Legal AI Primer
The emergence of artificial intelligence (AI) and machine learning technologies in the legal industry has been nothing short of radical. Given that these technologies have been the subject of much attention lately, it makes sense to first explore what legal AI is, its history, and its current state before delving into specific trends and case studies. According to legal futurist Richard Susskind, AI refers to a computer program or machine that "displays human attributes such as reasoning, learning, planning, problem solving, perception and the capability to move and manipulate objects." Notably, AI is now only one of many subfields found within the body of computer science, with numerous artificial and cognitive computing technologies also emerging over the years. Some of these advances have found their way into the legal sector in recent years. These include: The use of machine learning to parse through legal documents and identify relevant cases and statutes AI-assisted contract review, AI platforms that generate basic legal pleadings and briefs, and AI to support litigation and e-discovery Predictive analytics for appeal success rates AI based on neural and deep learning for improved contract review, due diligence, and compliance AI chatbots to provide basic legal advice Expert systems that automate knowledge-based decisions Deep neural networks and deep learning for automated bidding Robotic process automation (RPA) for automating repetitive tasks Swarm intelligence for dispute resolution Overall, these technologies have allowed lawyers, legal professionals, and business firms to better use data and maximize efficiency . Not any one of these technologies is earth shattering. However, when used together, they offer the potential to revolutionize the practice of law and how legal services are delivered to consumers. Perhaps one of the broadest applications of AI in the legal sector involves the use of software bots that can comprehend, analyze, and understand large volumes of both structured and unstructured data. These systems can be used in just about any way imaginable, from drafting agreements and legal forms to performing complex legal research – all with very quick turnaround times. In fact, these types of programs can even translate foreign language legal documents, perform statistical and analytical tasks, and predict legal outcomes with impressive accuracy by utilizing data science. Despite obvious possibilities, AI in the legal sector is still in early stages. However, this hasn’t stopped numerous start-ups from developing their own legal AI technologies. In fact, the industry is now experiencing a wave of legal AI start-ups that have recently been attracting considerable interest from investors, legal professionals, and businesses. These include businesses such as ROSS Intelligence, IBM, Legal Sifter, Casetext, Luminance, Kira Systems, LawGeex, Legal Robot, and many others.
Prominent Legal AI Startups
Another startup leveraging AI for a niche aspect of law is ROSS Intelligence. Founded in 2015 and based out of San Francisco, ROSS aims to streamline legal workflow. Their software uses natural language processing and "sentiment analysis," much like many dating apps, to parse through billions of documents within their in-house database in order to find relevant results based on keywords and context. The log-in system is similar to Lexis/Nexis and Westlaw in that you enter a search term or terms. However, ROSS takes it a step further in that it dissects the language and chooses the answers that are the most relevant. The predictive technology can even suggest which attorney at the firm might be the best fit to take on the case, based on previous information inputted onto the system.
Often associated with ROSS is LawGeex, another AI-powered new-comer that focuses on a similar area of law. Founded in 2013 and based out of Tel Aviv, Israel, they tap a powerful network of over 160,000 legal reviewers from around the world, whom review contracts and standard agreements. Using AI and machine learning, LawGeex can turn a 20-page contract into a succinct, one-page checklist that highlights issues that need to be addressed. LawGeex claims that their software is 10 times more accurate than the average in-house attorney, and filters through NDA’s, sales contracts, and non-competes at a rate 50 times faster.
Heretik is another emerging name in the legal AI market. Founded in 2015 and based in New Orleans, Heretik is an artificial intelligence document review solution that provides e-discovery software to lawyers. Heretik boasts innovative technology that allows both attorneys and contract managers to dramatically cut review times while maintaining precise accuracy. Artificial intelligence is being used to annotate and categorize contracts, extracting certain data points and utilizing advanced machine learning algorithms.
The last startup we will touch on here, also hailing from outside of the United States, is Neota Logic. Founded in 2010 and located in London, Neota Logic (and its US office) provides clients with no-code automation technology to create their own applications, just as Powersoft does. Their platform combines machine learning, workflow and experts systems into one platform known as Neota Logic 5.
Key Tech Used
As with any other industry, AI front-runners such as IBM and Google Cloud have a hold on the top tier of legal AI tools. But outside of these players, dozens, if not hundreds, of small scale legal AI startups are emerging. They are leveraging the same technologies to automate processes and help attorneys manage information. Natural language processing (NLP) drives the same political lexicon and translation applications that allow legal AI startups such as LegalSifter and Consultwebs to categorize and determine the most relevant information found in contracts by allowing computers to understand the context of words. Machine learning (ML), which allows systems to learn and adapt over time from data, is now used by Autodesk and Everlaw to identify and rank evidence on digital forensic and case management platforms.
Predictive analytics, a cornerstone of many legal AI solutions, enables systems to analyze previous case results to predict future outcomes. Companies such as Premonition and LexPredict are using past cases to build a predictive model of how judges and attorneys are likely to rule in current cases. Rule engines allow attorneys to mitigate risk and ensure compliance on a wide variety of legal issues by applying logic to legal rules and regulations. For example, cognitive computing and NLP allow websites such as ROSS to sift through complex legal issues and find the right results. Dashboarding and data visualization technologies are helping clients and attorneys quickly identify the financial impacts of legal issues (as with Mindcrest and Afina). Virtual assistants and chatbots are helping clients, attorneys and judges with common issues (as with Neota Logic and LawDroid). And now Cloud computing and blockchain are enabling secure, seamless integration of AI capabilities and data for legal applications.
Challenges for Legal AI Startups
Despite the significant opportunities, legal AI startups also face numerous challenges during their journey from inception to the eventual scaling phase. As mentioned previously, these challenges are mostly centered around market acceptance, regulatory barriers, and perceptions of ethical dilemmas.
For one, many legal AI startups can face hurdles in gaining buy-in from legal professionals who are often reluctant to embrace new technology. Lawyers may be hesitant to cede control over certain tasks to AI, and may worry about negative impacts on their profession. These concerns can disproportionately affect early stage startups that are more focused on product development and building effective AI models than on marketing, sales, or business development strategies. As such, media attention and feature coverage are essential during this phase to bridge the gap between legal professionals who do not yet understand the intricacies of AI and its impact to the legal industry.
However, creating media coverage for a legal AI startup is twofold: (1) general media publications do not understand the deep nuances of the legal industry and are not familiar with the uses of AI in the sector; and (2) legal publications have an even more difficult time, as legal AI is still in its infancy and its applications are varied and unique to specific problem sets and market segments.
Creating market awareness and demand will be a slow process, and legal AI startups must demonstrate results before they can enjoy a competitive advantage over others. Thus, to get over this hurdle, legal AI startups utilize all means of digital advertising (e.g., Google AdWords/Search Engine Optimization [SEO], LinkedIn Sponsored Posts/SEO, Facebook Ads, etc.) to create an online presence that brings conversion and adoption.
Secondly, there are critical barriers to entry in terms of the requirements for regulation, compliance, or institutional legitimacy. As legal AI startups are technology companies at heart, it is not uncommon to see nimble and adaptable entrepreneurs in this space, but that agility must be tempered with an understanding that the sale of technology must be balanced with a healthy respect for the rules and regulations that govern lawyers and law firms. Without this balance, any added value of the technology gets lost when ethics and compliance issues come to the forefront of discussion , and the venture fails to gain traction in practice.
Here, the legal AI startup must become even sharper with their messaging to avoid any potential issues. For example, getting lawyers to adopt legal research-based AI is easy when technology can be sold as an alternative to using LexisNexis or WestLaw. However, selling the same technology to a law firm that presently has offices in both China and Canada when both jurisdictions have quite different rules on using blockchain-based legal AI means the venture must understand those rules and how they play into the technology.
Thirdly, some legal AI startups face initial difficulty with integrating their solutions with existing lawyer workflows. Many lawyers are creatures of habit. They have their tried and tested processes that they employ for every case, regardless of how well it works. As such, any new technology that seems to disrupt their workflow may feel more like a nuisance than an opportunity.
To overcome this challenge, legal AI startups must be prepared to educate and train their target users on the ease and benefits of their technology. They must be able to work with the lawyers to show how the technology improves their workflow, and be flexible to operate within the constraints of existing practices.
Furthermore, gaps in technological deployment may also result in problems for legal AI startups that do not have a clear idea of the end-user in their product design phase. For example, the design considerations for a product intended for lawyers working within IP or patent law practices is very different from the design for civil or criminal law practitioners, thus the features and deployment of the product must be differentiated and customized for each specific case.
Lastly, legal AI startups often face hurdles in determining the best business model to execute given the realities of a competitive marketplace. For example, charging hourly rates can be a trap for any legal professional, and as such, many legal AI startups offer free to low-cost versions of their technology as a means to offer a try and buy method for potential clients, which is common in the technology space.
However, there are other, more sustainable business models that can be adopted such as hiring junior associates that mirror the tasks of the technology as case studies for students.
The Future of Legal AI Startups
A significant influx of capital into legal AI and increased demand for tech-enabled law services augurs well for future growth in the legal AI startup space. However, one third of legal AI startups that received 2018 investment plans to shut down, a clear sign of the survival-of-the-fittest nature of the market. The challenge will be to increase the number of customers regularly using tech-enabled services to yield a sustainable revenue stream for each solution.
Several areas are ripe for continued expansion. First, basic document automation such as form generation and e-signatures is expected to continue to see substantial demand. Second, solutions applying machine learning algorithms to generate legal documents, due diligence reports and contracts and agreements are becoming increasingly popular. Third, AI-based translation and transcription tools show promise. Fourth, AI solutions for contract and policy analysis will be important as will tools such as chatbots. Fifth, blockchain technology, particularly smart contracts, could positively disrupt how lawyers work by enabling a more efficient and reliable process of enforcing contracts and closing transactions. Sixth, the emergence of legal analytics is moving downstream beyond the analytics needs of in-house counsel to the clients of law firms. Seventh, configurable AI-based analytics tools for legal services are expected to gain traction. Finally, greater use of APIs and cloud-based platforms will continue to fuel innovation.
Investing in Legal AI Startups
The investment landscape for legal AI is dynamic and ripe for exploration. With increasing activity and ongoing interest, investors are keeping a keen eye on the sector. Within the past year, there has been a noticeable uptick in investment activity, with several companies attracting significant financings. One of the most prominent players in the legal AI investment space is Evisort. In April 2017, Evisort announced a $2.5 million round of investment led by The Venture Reality Fund. The company has since raised an additional $6 million in funding through a Series A round led by Storm Ventures. Evisort’s co-founder, Professor Jerry Jin of Harvard Law School, was awarded the J. William Fulbright award for his studies into the role of artificial intelligence in the legal profession. In 2017, Diligen, a legal analytics platform, received $3.1 million in a funding round led by Serra Ventures. In October 2018, Co-Counsel, Inc., a company that uses AI to reduce due diligence time periods, startup and legal aid ProCheckUp, and contract management platform LinkSquares all announced funding rounds totalling over $5.5 million. Funding activity in the legal AI sector is also heavily concentrated in New York, Silicon Valley and San Francisco. A number of firms are exploring legal AI investment options, including Boston-based attorney search platform Lawdingo, which has raised $650,000 in angel funding, and Apperio, a billing tracker for lawyers. Further venture capital activity has been seen in the UK, with Lexoo, a platform connecting legal problems to solutions such as fixed fee offerings , and legal analytics platform CourtQuant, raising $1.2 million and £65,000 respectively. With these important developments, there has been mounting scrutiny into the return on investment in legal tech. In the US, Dan Lear leads the charge with Ascent (a collective of those interested in the development of innovative and effective products and services for the legal sector) who has undertaken a number of surveys into the topics of investment trends and rates of return. In a recent study, Ascent surveyed over 100 founders, investors and analysts in legal tech revealing current industry sentiment, rate of return on earlier-stage companies, and their expectations for rate of return for companies currently being funded. Ascent’s survey showed that venture capitalists generally have a positive view towards investing in legal tech, with the majority of all respondents agreeing that investing in legal tech is potentially quite lucrative. The results show that, as with other sectors, seed and angel investments in legal tech are illiquid and thus the time-frames for investment success are longer. Traditionally, returns from legal tech investments have been similar to media and communications investments, though the majority of those surveyed agreed that legal tech has greater revenue potential but less predictable outcomes. Legal tech success stories include Axiom and LegalZoom, both of which have been acquired. The majority of respondents reported that their advice to new-found legal tech start-ups would be to consider investing in product innovation to remain competitive.