A recent study by the McKinsey Global Institute estimates generative artificial intelligence could generate between $110 billion to $180 billion in value for real estate in nearly every area of the industry. The dozens of different applications are only just beginning to show how versatile this tech wave will be. With all the real estate tech trends I’ve seen come and go this one feels different. It has the nimbleness the industry requires for the breadth of use-cases and it returns its output quickly so agents can adjust to the constantly changing landscape of buying and selling homes. AI just might have the staying power to become part of industry norms. With that in mind, I’ve looked at a few different ways it is being used that go beyond what we’re already used to using. This is by no means an exhaustive list, but aims to highlight some of the more cutting-edge examples just beginning to take hold.
Virtual staging, or the practice of digitally adding furniture or a perfectly roaring fire to a house picture has been around for years. This same concept of digitally altering photos has become much more sophisticated recently in real estate, such as a tool called REimagine Home AI that can “declutter” rooms by deleting images of objects in a photo. It can also help homeowners imagine what a renovation would look like by taking a picture of an old, rundown house and sprucing up the exterior on a computer screen so someone can try out different looks for the facade. One of the largest multiple listing services in the country, California’s CRMLS, recently announced a partnership with the company. Another platform called Revive Vision AI takes it one step further by analyzing pictures to give you a detailed outline of the renovation costs and an estimate of the profit potential for each project. It can also provide an automatic valuation of the property based on the condition it analyzes from the pictures.
Similar AI image analysis is being used by insurance companies to assess property damage or vulnerability risks. CAPE Analytics, which was recently acquired by Moody’s, is one of several examples that uses a combination of property photography, aerial images, and weather data to predict the likelihood of a natural disaster (see picture below). CoreLogic, which powers many of the multiple listing services around the country, has added AI tools to analyze pre- and post-disaster images to enhance their reporting.
The intersection of AI and photography has also reached natural language searches. A company called Restbi.ai automatically tags photos to a greater degree of granularity than you would typically find in the property description. It can automatically detect everything from natural light to stainless steel appliances to the style of the home’s exterior. So if you want to search for all the Tudor style homes in a certain area you would be able to find them even if it wasn’t listed in the description. The major search portals like Redfin, Zillow, Homes.com, Trulia and others have started incorporating language models to make searching for a home feel like you are “talking to a friend.”
Search sites often combine this with other predictive AI tools to enhance the parameters visitors to their site use to search for homes. Redfin found its Redfin Matchmaker tool led to users clicking on homes four times more often than ones from their own saved searches.
Before a homeowner even steps foot in a house there are AI tools being put to work at the very beginning of a purchase process. Compliance with national Fair Housing laws and the hundreds of local laws and guidelines has been a challenge for the industry to find an automated solution for. A new attempt by a company called Ocusell uses AI to “learn” the rules of a multiple listing service and scans what agents enter into a listing form to catch potential violations before they go live. Their AI platform has generative capabilities to create descriptions without the risk of running into a compliance violation in the areas where there is more free-form content.
Lenders are using AI to streamline the mortgage process too. Rocket Mortgage is just one example of a company that uses it to scan the mountain of paperwork involved in an application to pull out relevant information. They also have an AI tool that listens into phone conversations with a rep and can automatically input the necessary data as the customer says it out loud.
With all this talk of artificial intelligence questions about fraud are sure to follow. Mortgage fraud seems to be on an upward trend over the last decade, but AI tools are catching up quickly. Commonwealth Bank of Australia decreased its customer-reported fraud by 30% after implementing generative AI. Its tool flags thousands of suspicious transactions and proactively sends an alert to their customer before anything can proceed.
This is just the tip of the iceberg when it comes to how far-reaching AI can become within real estate. Looking for a home in your price range with a porch that gets great morning sun? It’s just a matter of time before you’ll be able to type those words and the AI engines will start whirring away to scan databases for homes with porches and then, knowing the geo-coordinates of the home, determine the angle of the sun throughout the year, the density of tree cover, the average cloudiness and any shade created by the surrounding buildings. Minutes later you’ll have an answer. A few minutes after that you might have a mortgage approved and the moving van on its way.
Read the full article here
