AI Is Speeding Up Commercial Real Estate… But Is Anyone Checking the Math?

Commercial real estate has never been about speed, deals take years to complete, projects take longer to build, and the buildings last sometimes for centuries. Capital decisions often outlive economic cycles that no one really controls anyway. The patience built into commercial real estate acts as both a safeguard and a constraint. At the very least, they take away the opportunity to make impulsive decisions that could prove very costly.

So how does AI factor into the safeguards?

Many have promised that AI will transform the commercial real estate decision-making process – faster underwriting, smarter forecasting, better risk assessment, real-time performance optimization, and more refined investment allocations. There are countless algorithms ready to rank buildings, dashboards full of all sorts of quantitative data, and predictive models evaluating every potential risk. On the surface – what a win! Who wouldn’t want all this delivered at hyper speed?

But at some point, you have to ask, if it seems too good to be true…

A couple of considerations. Are speed and intelligence the same thing? Or are these two things often in tension? Are all factors involved in any real estate transaction quantifiable? And is there a line that needs to be drawn? It’s easy to measure energy use, maintenance costs, occupancy trends, lease expiries, and capital reserves. But is that all you need to determine long-term asset value?

What about a building’s ability to withstand shocks – extreme weather, supply chain disruptions, regulatory shifts, or tenant turnover? In other words, resilience. Are these as readily quantifiable? Can they be turned into data points? How does AI effectively weigh these potential variables? These things are often shaped by design decisions made decades ago, the quality of the building’s current owners/managers, and sometimes a building operator who has spent years learning every nuance of a mechanical system. AI can analyze the patterns from the inputs, but not from the humans involved.

Should we be concerned with Ai-driven decisions over-prioritizing short-term efficiency at the cost of long-term stability, over-ranking a buildings short performance while underestimating a building’s vulnerability to a weather event that has never happened to this building? And what about a mid-tier building that is very well run, but the owner has yet to invest in sensors? Will that be downgraded?

Are too many people asking, “Does this building look good to the algorithm?”, rather than “Is this a good building”? These are not the same questions, yet it can’t help but impact where the money is flowing.

What about the cultural risk, are the traditional valued judgements, relationships, and contextual understanding being ignored and eventually lost? Are we too quick to favour AI “objective precision”, even though it’s built on an incomplete or biased data set? Unfortunately, a computer is often trusted to produce a risk score or recommendation over a human. Even when the human team is sensing that something is off.

Is there a balance to be found, perhaps using Ai as almost a decision support tool rather than a decision maker? Surely, at least right now, isn’t the best option to pair algorithmic insights with human expertise. Hooking up speed and wisdom.

This is where standards help, transparency, and a shared framework. Programs like BOMA BEST can help buildings become more “legible” to AI without reducing them to simple data points. This allows for consistent measures of performance, operations, and resilience. And AI models would be allowed to work better with more informed data.

The question of whether AI will shape commercial real estate is mute, it already has. The wiser question is whether we’ll allow AI to narrow our perspective, or will we use AI to deepen it.

Real estate plays a long game. It is up to the industry to make it smarter, not just faster.