AI in CRE Today is Mundane. How and When Will That Change?

A new report by PwC and the Urban Land Institute touches on many topics, one of them being artificial intelligence.

The opening is far from the excitement you might hear from vendors and consultants. “Despite the hype and popular attention on artificial intelligence, actual CRE uses appear to be limited and most mundane to date,” they wrote. “The range and sophistication of industry uses are likely to expand quickly, given the promise of the technology and the volume of venture capital investment going into the sector.”

This view is something that comes from experience. New promising technologies often have limited and dull applications when they first come out. When the web came to the Internet, for example, initial sites often took the form of what was called brochureware, which was the reproduction of printed materials in digital form. Dull and in hindsight predictable because fully using new capabilities means time to understand what they are.

Trying to look forward, they wrote, “The exciting potential includes probabilistic models to help predict property climate risks, identify investment opportunities, and construct higher-performing portfolios. AI adoption could replace many types of routine white-collar work, but jobs losses could be offset by greater overall economic growth as well as space demand from AI firms.”

Ultimately, no one knows. We’ve seen claims of probabilistic and predictive models in the mid-to-late 1990s to improve supply chains, for example. There was some progress, but nowhere near what was promised, and look at what happened during the pandemic. It was the type of widespread disaster that experts had predicted for 20 to 30 years unless companies did the hard work to take risk out of their operations and get a view into how their chains worked. The paperless office was supposed to come in the 1980s. It never did.

Inevitably technologies are tools, and people have to learn to use them, which always takes time and companies often want instant results without spending the time for learning and experimentation.

We’ll look at more of what the report had to say in some subsequent posts, but what is most important to realize at first is that everything takes more time, more effort, and more patience than those trying to make sales of software and hardware would want potential customers to think.