Since the days of ancient Rome, people have relied on paper documents and manual reviews to interact with other humans through the law; with the advent of digital documents and e-signatures, these processes have fundamentally changed.
Documents are now drafted using keyboards, signatures are written electronically, and contracts are stored inside servers. Legal and business professionals have adapted to these technological changes throughout recent years, and in contract management, some of the prevailing solutions involve storing files in data rooms and manually keeping track of important document milestones in spreadsheets and word documents.
Artificial Intelligence (AI), specifically advances in Natural Language Processing (NLP) and Computer Vision, have begun to accelerate changes to many legal processes. Recent progress in NLP and Deep Learning has opened new possibilities within the legal sphere, many of which have gone unexplored thus far.
Lawyers have started to integrate Machine Learning and AI into their practices, as they are under pressure from clients who are not happy with having to pay high-billable hours for routine tasks such as document review, and lawyers’ human capabilities are partly wasted when devoted to tasks that can be automated.
We have seen this trend also within enterprise organizations with the rise of legal operators, a new type of role that stands between legal (i.e. the general counsel) and the business owners. Legal operators streamline everything legal within the company.
I believe that AI should be an assistant to lawyers and consultants, not a replacement. State-of-the-art Deep Learning models still make errors and cannot currently compete with humans in terms of providing good judgement and sound decision-making. A good use for AI technology is the extraction of important parts of a contract, the highlighting of non-standard verbiage, and the generation of summaries. Time-saving is the biggest win: AI use can save a legal professional up to 80% of the time manual review would take without sacrificing quality (i.e. better-than-human-level accuracy, which has an error rate of about 11.3%, according to the American Bar Association).
I also believe that AI applications should also work “off the shelf”, meaning that users should not have to spend time training and improving models in order to reach their desired level of accuracy.
Growing up, I would spend weekends and school breaks helping my parents (both consultants) prepare business reports, review contracts, and check standard provisions so they could be presented in front of their clients.
This work was tedious, very time consuming, and I knew there was a better way.
I started to look for tools to simplify and automate the review process, and even spoke with a couple of vendors advertising such solutions, but they required a lengthy integration and a high up-front cost before we could even try their products. Upset from the results of my research, I decided to create something new.
When I started to talk to other analysts and contract managers, I found that I was not the only one with this problem.
As startups and SMEs grow, their small in-house legal departments need to keep track of the increasing number of contracts and extract key information such as renewal dates, terminations, and notice periods. Talking to lawyers and investors, I also found that legal due diligence is a process that takes a lot of time, during which junior attorneys and legal operators need to review hundreds of documents in a short amount of time without making any mistakes.
At this point, an augmented contract management and review tool becomes a necessity.
A good Contract Management Tool (CMT) would let users easily create contracts through a dynamic form; send the contracts for review to the other team members; track modifications; and send for signature; the ideal CMT would also suggest which provisions are the most appropriate for that specific type of agreement and jurisdiction, while providing relevant trends in the geographical area at the time of drafting.
A good CMT would let users organize documents into folders, share permissions and track access; the ideal CMT would also automatically organize documents into smart folders based on type of agreement, parties involved, and agreement date.
A good CMT would let users tag important keywords and create reminders for renewals, notice periods, terminations, survivals, etc; the ideal CMT would find those keywords automatically and send reminders to relevant users.
A good CMT would provide redlining and versioning; the ideal CMT would find non-standard verbiage, highlight high-stakes provisions, and spot red flags.
Data is only valuable when it is usable in a workflow. So, integration with upstream and downstream applications is essential. A good CMT would integrate with several data rooms, e-signature software, and CRMs; the ideal CMT would also have every app and feature available within two clicks to optimize human-machine interactions.
At SpeedLegal, we have made it our mission to build this ideal contract management tool and have already taken the first few steps towards the future ideal. We are creating an AI-driven application that empowers legal and business professionals to review complicated contracts in a few minutes instead of a few hours.
Today we use Machine Learning, Deep Learning, and NLP; tomorrow we may also use Smart Contracts and find the perfect combination of these technologies.
In the next post, I will explain what SpeedLegal is in detail.
I believe in a future in which people will be able to understand contracts without having to read them line by line. It’s a long path and we will get there.
The advent of Blockchain and Smart Contracts is helping automate contract execution thanks to if-then clauses that don’t require human approvals to be triggered.
Some good applications of such technology:
In the future, we will probably see forms and questionnaires connected to code snippets instead of lengthy documents with legalese verbiage. This will take some time and as of today, most of the contracts use spoken words and tend not to be very user-friendly.
The practical improvement is to use AI technology and the huge number of contracts at our disposal to augment contract managers, lawyers, and key decision-makers so they can spend more time on high-value tasks such as strategizing, building trustful relationships with their customers, creating new opportunities, and closing deals.
Let’s allow anyone to understand long and complicated contracts in a few minutes instead of several hours!