AI deserves more attention and perhaps more cautious adoption.
Real estate has gone virtual as more companies incorporate artificial intelligence into several stages of the buying, selling, and mortgage processes. Companies like Compass, Zillow, and LoanSnap are using AI to assist buyers in finding the right mortgage and the perfect house.
Although most real estate information is available to the public, sifting through public property records to find title documents, purchase prices, and even mortgage liens can be an arduous task. It often requires record seekers to travel to local offices and physically look through filing cabinets. Many real estate companies believe that artificial intelligence eases the home buying process.
This time-consuming operation has been simplified with the help of computer algorithms that can process millions of records in seconds. Artificial intelligence helps real estate professionals access property values, debt levels, home renovations, and even some of a homeowner’s personal information in the blink of an eye.
While several real estate companies now rely on AI technology, Zillow is at the forefront with its Zestimate method. A Zestimate uses AI, machine learning, and neural networks to analyze property data, such as sales transactions, tax assessments, public records, and home features such as square footage and location.
What is the Zestimate algorithm, and how does it work?
The Zillow estimate — also called the Zestimate — is an algorithm that uses unique valuation models based on information about the property, tax assessments, and previous and current transactions to calculate an automated appraisal for a property.
Zillow launched Zestimate in 2006, and today, it tracks the value of over 100 million homes. It is different from a traditional real estate agent’s comparative market study because a Zestimate is based on all the data in a county. In contrast, a conventional realtor determines a home’s worth based on data inside the home’s area.
Before its recent updates, Zillow determined the value of homes using nearly 1,000 different algorithms created for local markets. Zillow now uses a single neural network model to correlate home facts, location, housing market trends, and home values with property data submitted by property owners, real estate experts, or public sources.
Zillow started to develop a single neural network in 2019 when it launched a public competition to make needed improvements in its estimates. More than 3,000 teams competed for a $1 million prize. Today, Zillow’s neural network-based model allows the site to more frequently updating its estimates of the value of 104 million properties in 23 US markets.
The Zestimate for on-market properties has a nationwide median error rate of 1.9%, while the Zestimate for off-market homes has a median error rate of 6.9%. Zillow says the new algorithm will reduce its price estimate errors by 11.5% for off-market homes in nearly 30 regions across the US.
Zillow and other iBuyers are making cash offers on houses
In 2018, Zillow introduced its Zillow Offers program, a direct cash-offer homebuying business. Zillow Offers will buy a home in cash, streamlining the process significantly. Zillow will then manage any repairs or quick renovations and then sell the house itself.
Zillow is not the only real estate technology company doing this. Opendoor, RedfinNow, and Offerpad operate similar businesses and are sometimes called instant buyers or iBuyers.
With the help of data and AI, iBuyers make predictions of specific real estate markets. This AI-powered look into the future helps them make attractive offers on off-market houses directly to homeowners. These offers avoid the need to pay an agent or prepare a home for viewing, making them especially appealing to homeowners.
While iBuyers simplify the home buying and selling process, some believe an already volatile real estate market will not benefit from tech or investment conglomerates getting involved. Others worry that iBuyers are overpaying for off-market houses, which drives up property values.
For now, iBuyers make up a sliver of the real estate market and are still too small to manipulate the market significantly, but the skeptics aren’t entirely wrong. In November 2021, Zillow announced it was winding down the Zillows Offers program and was looking to offload over 7,000 homes it had bought, not to individual buyers but investors. Zillow estimates that it will lose over half a billion dollars in value on the houses it owns.
Despite the failure of the Zillow Offer program, other iBuyers are finding success. Based on their real estate transaction volume, the biggest iBuyers are Opendoor and Offerpad. Opendoor reported purchasing 15,181 homes and selling 5,988 homes in the third quarter of 2021. Offerpad reported purchasing 2,753 homes and selling 1,673 homes in the same period.
Artificial intelligence is not accomplishing more than traditional research. Arguably, its capacity to accelerate processes may sacrifice needed due diligence measures. However, it seems likely that in the fast-moving and ultra-competitive real estate market, its role will only grow as it “learns.”