Scott Moore is the chief technology officer at Cortland, a global, vertically integrated multifamily investment, development and management firm based in Atlanta. Opinions are the author’s own.
Nearly every multifamily operator is experimenting with artificial intelligence. Far fewer have moved it from experiments into the everyday decisions that shape portfolio performance.
When I ask why, the answer almost always comes back to the data underneath, which is fragmented across platforms, defined differently by each team and too short on history to tell whether anything is actually improving.
Before I ran technology at Cortland, I ran a portfolio on our investment team and the question I still ask of every technology decision in our industry is, “what does this do for the asset?”
That’s why I’ve come to see AI readiness as less a question of which tools to buy than of whether an operator owns the data those tools depend on.
An operator that owns its data decides what to capture and can build on it. In the multifamily industry, the data an AI system needs is often the data an organization needed to start collecting years ago.
The model is rarely the bottleneck
Capable AI models are now widely available, and most operators access similar ones through the same vendors. The model is the common part.
What separates a general-purpose model from one that produces something useful is the data and context it works from. A model is only as useful as the history it can see, and that history has to be clean, consistent and connected across the business.
For many operators, leasing, maintenance, renovation and underwriting data sit in separate systems or with external vendors, and teams can define the basics differently. Layer AI across that, and the result is an output no one fully trusts. So nothing changes in how decisions get made.
Technology is the means. The decision it improves is the goal, and the business has to name that decision before it chooses the technology.
Why time makes the problem harder
Most technology gaps can be closed with increased investment. This one can only be narrowed. If an operator didn’t capture lease-level outcomes, resident interactions and renovation results cleanly five years ago, recovering that history is difficult and costly.
That is why multifamily’s AI readiness gap is likely to widen before it narrows. Every year on disconnected systems is a year of context and learning an operator’s future models may not have, while operators with a connected history keep adding to a record that makes each insight more valuable.
At Cortland, we decided to own our data rather than rent it in 2016 — years before most of the industry recognized what that data could be worth. We made this decision by focusing on business needs and goals rather than short-term technology capabilities.
In 2019, we rebuilt our data model around the customer. Until then, like most of the industry, it was built around the property. In 2021, we brought investments, construction, property management and the details of the resident experience onto one platform. Today, that platform holds roughly a decade of history across more than one million Cortland leases.
When generative AI and more accessible machine learning arrived, our teams didn’t have to spend years reconciling data first. We moved quickly into areas like AI-assisted leasing, revenue management and payment-risk prediction, and we could measure results against years of our own baseline.
By starting with narrow, measurable problems and expanding from there, we have seen improvements in lead-to-tour conversion, application velocity and resident retention, alongside declines in fraud and delinquency. The operators I see as best positioned to benefit from AI over the next five years have largely already made the choice to build that foundation.
Readiness is an operating decision, not an IT project
Looking back, the decisions that mattered most weren’t purely technical ones, but also choices about how the business would run.
The first key choice was treating data as an enterprise asset rather than a byproduct of individual systems. In a vertically integrated company, investments, construction and property management each generate information the others need.
A leasing trend on-site should inform how we underwrite the next acquisition, and renovation results should shape the next value-add budget. That connectivity only happens when leadership across those groups agrees the data belongs to the enterprise, not to the team or vendor that produced it.
The second, and the more significant, was choosing to own that foundation rather than rent it. It followed from vertical integration: We already owned the parts of the value chain that many operators assemble from third parties. So owning the data that runs through all of them was a logical extension.
Licensing a platform is often faster in the near term, and outside tools still have a place in our business. However, a rented system decides what gets stored, at what level of detail and in what structure, based on the vendor’s product rather than the operator’s future questions. Owning the foundation means we make those choices.
Where judgment still sits
We made the same kind of choice about people. Choosing where to live can be an emotional decision and residents want to feel part of their community and know the people who manage it. Better technology doesn’t change that.
We centralize and automate the transactional work, such as routine inquiries, tour scheduling and payment follow-up, to give on-site teams more time to build the experiences that shape whether someone renews. We chose to design our models to hand a better decision to a person, like a renewal-risk flag to a community manager who knows the resident behind it.
Data can show what is happening across a portfolio sooner than a traditional report, but the people closest to a community can contextualize why and what should happen next. In our experience, teams tend to adopt AI more readily when they see it as something that gives them more time with residents rather than something that second-guesses their judgment.
The choices that outlast the technology
Much of the industry’s AI conversation centers on which product comes next. The more consequential conversations are less visible but should focus on whether an operator owns the data its business runs on and whether it uses AI to replace relationships or support them. Those choices compound over years before they show up in results.
Renting a home is a relationship decision, and building that relationship is what our teams are there to do. Owning the data is what lets us see those relationships clearly and put the right information in front of the right person.
The operators who get both of those choices right will get the most from whatever comes next.
Click here to sign up to receive multifamily and apartment news like this article in your inbox every weekday.