Commercial Real Estate AI Strategy
The Model Just Became a Commodity. Here’s What Actually Compounds in Commercial Real Estate.
Stanford’s 2026 AI Index put a number on something CRE owners need to feel: AI capability is converging fast. The competitive moat has moved up the stack, and most portfolios are not built for it yet.
A number landed in my inbox last week that should make every commercial real estate owner sit up.
Per Stanford’s 2026 AI Index, the performance gap between open-source AI and the most expensive frontier models shrank from roughly 8% to roughly 1.7% in a single year. The cost to hit benchmark performance is falling 5 to 10 times annually. Most enterprises now run three or more model families at once. Multi-model is becoming the default, not the exception.
In plain English: the AI model is becoming a commodity. Faster than almost anyone forecasted twelve months ago.
For CRE owners and operators, that shift rewrites how you should think about AI in your buildings. The “which AI tool should we buy?” conversation that has dominated boardrooms for the last 24 months is already outdated. The better question is different, and most portfolios are not built to answer it yet.
The wrong question
Here is what most owners are still doing.
You get pitched a new “smart building” tool every quarter. Each one brings its own dashboard, its own AI claim, and its own model under the hood. You evaluate the features. You ask which tool is “best.” You pick the winner, sign the contract, and integrate it into one building.
Then a better model comes out. Or a better tool. Or your vendor gets acquired. Or pricing changes. Or the next building does not fit the same vendor profile. So you start over.
Meanwhile, every dollar of insight that tool generates becomes part of the vendor’s value, not yours. Your data trains its model. Your operational patterns refine its roadmap. When you eventually move to a different vendor, and you will, your building’s “intelligence” walks out the door with it.
That is the trap. In a market where the model is becoming a commodity, the trap gets more expensive every quarter you stay in it.
What actually compounds
When AI capability is converging across providers, the LLM model is no longer where the competitive value lives. The value moves up the stack into four things only your organization can build.
This is especially important in CRE. Buildings generate operational data every minute of every day, across every property type, in formats that do not naturally appear on the public internet. That makes the data you already sit on uniquely valuable for AI — but only if it is captured in a form you own, govern, and can reuse. Otherwise, you are a tenant in your own intelligence stack.
Your proprietary data
Models are trained on the public internet. What is not on the public internet is what your buildings generate every day: utility patterns, tenant ticket flows, access events, maintenance cycle times, occupancy signals, and vendor performance over time. That data is yours by right. The real question is whether you actually have it in a form you can use, on terms you control, with enough history to make it useful for analysis.
Your operating workflows
A frontier model dropped into a broken process produces broken outputs faster. The leverage is in the workflow itself: where the data flows, where the human reviews, where the decision sits, and what the rules are for taking action. That is an operational design problem. It does not come in a box.
Your orchestration layer
When owners eventually run three or more AI models across their buildings, and many already will without realizing it, the differentiator is the layer that decides which model gets to act, when, on which data, under whose permissions, and with what audit trail. That layer includes governance, identity, retention, lineage, and rules of use. If you do not own it, you have ceded the most valuable real estate in the stack.
Your institutional knowledge encoded into systems
The way your best operators make calls. The patterns that distinguish strong renewals from weak ones. The “how we do it here” that has historically lived in the heads of asset managers and senior property managers. Owners who turn that knowledge into structured decision logic — prompts, agents, workflows, and rules — build something no competitor can copy, regardless of which model wins this month.
That is your competitive moat. Not the AI tool. The four layers above the AI tool.
The OpticWise translation
This is not a new idea for OpticWise. It is exactly what we have been building for CRE the entire time.
We deliver a two-layer model designed for owner control.
Layer 1 — Managed data and digital infrastructure. This is the owner-controlled foundation. We design it, implement it, and operate it across your properties so it is repeatable, governed, and structured. On-site engineers and property managers should not have to become technologists. This is where your networks operate and where your proprietary data actually lives, captured in a form you can use.
Layer 2 — Vendor- and LLM-agnostic intelligence layer. This is the governed data plane and trust plane that lets any decision engine, AI model, or vendor platform plug in under your permissions. It also lets you swap any of them over time without losing your data, governance, or portfolio intelligence. Standardize it once at one property, then scale it building to building until intelligence compounds across the portfolio instead of restarting at every address.
The path from where most owners are today to where they need to go is the PPP 5C™ plan:
- Clarify — define success metrics, map who owns what, identify leakage, and document what is trustworthy and portable.
- Connect — establish secure, owner-controlled connectivity that is repeatable from property to property.
- Collect — capture and normalize the data your buildings already generate into a consistent model you can reuse.
- Coordinate — govern identity, access, privacy, lineage, retention, and rules of use.
- Control — enable any decision engine or workflow — vendor platform, internal analytics, or AI model — to act under your permissions.
What it looks like when owners get it wrong
Owners who do not make this shift get something predictable.
Every new tool becomes another silo. Data stays inconsistent, hard to trust, and locked inside vendor contracts you cannot easily read, let alone control. Every property needs custom integration: slow, expensive, brittle. The portfolio becomes a patchwork of “smart buildings” that cannot compound value because nothing transfers between them.
Insurance underwriters cannot see a clean operating story. Diligence teams find recoverable NOI you were not capturing and price it into the next transaction at your expense. AI quietly becomes automation without governance, which is worse than no AI at all.
You end up with dashboards. You do not end up with durable capability.
The villain is not any single vendor. The villain is fragmented, vendor-controlled data and digital infrastructure that turns every property into a one-off and lets your building’s intelligence become someone else’s asset.
The model dropping in price does not fix that. It makes it more expensive to ignore.
What it looks like when owners get it right
Properties become portable intelligence assets. The portfolio gains a shared, governed data foundation that is clean, reusable, and auditable. You can swap decision platforms without rewiring buildings. You can swap vendors without losing history. Risk drops across privacy, security, compliance, and audit. Outcomes accelerate across utilities, operations, tenant experience, leasing, security, and capital planning.
And the part that matters most at refinance, recapitalization, or sale is this: the portfolio’s intelligence shows up as enterprise value, not as a vendor subscription you are trapped inside.
The owners who get this right in the next 24 months will not have the most sophisticated AI. They will have the most differentiated. In a market where model performance is converging toward a 1.7% spread, differentiation is the only thing that compounds.
Where to start
Start with one building. Map who owns what, where the data lives, where operational burden is stacking up against your KPIs, and what would need to become portable. That is the Peak Property Performance® Review — the cleanest, lowest-risk way to see where your portfolio stands today and what it would take to build the layer that actually compounds.
It is not a sales pitch wrapped in jargon. Your team can even do it themselves. It is a one-building diagnostic that gives you a clear picture of your current digital control posture and your data ownership reality. The bonus is a set of specific monthly plays you could run in the next 90 days if the foundation were right.
The LLM model is becoming a commodity. That is not the threat. The threat is staying built for a world where it was not.
Own your data and digital infrastructure. Operate with strategic foresight. Build for the long game.
