In 2016, the managing partner of the litigation boutique I was working at had just returned from a conference at Harvard for law firm managing partners. He’s panicked. Big firms are adopting AI, and he believes if we don’t do the same, our firm won’t be able to compete against their tech advantage. My background is in technology and computer science, but I hadn’t coded anything in years, and I was too busy litigating to have time to follow the latest developments in legal tech. In my mind, I had seen IBM’s Watson beat Jeopardy a few years earlier, been terrified as a junior associate about what legal AI products would do to my job, and was getting pitched by legal tech startups about their latest tech nearly every month. It seemed all too realistic that we’d be out of work if we didn’t embrace AI.
How do you start an AI company back in early-2016? For us at my then-law firm, we had a contact at IBM so we started there. Through that contact, a group of us met with IBM’s Watson team, we pitched our initial idea of what we wanted to build with their platform. They laughed. Their words: “maybe with $100 million and 10 years.” They were right. These initial ideas were based more in science fiction and less in the reality of where AI tech actually was in 2016. Based on that feedback, it was back to the drawing board to simplify the initial product. We came back with an AI system that analyzed legal complaints and determined the allegations and causes of action alleged within. That turned out to be doable. LegalMation launched in December 2016 on a shoestring—all of the founders kept their day jobs. A little over a year later, myself and the other founders had all left our day jobs and begun working full-time for LegalMation. In April 2018, the company’s signature Complaint Analyzer product launched. LegalMation’s first two customers were one of the largest retailers in the world and one of the largest law firms.
Over the next several years, I transitioned from attorney to tech executive, leading the product, analytics, and AI teams in building LegalMation’s product suite. I became an expert in all things legal tech, I became a SCRUM Master, I became religious about requirements and QA testing, and I lived my life in two-week sprints. I also saw firsthand how the legal world’s attitude towards AI was beginning to change. In the beginning, LegalMation’s sales pitch was “we can do in 2 minutes what takes your associate 8 hours.” The firms’ response was that they could bill for those eight hours, but they’d have to pay us for the two minutes so why would they want our product? The sales pitch never changed, but by 2019, the attitude in law firms had changed. I will never forget when I was on a sales call with a law firm. I had just finished a demo of the latest release of the product, and they were not impressed. They said that what I had just showed them was table stakes (i.e. the bare minimum for non-gamblers). They may have meant it as an insult, but I took it as one of our biggest compliments. The product that just two years earlier was revolutionary and award-winning was now considered table stakes.
This change that I witnessed firsthand can be seen throughout the legal world. Five years ago, the only AI that 99 percent of lawyers had ever experienced was via e-discovery platforms (predictive coding and similarity scores). Nowadays, courts have issued sanctions for not using TAR (technology assisted review) in document productions, and every attorney that does legal research has had their research dramatically improved through legal research platforms’ investments in AI. Moreover, every decent sized law firm has embraced analytics and automation and is starting to wade into the possibilities that AI will provide. My current firm, Ogletree Deakins, is a prime example as the firm has undertaken a number of efforts to improve its AI and analytics capabilities, the latest of which was creating a position for me to lead a number of the firm’s AI, analytics, and automation projects.
Having lived this change firsthand, what do I predict for the future? Both a lot of change but also not that much. The low hanging fruit of AI, analytics, and automation is rapidly being picked by the current crop of legal tech companies, leaving the tough stuff for the next generation of entrepreneurs. Over the next five years, adoption of the current set of technology will rapidly increase. We saw it with e-discovery in the mid-2010s, when there was initial resistance to the additional cost of using the latest technology as well as the decrease in billables from document review. Then, in what seemed like overnight, every client required an e-discovery provider with TAR. We are currently seeing a similar situation occur with contract management. CMS (Contract Management Systems) are the latest darlings of legal tech fundraising with a new blockbuster round announced every other month. Over the next two years, CMS will be mandatory for every law firm or corporation of any significant size. Next on the horizon are the “no code” and document automation platforms. A new legal tech company appears every month offering some new iteration of one of these technologies. As this area matures, just like e-discovery and CMS before it, clients will soon require firms to utilize “no code” and document automation technologies.
Through all of this, legal research will continue to improve. Legal research providers either have deep pockets or big imaginations for how to use AI to improve research and will continue to make research more efficient and accurate. Analytics will also continue to see a big bump. As more and more state courts adopt e-filing and digitize their historical documents, there will be a huge boon in analytics and predictive technology regarding judges and litigation. I am excited to see what comes of the potential of the free PACER movement that seems to be on the horizon. The amount of data contained in PACER alone is enough to keep researchers and academics busy for the next 20 years trying to discover all of the interesting connections. However, like so many promising technologies, we will also start to see the limits of legal analytics over the next few years. Much like predicting the stock market, it is much easier to describe what happened in the past than it is to predict what will happen in the future, and just because you know what a judge has done in the past across thousands of motions, I would hesitate to try to predict what he or she would do with my client’s motion.
That’s not to say that AI and analytics won’t change the practice of law over the next five years—they definitely will—but law students shouldn’t be worried about their future in the profession. There’s plenty of room for more lawyers and plenty of legal work still to be done. Rather, through all of these innovations, I would expect over the next five years that the practice of law becomes 20 percent more efficient and a lot more data-focused. “No code” and document automation are here to stay, and they will make everyone’s lives easier once we fully embrace them. The 20 percent of work that they will automate—like preparing proofs of service in triplicate for 10 motions—is not work that anyone would miss doing by hand.
Similarly, in five years, you won’t be able to pitch a case or respond to a client’s RFP (Request for Proposal) without including due diligence data on opposing counsel, the court, and the subject matter at issue along with the analysis and visualizations to back it up. This will naturally benefit those firms that have more data and have that data better organized. Small firms, like always, will need to be creative in bridging the data divide. In fact, I wouldn’t be surprised if we start to see large firms and clients start licensing their troves of data in the near future once they realize that they can make more on licensing than on the competitive advantage the data provides. Lastly, more and more clients will stop paying for research. As a big firm, you’ve either written this motion before (and can now find it in your DMS with AI-assisted search), or you have near instant access to the most relevant cases via your AI-assisted research platform, or your analytics provider will give you the top 20 similar motions your judge has granted in the past five years. It will be interesting to see how the legal field adapts to more and more expensive research products that clients will pay less and less for.
Notice what’s missing from the above: lawyers are still doing the vast majority of the work. Over the next few years, the practice of law will accelerate with technology, but it will mainly be a blue-shift of the work. AI is not going to be writing your briefs for you in five years, it will not be taking your depositions (but it will hopefully be summarizing them for you), and most of all, AI still will not be making novel arguments or performing critical thinking. Like the present, AI will be a tool to help supplement an attorney’s skillset, but it is still a long way off from automating the substantive parts of an attorney’s job.
But, who knows, ask me again in five years, and maybe I’ll have my AI assistant write my follow-up.
Tim Fox is the Director of Practice Intelligence at Ogletree Deakins. Prior, he helped found and lead LegalMation, a legal tech company that uses AI to automate early-stage litigation tasks. Earlier in his career, he was a litigator at a variety of firms in California, focusing his practice on patent and tech-related litigation. Before practicing law, he was a DBA/systems integrator for Nestle/Dreyers. Tim obtained his B.S. in Computer Information Systems from Cal Poly and his J.D. from the USC Gould School of Law—where he also is currently a lecturer-in-law.
Interested in having your own Viewpoints article published? Submit your proposal here.