends. Nothing portable was built. The owner is one more quarter into letting the AI roadmap be set by whichever vendor showed up that month.

This is the wrong shape of conversation. The right shape starts from a different question. Not which AI tool. Not which model. The board's real question is: when AI becomes table stakes in our industry, what part of our AI capability will we actually own. The cost of running AI is collapsing. Per Stanford's 2026 AI Index, the gap between open-source models and the most expensive frontier models shrank from roughly 8% to 1.7% in a single year. Cost to hit benchmark performance is falling 5 to 10x annually. Most enterprises now run three or more model families at once. The model is becoming a commodity.

When the model is the commodity, the moat is the layer above it. That layer is owner-controlled data, owner-controlled workflows, owner-controlled orchestration, and the institutional knowledge of how your buildings actually run, encoded into systems you control. Those four things are not for sale from a vendor. They have to be built. And they are exactly what most CRE owners do not have today.

Why the buyer profile changed overnight

The reason this conversation moved to the boardroom is that the buyer changed. Three years ago, the AI-in-CRE buyer was a director of innovation or a head of operations evaluating a specific tool for a specific workflow. Today the buyer is the CFO and the asset management leadership, because the question is not "should we automate lease abstraction" — the question is "what does an AI-ready portfolio look like at refinancing, at sale, at the next debt conversation." That is a capitalized-value question. NOI, cap rate, DSCR, refi math. The CFO does not care which model summarized a lease. The CFO cares whether the buyer of the portfolio can plug their own AI into the operating data and value the building accordingly — or whether the data is trapped in a vendor stack that does not transfer in the trade.

This is also why the trade press is full of M&A announcements right now. Sachem Capital and IRG just announced a $3.4B industrial REIT combination, according to Connect Media this week. Scale is becoming a precondition for the kind of standardized, repeatable digital infrastructure that makes AI actually work across a portfolio. Bigger owners are buying their way to a coordination layer. Smaller owners need to build one before they sell.

The four AI vendor pitches in your inbox

Look at the four AI pitches that landed in your inbox this month. I will bet you a coffee they fall into the same four buckets every owner is seeing. One is from your BMS or controls vendor announcing AI optimization inside their proprietary platform. One is from your property management software vendor launching an AI co-pilot that only reads their own database. One is from a smart-building analytics startup with a new agent layer. One is from a leasing CRM bundling an AI tour qualifier into the next renewal. Every one of those is real. Every one of those is good news for the vendor. And every one of those, if accepted on the vendor's terms, locks another piece of your portfolio's intelligence inside someone else's environment. You end up with four AI brains, four data planes, four sets of permissions, and no way to make them answer the same question on behalf of the owner.

The board does not want to hear that you piloted four things. The board wants to hear that you have an AI strategy.

What an AI strategy actually looks like for a CRE owner

An AI strategy for a CRE owner is not a model selection. It is two things in a specific order.

First, an owner-controlled data plane. The data your buildings already generate — leasing, work orders, energy, access, IoT, tenant communications — collected and normalized into a consistent model under your governance. Not as a side effect of a vendor purchase. As an explicit asset. This is what the PPP 5C™ plan calls Clarify and Collect. Without it, every AI you run is at most useful inside one vendor's silo and at worst dependent on data the vendor decides whether to give you.

Second, a vendor- and LLM-agnostic intelligence layer that sits above your data plane. Property Brain™ at the property scale, Portfolio Brain™ at the portfolio scale. The intelligence layer is where models, agents, vendors, and decision engines plug in under your permissions and your governance — and where any of them can be swapped over time without losing data, without losing institutional knowledge, and without losing portfolio intelligence. Property Brain™ → Portfolio Brain™ is built to outlast any single model, any single vendor, any single platform. That is the strategy. The vendor pitches in your inbox are just inputs.

When you have that layer, the board AI question becomes easy to answer. You point at the data plane and the intelligence layer and you say: every AI capability we deploy runs through this, on data we control, under permissions we set, and we can change any of it without breaking the portfolio. That is the answer a director with a fiduciary obligation actually wants to hear.

Why this is a Coordinate and Control conversation

In PPP 5C™ language, this is a Coordinate and Control story. Coordinate is identity, access, lineage, retention, rules of use. Control is the layer that lets any decision engine or AI agent act under owner permissions. AI without Coordinate is what creates the shadow-IT problem at every portfolio I have seen — a tool nobody chartered, running on data nobody mapped, producing recommendations nobody can audit. AI without Control is what creates the dependency problem — the vendor decides what the AI can do, when it runs, and what happens to the patterns it learns when you leave.

If you don't own your data & digital infrastructure, your vendors do. And when the AI brains in your buildings know more about how your portfolio operates than your own team does, your operating standard just became someone else's proprietary asset.

What to put on the board agenda before the next meeting

Three items, in this order.

One. An honest read of where your data plane is today. Not vendor by vendor — but as the asset manager would view it. What data do you actually have, in what form, under whose terms of use, and how portable is it if the vendor relationship changes. Most owners discover the answer is "less than I thought" and "harder to move than I thought."

Two. A named owner for the intelligence layer. Not the BMS vendor. Not the PMS vendor. An owner-side accountable function — internal or partnered — that holds the standard for how AI gets plugged into the portfolio. Most CRE org charts do not have a slot for this yet. That is the actual gap most boards are trying to name when they ask about AI strategy.

Three. A 90-day data and digital infrastructure review of one stabilized property in the portfolio. Pick the one with the most complex operating stack. Map the data, the permissions, the vendor terms, and the AI surface area. Report it to the board with cost, NOI impact, and risk in the asset manager's language: NOI, OpEx, DSCR, exit. The board does not need a multi-year plan in the first meeting. The board needs to see that you have an actual approach.

The PPP path forward

The PPP 5C™ plan was built for this exact moment. Clarify (define success and map ownership). Connect (secure, owner-controlled connectivity property to property). Collect (capture and normalize data). Coordinate (govern identity, access, lineage). Control (run any decision engines or AI agents under owner permissions). The plan does not care which AI model you use. It cares whether the layer underneath the AI is yours.

That is the answer to the board's new AI question. Not a model name. Not a vendor logo. An owner-controlled data and digital infrastructure layer plus a vendor- and LLM-agnostic intelligence layer above it, governed by the owner, repeatable property to property. The AI strategy is the layer, not the tool.

If you are walking into a board meeting in the next 30 days with no AI strategy and four vendor pitches in your inbox, the first conversation to have is the data and digital infrastructure review. That is what makes the board AI question answerable. Own your data & digital infrastructure. Operate with strategic foresight. Build for the long game.

References Cited

• AI Leadership (Geoff Woods) — "Your board just changed the AI question" — 2026-05-19 (email newsletter; URL: verify before publishing)

• Realcomm — "View from the Top: Enterprise Platform Strategies for the Future — Yardi, MRI, VTS, Altus CEOs" — https://www.realcomm.com/realcomm-2025/program/

• Credible CRE — "New Book Release: Artificial Intelligence in Commercial Real Estate" — 2026-05-18 — https://news.crediblecre.com/

• Fortune — "How the AI data center boom is transforming CBRE" — 2026-05-20 — https://fortune.com/

• Stanford 2026 AI Index — model performance gap and cost decline figures — https://aiindex.stanford.edu/

• Connect Media — "Sachem Capital, IRG to Form $3.4B Industrial REIT" — 2026-05-18 — https://www.connectcre.com/