Investment

Multifamily AI Readiness Depends on Years of Organized Property Data

AI can support apartment leasing and operations when it draws on consistent, long-term records. Scattered or missing data makes reliable decisions harder.

Nov 16, 2026 ● 3 min read
Sign in to save

AI’s real constraint is the history behind it

Apartment operators can buy access to capable AI models. They can’t buy years of well-organized operating history on demand. That difference is more significant than it looks: nearly every multifamily company is testing AI, but far fewer are using it to guide routine business decisions. The sticking point, in my experience, is often data scattered across systems, defined differently by different teams, and too limited to show whether a change is working.

I came to this view after working on both sides of the technology question. Before becoming chief technology officer at Cortland, an Atlanta-based, global multifamily investment, development and management company, I managed a portfolio on its investment team. I still judge technology by asking what it does for an asset. AI readiness starts with whether the operator controls the data the tools need—and whether it has collected enough of it to make a sound comparison.

A person reading a floor plan on her digital tablet.
Getty Images

The model is the easy part

Operators often get access to similar AI models through the same vendors. The useful difference comes from the information and business context those models can draw on. Leasing, maintenance, renovation and underwriting records may live in separate systems—or with outside providers—and even basic terms can mean different things to different teams. Put AI over that patchwork and you get results people have reason to distrust. Decisions then carry on as before.

Here’s the thing: choosing a tool before naming the decision it should improve puts technology ahead of the business problem. If you’re evaluating AI, start with a specific decision and ask what history would make it better.

Lost history takes time to rebuild

Some technology gaps shrink when a company spends more. Missing data is harder to fix. If an operator failed to record lease outcomes, resident interactions or renovation results consistently five years ago, reconstructing that record now can be difficult and expensive. The data doesn’t fully explain how quickly each company can catch up, but it does show why the gap may widen: disconnected operations keep losing years of context while connected systems add to a history that can improve future analysis.

Cortland began building its own data foundation in 2016, before much of the industry recognized its potential. The company rebuilt its data model around residents in 2019; previously, it centered on properties. In 2021, it brought investment, construction, property management and resident experience information onto one platform. That platform now holds about a decade of records covering more than one million Cortland leases.

When generative AI and more accessible machine learning arrived, Cortland says it could move into applications such as AI-assisted leasing, revenue management and payment-risk prediction without first spending years reconciling its data. It tested narrow, measurable uses against its historical baseline. The company reports improved lead-to-tour conversion, application speed and resident retention, as well as lower fraud and delinquency. Those results are the company’s account, but the sequence is instructive: build the record first, then test specific uses.

Data ownership is an operating choice

For a vertically integrated business, information from investment, construction and property management has value across the company. A leasing pattern at one community could inform underwriting for a future acquisition; renovation results could shape the next value-add budget. That connection depends on leaders treating data as an enterprise asset, rather than something belonging only to the team or vendor that generated it.

Owning the underlying data also gives an operator more say over what gets recorded, how much detail is kept and how information is structured. Licensed platforms can be quicker to adopt, and outside tools can still serve a purpose. But a vendor’s product design will influence what the system preserves. If the operator’s future questions differ from the vendor’s assumptions, that choice can become a constraint.

Keep people in the decisions that call for judgment

Renting a home is personal. Automating routine inquiries, tour scheduling and payment follow-up can give on-site staff more time to build relationships, but technology can’t replace the human context behind a resident’s choices. Cortland says its systems are designed to flag decisions—such as a resident who may be at risk of not renewing—for a community manager who knows the person involved.

Portfolio data may reveal a trend sooner than a traditional report. People close to a community can help explain what’s driving it and what to do next. And yet, teams are more likely to use AI when they see it as giving them time with residents, rather than second-guessing their judgment.

The industry’s focus on the next AI product misses the longer-term decisions: who controls the data, and whether automation supports resident relationships. Those choices build over years. The operators best positioned to use whatever comes next are likely to be the ones that have already invested in connected data and kept people responsible for the human calls.

Scott Moore is chief technology officer at Cortland. The views expressed are his own.

Source: Scott Moore · www.multifamilydive.com

Comments

Sign in to join the discussion.