“Our job is to help agents grow their businesses, make more money, save time, and create great experiences for their clients,” says Rory Golod, president of growth and communications at Compass. 

The investments by Compass, the nation’s largest brokerage by sales volume, include “Likely to Sell,” an artificial intelligence tool that analyzes prospective client leads and makes recommendations on who may soon be ready to sell their home. More recently, the company rolled out Compass AI, a chatbot tool that can help write listings for properties, marketing materials, and agent profiles. 

Agents can have thousands of contacts at their fingertips, Golod says, and conversation rates are often exceptionally low for traditional marketing tactics like email and social media. But Compass says that since Likely to Sell launched in the summer of 2020, nearly 8% of recommendations given through the CRM (customer relationship management) tool each month are listed on the market within 12 months.

“We want to use AI to help an agent say, ‘If I’m going to reach out to someone, I want to reach out to the people who have the highest propensity to maybe transact,’” says Golod. 

There’s an estimated $180 billion in value that could be unlocked from the advancements of generative AI, according to McKinsey estimates. The industry could certainly benefit from such a jolt, as U.S. home sales declined to their lowest level in nearly 30 years in 2023 due to lofty mortgage rates and low inventory that has made home buying a lot more expensive. The industry is due for major disruption to commission fees after the National Association of Realtors struck a deal that could lead homebuyers and sellers to negotiate lower agent commissions. There is also a major problem with construction, because the nation just isn’t building enough new homes to meet demand. 

But there are a lot of complicating factors that make AI adoption within real estate especially difficult. Experts say there are massive amounts of unorganized data, ranging from leases to contracts, from investment documents to design plans. Construction operates on razor-thin margins. The median age of a real estate agent skews older than that of workers in most industries, and those in the sector are notoriously tech-averse. And the highly physical nature of the industry means many technology advancements are still in their relative infancy. 

“I would say, historically, real estate has always been a bit of a laggard, in terms of use of AI,” says Alex Wolkomir, a partner at McKinsey. 

Wolkomir says commercial real estate is further along than residential when it comes to AI adoption. The top challenge he thinks the industry faces is making sure members of its workforce—construction workers, real estate agents, designers—are properly trained and understand the capabilities of the AI tools they are given. He is encouraged by some forward momentum in the industry’s AI journey over the past five years.

“I think a lot of the [generative] AI use cases are kind of opening up new areas that are very valuable to real estate,” says Wolkomir.

Yao Morin, chief technology officer at JLL, says one of the challenges that commercial real estate faces is the abundance of unstructured data, in the form of leases, contracts, and invoices. “I believe in this era of AI, the barrier of using AI will continue to go lower,” says Morin. “And then you ask yourself, ‘If using AI is not a competitive advantage, then what is?’ The answer is absolutely your data.”

Last year, the company unveiled JLL GPT, a generative AI model that provides insights to clients based on JLL’s proprietary market research, alongside externally available market data. Morin says 20% of JLL’s 103,000-person workforce is using JLL GPT on a weekly basis because the technology enables staff to complete repetitive tasks more efficiently.

JLL is also using generative AI to better predict building-maintenance needs, research investment opportunities, and implement sustainability initiatives. “If you think about classic AI, it takes a higher learning curve to understand it and be able to trust the results,” says Morin. “But with generative AI, it is much easier for us to adopt and for people to see the value.”

The startup Higharc has launched a homebuilding automation platform that aims to turn home construction into a faster and more affordable process. 

“What we do is make data available about houses that are going to be built,” says Higharc CEO Marc Minor. “And when I say, ‘make data available,’ I mean every part and piece of the building, and where it belongs, and when it needs to be built, and who’s responsible for that segment of the building. We control all of that information in an automatic way.”

Last month, Higharc raised $53 million in Series B funding, including from retailer Home Depot, the venture arm of France’s Schneider Electric, and others across the construction, building products, and manufacturing industries. Minor says the greatest possibilities lie in both improving the way homes are built alongside access to data from distributors and suppliers.

“If you build the right software layer to systematically change housing, in terms of the designs of the houses, that gives you the capability to more easily understand the ways to leverage the hardware side,” says Minor.

Prologis Ventures, founded in 2016, has invested $250 million into over 45 supply-chain and logistics-focused startups, including AI-enabled companies like TestFit, Altana AI, and Logiwa. 

“People have always used intuition as a way to make real estate decisions,” says Will O’Donnell, managing partner of Prologis Ventures. “But there’s just reams and reams of data that if you could pull together and do analysis on [it], you would be able to have better insight to make that decision.”

Prologis, as an example, uses TestFit’s AI to better judge the feasibility of new warehouse sites. Information on specific zoning regulations, environmental conditions, transportation around a site, and labor can be integrated to improve decision-making. TestFit can also produce dozens of project renderings in as little as an hour and will make suggestions based on past metrics. 

“As a company, one of the things that we’ve been spending a lot of time on is, what information is important to our customers when they make a decision?” asks O’Donnell. “What’s important to them as they’re driving their business, and how do we empower both our people and our customers to better understand that information?”

Augmenta, meanwhile, automates designs for electrical systems, meaning all the parts and pieces inside a building that take electricity from one point and deliver it to another. “The process of ideation to the full, detailed plan is fraught with challenges,” says Francesco Iorio, cofounder and CEO of Augmenta. 

The design process, Iorio explains, is extremely complex, because to go from sketch, to a list of parts, to construction, there is a long list of considerations before there’s an actionable plan to construct. The greatest benefit of AI, he says, is to automate the preconstruction phase for electrical systems. 

“Giving them the ability to design at the highest level of detail, with cost and time being at the forefront very early in the design stages, allows people to experiment and answer those questions that would be costly to answer downstream,” Iorio says.