A version of this article first appeared in the CNBC Property Play newsletter with Diana Olick. Property Play covers new and evolving opportunities for the real estate investor, from individuals to venture capitalists, private equity funds, family offices, institutional investors and large public companies. Sign up to receive future editions, straight to your inbox.

D.R. Hortonthe nation’s largest homebuilder, is tapping an artificial intelligence tool from Portland, Oregon-based startup Prophetic to build more homes and address the country’s housing shortage.

Chronic underbuilding since the Great Recession has caused a deficit of roughly 4 million homes, according to analyses from several sources, including Zillow. The supply-demand imbalance has caused prices to rise over 50% from pre-pandemic levels.

Homebuilders are trying to respond but say that the cost of construction, along with the difficult and costly process for acquiring and developing buildable lots, is making that difficult.

“One of the largest challenges to providing affordable housing is the identification, acquisition and entitlement of land suitable for development. We are confident the insights provided by Prophetic are going to help us expand homeownership opportunities for hard-working American individuals and families,” said Jason Jones, vice president of data analytics at D.R. Horton, in a release.

Prophetic has developed an AI-native platform for land acquisition and development analysis. For any potential parcel of land, Prophetic’s software will pull every single zoning manual from every city and county in a state. The company said it is currently operational in 25 states and expects to be in all 50 by June.

“It’s an incredibly large, tedious, detail-oriented process to take tens of thousands of these zoning documents and extract the rules, not only efficiently, but correctly,” said Oliver Alexander, founder and CEO of Prophetic.

Among other things, the system looks at minimum lot size and minimum or maximum density setbacks, which differ by municipality and zone. It updates those quarterly.

“Then it tells you where that information came from, which is the key differentiator,” Alexander explained. “When you have that section title and the page that it came from, that builds trust, and then it becomes ultra-efficient, where you can analyze development potential in 30 seconds instead of two to three hours.”

Alexander said there are a little over 440,000 different ways to describe what you’re allowed to do on a piece of dirt in the states Prophetic has analyzed. Developers need to go through all of that information to figure out if they can build a single- or multifamily housing development on it.

The AI’s large language model-based analysis of these documents at scale can answer the questions and then feed that into search AI, which Alexander calls “the major unlock” – search plus the zone AI information together. At the ground level, with this AI, builders can figure out what they can build, where and how much at a much faster pace, making them more competitive with landowners.

“If you have that much of an edge in your speed to decision, you effectively control your entire market, because before anyone else can decide, you’ve tied it up,” said Alexander.