Lease Abstracts: Artificial Intelligence or Document Automation? What’s the Difference?
Few CRE tech buzzwords command more investor interest than Artificial Intelligence (AI). It’s the new hot thing. And while AI is an undeniably exciting technology, does it deserve the hype it gets? For some applications, the answer is unequivocally, “yes.” For others, AI may be more trouble than it’s worth. In the case of lease abstracts, the answer is, “maybe.”
There’s probably nothing more loathsome than finding all the key details buried in a 60+ page lease and formatting them into an abstract. Yet, since a lease abstract is such an integral part of business operations, it’s something that must be done. Tell a lawyer or leasing admin that there’s an easier way to abstract a lease, and there’s a good chance you’ve just won his or her undivided attention.
In a nutshell, AI approaches the abstracting process by brute-force data extraction and analysis. Armed with programmatic logic and a language algorithm, AI-based lease abstraction software approaches the abstracting process by tirelessly combing through every word, clause, sentence, and section of a lease in order to make sense of it. After some intensive data-crunching, AI software delivers a report of its findings which, hopefully, offer an accurate summary of a lease. It’s important to note that AI still has a ways to go before it can produce abstracts independently; it’s not completely error-proof (and may not ever be). To ensure total accuracy, the results will require some degree of human audit.
A good rule of thumb is that AI will generally suss out 80% of the key info in a lease. The rest still requires manual review by an expert. Despite the ‘babysitting’ AI-based lease abstraction requires, it’s a step in the right direction.
While it may not be perfect, AI as an abstracting solution is well-suited to address the problems companies face in two types of situations:
- A company that’s just acquired an asset and inherited the existing leases
- A company that needs to collect and parse data from a library of already-existing but loosely organized leasing agreements
For a CRE company facing one or both of these challenges, AI promises to be a time-saving solution. But when it comes to working with new leases, AI is an unnecessarily complex solution that still requires human oversight. In essence, it’s a band-aid solution to a latent problem inherent in the leasing process. It doesn’t address the root cause of the problem: an antiquated drafting workflow.
A Different Approach
When a lawyer looks at a lease in MS Word, s/he doesn’t need to read from beginning to end to understand it. S/he has the training and experience to see the interconnectedness and dependencies (e.g. if A changes, then B and C must also change) that form the lease’s underlying structure. That’s more or less how AI approaches the lease, too: over time it ‘learns’ the structure of the lease.
But, as we’ve already noted, there’s an inherent logical structure that forms the underpinnings of every lease (it’s not a James Joyce novel with layers upon layers of obfuscated meaning and endless scholarly debates). Leases are more or less straightforward and predictable. If this is the case, does it really make sense to approach the abstracting process with a brute-force learning algorithm?
That’s a rhetorical question. Of course it doesn’t.
Traditional, logic-based software excels at tasks with a high degree of predictability. So rather than try to build a solution for lease abstraction that fits into an inherently broken process, we designed LeasePilot to eliminate the shortcomings in the leasing workflow from the start.
LeasePilot is aware of the interconnectedness of every part of the lease. Simply put, if one part of the lease is modified by the user, LeasePilot will update/add/remove language elsewhere in the document if necessary.
For example, if a landlord isn’t able to complete a credit review of the tenant before sending out the first draft and subsequently determines that guaranty of the lease is needed, adding one would modifications to several different parts of the lease: provisions such as default, notice, and financial reporting would need to be modified, and a completed form of guaranty needs to be added as an exhibit. LeasePilot handles these changes instantly and automatically; all the user has to do is check a box and input guarantor details one time. LeasePilot modifies the lease is accordingly.
But sometimes, a lease needs to be changed in a unique way that deviates from the standard terms. In these cases, LeasePilot’s in-browser text editing gives users total control over every sentence, word, and character that appears in the lease while still maintaining the underlying data integrity. It’s this combination of automation and granular control that sets LeasePilot apart from other more generalized document automation solutions.
When a lease is built from this kind of programmatic, structured data, generating an abstract is a trivial task. In many cases, the abstract isn’t even necessary. The important information from the lease can be automatically injected into other software systems like MRI, VTS, and others.
Abstracting: AI vs Smart Automation
At the end of the day, both approaches to lease abstraction are an enormous improvement over the status quo. As noted previously, AI excels in situations where the lease is already executed. On the other hand, software like LeasePilot is built with a different approach in mind; one where data about the deal is tracked and organized throughout the entire process. In other words, AI is a reactive approach to solving the problem of abstracting, and LeasePilot defines a proactive approach. Both options serve different needs.