Artificial Intelligence (AI) as an idea is more than 50 years old, though only in the last five years or so has AI come to dominate the world of business. Business communication, that is. Just as it is hard to overestimate the potential of AI to create change at massive scale, it is nearly impossible to exaggerate the degree to which talk of AI in both the news and in the marketing slicks created by PR agencies is hyperbolic and often disconnected from ground realities.
So what about AI in real estate? Over the past several years, many companies have made claims about using AI to “disrupt” the space; I am myself affiliated with more than one such company. So what, indeed, is the verdict: Reality or rhetoric?
I’ll use my authorial perch to play both judge and jury here, but, spoiler alert, the jury is hopelessly hung. As judge, I’ll say that so far we’ve seen far more rhetoric than reality.
To be fair to the AI community, though, let’s at the outset state areas in which AI has helped in fundamental ways:
- Automated Valuation Models (AVMs)and valuations In general
- Computer Vision (CV)
- Incremental process-oriented upgrades
With regard to AVMs and valuations in general, the ability to contextually analyze hundreds of millions of data points and associate them to particular clusters of variables is fairly new and certainly has come about because of advances in AI. Valuations are the core of real estate because so much emanates from the simple question, “What is this house worth?” While AVMs and the entire valuation industry took a hit with the rhetoric emanating from Zillow as they crashed and burned with Zillow Offers, this was the unfortunate byproduct of a bad business decision and as such cast unfair doubt on AI.
While primacy should be adapted to each financial institution’s needs, the industry needs standardization, which can only start with data. One of a bank’s greatest assets is the intelligence they have from a client’s transaction data.
Presented by: William Mills Agency
Computer vision allows us to enter the heretofore closed-off portal: The inside of the house. With device proliferation and fantastic upgrades in the power of edge computing and “AI on the edge,” true condition-adjusted valuations and service-offerings can be considered legitimate for the first time.
Finally, AI has helped make certain elements of the transaction process better — in an incremental and “point solution” fashion.
Still, the promise of AI remains in the domain of “potential.” For the most part, real estate is very much a human-led and still error-prone ecosystem. Buyers and sellers do not meet in an open market characterized by full transparency and instantaneity. Cumbersome processes and vested interests continue to play an out-sized role in residential real estate.
And with all the hubbub about “democratization” — a sine qua non of the AI story — homeownership has over the last five years become an impossible dream for about one-third of Americans. On these accounts, AI has failed to live up to the hype.
We all need to do more tire-kicking. We need to understand the exact offerings and their precise business value before rejoicing about the advent of AI. The cycle of hyperbole must end and we have to demand that those who say they are “game-changers” better really change the game. So far, some have, but most have not.
As such, for now, the game remains largely the same.