The Short Version

• CoStar is applying AI to its database of 1.8 million retail properties to help brokers, landlords, retailers, banks and other users search, crunch and analyze commercial real estate data instantly.
• CoStar acquired Visual Lease in 2024, to boost CoStar Real Estate Manager. Now, the lease administration, lease accounting and portfolio-management solution employs AI to extract information like rent schedules, renewal dates and financial terms and clauses for review and reports.
• CoStar’s year-old tenant portal tracks retailers with five or more locations to reveal retailers’ real estate footprints, portfolio sizes, transaction activity and lease expirations.
• Svec expects a large language model to sit on top of CoStar’s data so clients can pull up analytics and charts or query it, such as requesting every TJ Maxx with a CMBS loan coming due within two years.

Data Powerhouse CoStar Is Using AI to Search, Crunch and Analyze Information Instantly

Forty years after CoStar Group’s founding, any commercial real estate stat that brokers, economists and journalists cite — be it real estate vacancy, rent growth or capitalization rates — likely includes an attribution to the data and analytics firm. CoStar’s own in-house researchers, economists and reporters leverage the company’s reservoir of data to analyze local, regional and national market property trends from construction to investment sales activity. CoStar’s retail property database alone encompasses 1.8 million assets totaling nearly 10 million square feet, and it includes all manner of fundamental and financial metrics, said Brandon Svec, the company’s national director of retail analytics. Now, CoStar is leveraging AI’s ability to search, crunch and analyze data instantly for Marketplaces Industry clients from brokers and landlords to retailers and banks, Svec said.

Some of those clients also use lease administration, lease accounting and portfolio-management solution CoStar Real Estate Manager. To expand the tool’s capabilities, CoStar in 2024 acquired Visual Lease, another lease management and accounting solution. More recently, it added to both products AI-powered lease abstraction to extract information like rent schedules, renewal dates and financial terms and clauses from leases and organize it for review, including links to the underlying lease language. It also added to both products an Ask AI chatbot that lets users query that data in plain language and get quick answers. Users then can create reports summarizing key business, financial and legal considerations, among other benefits, Svec said. And they can use CoStar’s market data and analytics alongside the product to help make decisions about their portfolios, he added.

“The existing ethos of being the only real census-based database of all retail inventory and all the data underlying that inventory in the U.S. is still at the heart of CoStar,” Svec said. “But we’re creating filtering tools, additional slices of data and some pretty cool AI enhancements to help retail users, brokers and owners make all of our information more actionable.”

Svec spoke with Commerce + Communities Today contributing editor Joe Gose about AI-powered expansions to its products and services.

What are some client initiatives that CoStar has introduced recently?

Eighteen or so months ago, we began to include [commercial mortgage-backed securities] documents on all applicable retail properties in our database. That information includes the property’s financials, who the tenants are, rent rollovers, loan status, servicer notes and maturities. It also includes new issuances. It has become a great additional data source for our clients who are trying to understand the financing underneath retail assets.

We also introduced a portal about a year ago that lists tenants with at least five locations to provide a more complete view of a retailer’s individual locations and broader real estate footprint. The platform represents a substantial enhancement of our ability to track and organize what is a very large and fragmented universe of properties and tenants. At the company level, clients can evaluate locations, portfolio size and composition, transaction activity, upcoming lease expirations, news and other information. At the individual-location level, users can review a tenant’s occupancy and transaction history and, where available, information such as rents and escalations, concessions and tenant improvement allowances, among other lease terms.

Before we introduced the portal, we had about 200 people who were solely devoted to the task of collecting that data and we were comfortable conducting Google and other web searches to pull in data for clients to access. But in addition to better organizing and continuously updating that information, AI’s ability to monitor activity has helped our researchers better leverage their time. What used to be a part-automated and part-manual process has become an extremely automated process that enables humans to review and track the entire universe of tenants at a scale that simply wasn’t possible before.

It sounds like efficient collection of data collection has freed up resources. How is CoStar taking advantage?

We are devoting more time to finding new ways to provide our data to clients. Last quarter, we launched a report specific to the grocery-anchored retail sector, where clients can find information on how the category is performing and in-depth profiles of the 15 largest grocers within the country, what their portfolios look like and what their plans are. We think it’s a new and different way to look at what is a very important segment in the retail market. That was introduced on the back of a new net lease report that is breaking down all of the trends in that category.

This quarter, we’re launching another report focused on retail leasing that will give clients a much more comprehensive view of leasing activity, including where leases are being signed, the types and sizes of the spaces being occupied, which tenants and sectors are driving the market and how activity differs among local, regional and national tenants. We’re also systematically tracking retailer announcements, earnings reports and other public disclosures to build an across-the-board view of store openings and closures and how retailer footprints are changing. That work will ultimately support a dedicated CoStar store-opening and -closure report.

What’s the next innovation CoStar users might expect?

I think there will be a time in the near future when an AI large language model will effectively be written over CoStar and our clients will be able to prompt it to extract information from a database or pull up analytics or charts or find all the properties that have a TJ Maxx with a CMBS loan coming due in the next two years. In other words, clients will be able to do any of the things that they can do right now, but they’ll be able to do them faster when they have a bot working on their behalf.