The CEO of legal tech firm LawGeex — which helps automate contract review — lays outs how its new AI patent could turbocharge demand


Summary List Placement

LawGeex, a startup that automates the contract review process, has patented its artificial intelligence technology — a milestone that signals growing comfort with machine learning in law and could accelerate growth for the business. 

The AI analyzes documents based on the contract’s context with “the mind of a human lawyer,” according to Noory Bechor, the company’s co-founder and CEO. This contextual approach takes into account surrounding factors like company policies, the document type, and whether the client is the buyer or seller.

Risk-averse lawyers and their clients can have millions of dollars hinging on specific wording or a single clause in a contract, so accuracy is top of mind. However, humans can actually only reach about 85% accuracy, according to Ilan Admon, co-founder and CTO of LawGeex.

“If you put ten lawyers in a room, they often don’t agree on everything,” added Bechor. These differences in interpretation prolong contract negotiations, leading to time and money wasted.

Founded in Tel Aviv in 2014, LawGeex is used by large investment banks like UBS and law firms like Simpson Thacher to automate contract review. It’s since raised a total of $45 million in funding from investors like Aleph, Corner Ventures, and Lool Ventures.

LawGeex’s patented technology boosts the accuracy of contract review — anywhere from a 90% to 100% accuracy rate, depending on how common the contract type is — speeding up the closing deal time by threefold. This translates to 80% time saved reviewing and approving contracts and 90% cost saved compared to manual processes, per the company.

The patent not only serves as a “proof point” for the AI’s prowess but also advances LawGeex along its broader plan for growth in the contracts space, said Bechor.

The five-year brainchild

It took five years for LawGeex to have its patent approved. The initial resistance came from the fact that the United States Patent and Trademark Office said that the tech mimicked human behavior.

“The reviewer decided that it’s a force of nature, that we’re building a brain. And you can’t patent thought,” said Ilan Admon, co-founder and CTO of LawGeex. “We initially had trouble explaining that it’s not a brain, it’s an algorithm — it just works the same way.”

LawGeex’s AI is pre-trained by screening thousands of documents of different types, from non-disclosure agreements to sales contracts. The technology mimics what a lawyer would do by looking at the document at a high-level and deciding what the context is. It’ll then use the algorithm best suited for that context to drill down into each provision of the contract.

“Think of the algorithms as different glasses: You have one for near-sightedness, another for reading, and so on. Each algorithm switches depending on the context of that particular contract,” explained Admon.

Over the course of the five years it took for the patent to be approved, LawGeex continued to hone its technology and improve its accuracy.

Differentiating from other tech

Although there are open-source tools for language analysis, like Google Cloud’s AI, LawGeex’s patented contextual tech is tailored to the specific way lawyers think.

“We originally thought …read more

Source:: Business Insider


(Visited 7 times, 1 visits today)

Leave a Reply

Your email address will not be published. Required fields are marked *