Using Artificial Intelligence for Legal Contracts
Human history is full of bad deals. We’re all familiar with one of the world’s worst dealmakers whose ambitions for power and pleasure are famously nearly limitless. We’re talking, of course, about Faust, who trades his soul for magic and mischief. It’s really a bad deal considering that he gets a few years of fun in exchange for eternal torment. What Faust could really have used was an expertly drafted contract that guaranteed his damned soul certain rights, such as not being buggered by the thorned phallus of a cacodemon for all eternity. Yes, that’s how we’re introducing today’s topic on using artificial intelligence for legal contracts, because dealing with some attorneys must be what it feels like to be sodomized by Satan.
Some of the first legal contracts appear in ancient Mesopotamia, covering everything from the sale of a slave for 10 shekels to the one-year rental of a Babylonian house with a nice view of the Euphrates River. Today, contracts dictate our online life through countless user agreements that we blithely click through, not to mention the stacks of paperwork that international corporations must manage to operate their financial empires across vast supply chains. If only there was a way to automate legal contracts, right?
Using Artificial Intelligence for Legal Contracts
We made our first dive into how artificial intelligence is changing law practices a couple of years ago. At the time we noted that the $437 billion U.S. legal services market and the 1.3 million attorneys in this country aren’t going away just because of artificial intelligence and a few handy chatbots. Instead, AI will pick up the slack – initially – in tedious, repetitive jobs where it can be more efficient than humans. That brings us to legal contracts.
In an excellent analysis in the Harvard Business Review on how AI is changing contracts, writer Beverly Rich notes that the main challenge is the sheer number of contracts that law firms must track, many of which “lack uniformity” and are “difficult to organize, manage, and update.” That sounds like a job for natural language processing (NLP), a field of AI that deals with how machines understand (NLU) and generate language (NLG). The combined technologies help AI extract and understand unstructured and structured data. Part of the goal is for the machine to recognize context, which is a very hard thing for a machine to get because of the subtleties of our different languages.
Rich goes on to spell out the use cases and applications in using artificial intelligence for legal contracts:
“It can let companies review contracts more rapidly, organize and locate large amounts of contract data more easily, decrease the potential for contract disputes (and antagonistic contract negotiations), and increase the volume of contracts it is able to negotiate and execute… The use of AI contracting software has the potential to improve how all firms contract – and it will do so in three ways: by changing the tools firms use to contract, influencing the content of contracts, and affecting the processes by which firms contract.”
AI Beats Human Lawyers at Their Own Game
Last year we published an article about how AI startups are coming for some white-collar jobs, including lawyers. We profiled an Israeli company called LawGeex, which has since raised another $12 million, bringing its total funding to $21.5 million for its AI-driven contract reviewing software. In a study sponsored by the startup, it pitted its AI against 20 experienced, U.S.-trained lawyers on reviewing five non-disclosure agreements. The contest showed that the AI was better at identifying risk, 94% versus 85% for the human attorneys. The machine was also a wee bit faster reviewing the contracts – less than 30 seconds against an average of more than 90 minutes. We wonder if LawGeex handles divorces online.
AI Startups Using Artificial Intelligence for Legal Contracts
Now let’s take a look at some other AI contracts lawyers trying to show up their human counterparts.
AI Search Engine for Legal Contracts
Founded in 2010, San Francisco-based Seal Software has raised $58 million in disclosed funding, and just today announced a $15 million investment by DocuSign, an eSignature and documents management company that Seal Software has worked with since last year. The company had also raised a $30 million round last June. The next month, the startup made its first acquisition – a Charlotte, North Carolina startup called Apogee Legal that had developed contract analytics using AI for specific applications such as procurement and data privacy, as well as more niche products like Brexit. Seal’s own flagship product, Intelligent Content Analytics AI solution, works like a search engine. Users can ask it various questions about contracts and get an answer to queries about legal contracts ranging from payment terms to liability and incentives to lease agreements.
A recent study sponsored by Seal Software, found that major corporations – think American Express (AXP), Hewlett Packard (HPE), Nokia (NOK), Novartis (NVS), among others – are increasingly using AI for legal contract review. Specifically, more than one-third of respondents are currently using AI for analysis and review of contracts and agreements, and half expect their spending on contract AI to increase in 2019. As they say in the movies, case closed.
Learning From the Best
Launched in 2016, UK startup Luminance has raised $23 million, including a $10 million Series B in February. Based on research in machine learning and pattern recognition techniques developed at the University of Cambridge, the platform reads and understands legal documents, while actively learning from the interactions between the lawyer and his or her documents. Luminance claims firms can save up to 85% of their time in the first two weeks of deploying the software.
Luminance says it has more than 130 customers, including 15 of the Global Top 100 law firms and three of the Big Four accountancy firms.
Doing the Work of a First-Year Associate
Boston-based LinkSquares, founded in 2015, has raised nearly $7 million over two Seed rounds, including a $4.8 million one earlier this month. Its platform seems to be another sort of legal contracts search engine that uses NLP that has been trained on “tens of thousands of variations of sentences that live inside contracts to surface accurate and meaningful results.” Instead of “paying a first-year associate at an outside law firm to manually read through” documents, LinkSquares manages all the legal contracts in the cloud. Among its customers is DraftKings, a popular fantasy sports startup that was growing so fast that it just put all of its legal paperwork onto a shared drive. LinkSquares got all that organized in a month, and now thousands of PDFs are easily searchable within seconds.
Update 02/26/2020: LinkSquares has raised $14.5 million in Series A funding to help even more companies move beyond contract management struggles and realign their priorities, processes, and outcomes. This brings the company’s total funding to $21.5 million to date.
Sifting Through Legal Documents Using AI
Founded in 2013, Pittsburgh-based LegalSifter has taken in $6.2 million for its platform, which uses machine learning and NLP to sift through documents and identify important concepts that demand attention or are missing entirely. Its own lawyers find hundreds or thousands of examples of a specific concept, then the data scientists use machine learning to find the patterns that link the concepts, parsing through the different writing styles to locate the nuggets of important information:
The company offers specific “Sifters” for different legal contracts, from leases to invoices and everything in between.
Humans Still in the Loop
Founded back in 2006 and headquartered near Boston, Brightleaf has raised $3 million in disclosed funding, though nothing since 2010. Brightleaf refers to its use of AI for legal contracts as a semantic intelligence engine that analyzes “any and all commercial terms, legal provisions, and obligations from any text-based legal document.” The data are eventually put through a multi-step Six Sigma process using additional automation, before being reviewed by a team of legal and financial experts. The startup claims its process provides 99.99966% accuracy. We’ll blame humans for not getting a perfect score.
Translating Legalese Into Simple Language
No word on the financial backing behind Legal Robot out of the San Francisco area in 2015. We do like its origin story as told to the New York Times: The founder decided to start the company after six corporate lawyers, each billing at hundreds of dollars an hour, were inspecting a contract looking for capitalization errors. Legal Robot is built to translate legalese into plain English for those who haven’t sold their souls to become an attorney. Deep learning helps “measure the complexity and readability of the language, and identifies the responsibilities, rights, and terms of an agreement.”
It then presents the information in a way that is “far more interesting and understandable than legal language” so users will understand their own deals with the devil.
AI is quietly but effectively invading the legal industry, starting with one of the more mature and proven artificial intelligence technologies – NLP. One of the other results of that Seal survey we mentioned earlier is that of those participants who said they are not using artificial intelligence for legal contracts, nearly half said it’s likely or highly likely their organization will implement it in the coming year. That’s serious penetration in just a handful of years, which one would hope would be reflected in the cost of legal fees to consumers. But the jury is still out on that one.
Tech investing is extremely risky. Minimize your risk with our stock research, investment tools, and portfolios, and find out which tech stocks you should avoid. Become a Nanalyze Premium member and find out today!