The Short Version

• AI tools can help commercial real estate professionals review documents, visualize property information and identify potential issues during due diligence.
• Orbital’s Justin Lischak Earley argues that purpose-built AI may be better equipped than general-purpose models to handle specialized real estate documents, terminology and geographic nuances.
• Earley sees AI as a tool for making due diligence faster and more efficient, not as a replacement for the legal and real estate professionals responsible for assessing risk.
• Specialized systems could help transaction parties establish a common set of facts more quickly and focus their attention on material issues.

Specialized AI Tools Could Help CRE Professionals Conduct Due Diligence More Efficiently

AI is changing the landscape of due diligence in commercial real estate. The best AI for the job, however, may not be the AI that you can get off the shelf.

Artificial intelligence is rapidly changing how legal and real estate professionals handle the document-heavy work behind transactions. According to Justin Lischak Earley, head of real estate innovation at Orbital, the value of AI lies in augmenting — not replacing — the expertise of lawyers and real estate professionals. The key is to use AI built for the task.

Orbital, founded in 2018, provides AI designed for real estate legal and property workflows, connecting documents, maps, historical deeds and other property information to help professionals conduct due diligence and identify and manage risk.

Earley, who has worked as a lawyer at the intersection of real estate and tech, spoke with Commerce + Communities Today contributing editor Halley Bondy about how AI is reshaping real estate due diligence. He explained why general-purpose AI models can fall short in commercial real estate and how specialized technology could help shopping center owners, developers and their legal teams move transactions forward with greater speed and efficiency.

Where do you see the biggest opportunity for AI in commercial real estate?

The most obvious place is due diligence, especially around land titles. You also see large amounts of due diligence work when you’re dealing with more property or more complex property. Take assemblages, where people are putting together different parcels of land to create a larger project. You have to make sure there are no gaps between those pieces. In legal parlance, we call those [strips of land] gaps and gores. I’m seeing a lot of progress around the ability to mathematically and visually plot those things to make sure the pieces fit together.

How has the landscape changed from recent history?

When someone is developing a new center, often those are assemblages. Or, if they’re not assemblages, there are outparcels being carved out for the fast-food restaurant, bank or whatever else is going there. There are often a series of reciprocal easements and rights that need to work together for that shopping center to function operationally the way you want it to. The ability to visualize where those things are — or will be — gives a developer much more visibility and comfort than would have existed when I was a young lawyer. I remember doing shopping center deals where there were hand-drawn maps showing where access and maintenance easements were going to be on the common area. We’re on the verge of advancing so far beyond that, with software helping us in the development process, that a lot of the legal and business risk around those issues can be drained out of transactions.

What does purpose-built AI understand about a commercial real estate transaction that a general-purpose AI might miss?

AI is largely built to mimic what humans do. If you have a lawyer who handles family law, criminal law and whatever happens to walk in the door, you’re probably not going to send them a complicated commercial real estate transaction. You go to a specialist. There are many lawyers who, like me, are commercial real estate lawyers. That’s what they do. Why? Because specialization has produced better results for clients. I don’t know why, given that humans have gone in that direction, AI would be any different.

How does AI actually work in the collaborative due diligence process in a commercial real estate transaction?

[For Orbital, for example], Parties upload the documentation that they want our product to analyze. ... The AI becomes another co-worker that they give direction to, in much the same way that they give direction to a junior associate.

In a real estate transaction, at least any commercial real estate transaction of size, you’re going to have a number of different parties who first and foremost need to have a common set of facts off of which to work. So if I am the buyer of a center, the seller and my lender are all going to need a common set of operative facts on which we can negotiate. ... It’s one thing to be able to grapple with text, see where text has changed, and it’s another thing to be able to know whether that change matters. ... People are beginning to infuse legal knowledge and legal reasoning into those machine models to be able to say not just, “What has changed and what is different?” but, “Does it matter?” To spot not just a change but a material change.

What should commercial real estate professionals understand about AI’s role in de-risking transactions?

Commercial real estate is a very risk-averse industry, and we believe in doing our homework. That is important to keep in mind when we talk about AI. Early in the technology revolution around real estate, there was a lot of discussion about whether we could simply predict the likelihood of a bad outcome well enough that we could decide not to do all of the due diligence. I think that was the wrong approach. It fails to understand the circular nature of how due diligence de-risks a transaction. The reason a bad outcome may be unlikely is often because we’re doing the homework in advance that helps prevent it from happening. You can’t say, “The probability of this problem is only 5%, so let’s not spend the time looking for it,” because the process of looking for those problems is part of what makes the transaction safer in the first place.

What I’m happy about now is that we’re getting tools that can help us do that de-risking work better, faster and easier. AI can help identify issues, analyze large amounts of information and get the parties to a common set of facts more quickly, but that doesn’t mean we’re eliminating the due diligence. We’re making the due diligence more efficient and, ideally, more thorough. It’s like long division. Do you want to do your long division by hand with pen and paper? Not really, but you still need the answer. We’re still doing the de-risking due diligence; we’re just doing it with better tools.

Where does the current generation of AI fall short?

Foundation models are based on data. If you don’t have the data, AI can’t discern the patterns. Commercial real estate transactions have very interesting, specialized and ultimately limited data. There are real estate-specific documents, terms and phrases that you aren’t going to find in typical training data. There are also state-specific nuances and sometimes city- or county-specific nuances. Commercial real estate practitioners know those things, and they know what they don’t know. An AI model doesn’t know what it doesn’t know in that sense. [Generative AI] foundation models don’t necessarily have that lingo or those geographic nuances.

How will AI change the relationship between attorneys and their real estate clients?

What I see AI doing is enabling lawyers to get to the work that really matters for their clients much faster. It takes the issue-spotting and the process of separating what’s material from what’s immaterial and moves that grunt work to a machine. That’s work that all of us as lawyers should be very happy to pass off. AI is also good at spotting patterns and enabling you to leverage those patterns to do work better, faster, easier. The human operator element very much matters because at the end of the day, AI is a tool. Even the best tool in the hands of someone who doesn’t know how to use it isn’t terribly helpful, right?

What are the biggest misconceptions commercial real estate professionals have about AI?

That it’s magic. There’s a psychological tendency to anthropomorphize AI. I sometimes call it the dancing bear fallacy: If you see a bear dancing, you think, “Wow, it’s trying to express some emotions.” No. It just does it because it was trained to do that. AI systems are great communicators. They’re eloquent, verbose and convincing because that’s what they were trained to do. They’re a tool.

The other misconception is that prediction is the same thing as reasoning. These systems are incredibly good at complex statistical analysis that predicts what should come back. That’s why having these tools in the hands of people who are skilled at using them is so important and why using the right tools for the job is so important. Something that is fit for use isn’t necessarily fit for purpose.