MIT Spinout Aims To Automate Contract Law

When you think of automation of certain tasks, you probably think of industrialized factories or the driverless car, but the legal industry has been one of the more interesting examples of automation.  I’ve written previously about attempts to automate e-discovery and even use AI to predict the outcome of cases.  The latest project to use AI in the field is a spinout from MIT, called Klarity.  The MIT service is aiming to make nondisclosure agreements easier to check and sign off on.

“There are only a few pieces or items that you care about, but there’s a labyrinth of clauses, so you don’t know what will trip it up,” the company say. “We decided to use natural language processing and AI to accelerate review.”

A process that usually takes days can now be done in minutes, with the AI underpinning the system determining whether the agreement should be signed, renegotiated or referred to the legal department.

Determining value

The system has been trained to accurately identify the terms in the NDA that are of crucial importance, both to spot it in the text but also to tell when it’s missing.  This enables the system to provide a risk profile for the contract, alongside individualized risk profiles for clauses within it.

“We ask our clients when deploying, ‘What’s your policy? What are the terms you are and are not willing to agree to?” the team say. “We take this policy and configure the system for the particular client. Klarity looks at contents of agreement in relation to company policy, then creates a risk profile.”

The company have already attempted to find their way to market, but after initially targeting law firms the company moved their focus towards helping companies manage the various contracts they have to sign, and especially the NDAs that startups commonly have to deal with.  This is typically the first document signed in any enterprise software sales process, so the team believe they have an instant opportunity to provide value, both in a financial sense but also in terms of showcasing the potential of the platform to automate a significant part of the legal process.

“We want our clients to trust our system; it’s a way to build that trust,” the team say.

Facebooktwittergoogle_plusredditpinterestlinkedinmail