We are back on track with our Lawyer to Legal Tech Co-Founder Series. Today, we welcome Pablo Arredondo, former litigator and Casetext Co-Founder and Chief Legal Research Officer. Last year, I wrote that Casetext is one of the companies that is truly innovating legal research. And in fact, since my last post, Casetext has rolled out even more features – searches now turn up case briefs, and users also have the option of accessing case holdings and black letter law gleaned from Casetext’s culling of caselaw in its data base. Let’s learn from Pablo what it’s like to found and work at a forward looking legal tech company. To access earlier Law to Legal Tech/Entrepreneur profiles, click here.
Q. Thanks for making some time for this interview. Tell us a little about Casetext – what it does, where it’s located and how long its been in operation and current company size.
Casetext is a start-up focused on building a new type of legal research. The company was founded in 2013. I have watched it grow from two people in Jake’s Sunnyvale living room into over thirty-seven employees in the SOMA district of San Francisco.
Q. Prior to starting Casetext, did you practice law and if so for how long and where? What was your area of practice? Do you still currently practice?
Prior to putting my shoulder to legal informatics I worked as a litigator in San Francisco and New York. I focused primarily on high stakes patent litigation, though I have worked on copyright and trademark matters and pro bono criminal cases as well. I don’t currently practice.
Q. What gave you and your co-founders the idea for Casetext?
Before we met, Jake and I independently concluded that better technology could help the legal profession, and that a start-up focused initially on legal research was the way to do it. For me, I was constantly comparing the technology of the companies I was representing, like Apple, with the technology we were using to represent them. I hated basically all of the technology we used to litigate, from the document management system to the e-discovery system. You won’t be shocked to hear that the partners at the firm had varying levels of interest in my views on the deficiencies of current legal technology.
It took me five years of this pain to “take the plunge” into the the world of start-ups. Jake got there a lot faster and, to his credit, was less concerned with vengeance against inferior user interfaces than with helping the legal profession as a whole be more democratic. In the early days in his living room Jake would talk about how it was a moral failing that some people couldn’t afford legal research tools, and I would talk about how Lexis’s “Sort by Jurisdiction [Z-A]” was a different type of moral failing. We were both right.
We were lucky enough to be joined by Laura, who had known Jake since high school and was willing to leave her high paying law firm job to join a nascent Casetext. I would like to think that it was my passionate views on how to make copy-with-cite functionality better, but I suspect it was Jake’s articulation of Casetext’s larger mission that convinced her. Either way, Casetext would not be where it is if she hadn’t been willing to forgo a skyscraper in Manhattan for a two-room office in Palo Alto.
Q. Before we talk about the product itself, I want to ask you specifically about the steps from idea to implementation. First, once you had the idea for CARA, did you take any steps to validate it? Second, after you decided to go forward with CARA, how did you find developers and how long did it take to get the site up and running?
At its heart, the idea behind CARA is that you can mine the information encoded in the litigation record (e.g. a brief or a pleading) and use that information to create a new type of legal research. I remember the “a-ha moment” vividly and the truth is once the idea hit me I never doubted that it would work. I worked with some friends and within a few weeks had a crude proof of concept that ran on a tiny database of cases.
But a proof of concept is not a product, and if Silicon Valley teaches you anything it is that ideas and prototypes are just the start. The CARA that exists today, powered by a full fledged case law database, is the result of years of effort by Casetext’s amazing team of data scientists, engineers, designers, and lawyers. And of course, the technology is always improving. I would like to give special shout-outs to the brilliant and indefatigable Ryan Walker, who more than anyone helped CARA become something lawyers should use, and to the immensely talented Anand Upadhye who tackled the daunting challenge of getting law firms to try (and buy) something new.
Q. All startups face bumps in the road. Can you share one or two of the hurdles that Casetext has encountered, and how the company addressed them?
One of the things we realized from the beginning was that having primary case law was not enough. Lawyers and other legal researchers needed analysis to help them understand what they were reading, to alert them to pitfalls, and to put the individual case they were reading in larger context. In short, the cases needed to be annotated with expert analysis.
Jake is a lifelong coder and he was familiar with platforms like Stack Overflow on which engineers volunteer their time to answer each others questions. Early on, we wanted to create the same type of system on top of the common law. What we discovered is that for various reasons the majority of attorneys were not amenable to this sort of open, crowdsourced platform. We tried a number of things, even going so far as to create a specialized writing platform tailored for legal analysis, but we didn’t see the volume of organic contributions we needed to truly annotate the common law.
So we did what entrepreneurs do. We found another way. We saw that many of the attorneys who didn’t want to contribute annotations directly into the case database were in fact generating fantastic, free, analysis through the client alerts they wrote to raise their profile. This analysis would often gather dust on a law firm website, but we realized they would provide immense value is properly integrated into a research platform. Suddenly the “ask” changed from “Would you take time to create new content on our platform” to “Would you like us to help you get more viewership for your pre-existing content?”. We partnered with hundreds of law firms and non-profits, bringing in tens of thousands of pieces of analysis which we datamined for case references. The end result was much in line with what we had originally sought: opinions on Casetext are annotated with analysis from thousands of attorneys.
We also applied work that I had done at Stanford to further annotate our cases with concise case summaries, this time written by judges. We harvested over two million case summarizing “explanatory parentheticals” directly from the case law. These valuable descriptions of a decision’s key holdings appear above the full text of the opinion. Recently we also made the summaries a searchable database in their own right.
Q. I remember that back in the dark ages when I was in law school that LEXIS and Westlaw provided free service and training to addict students to their services. Is Casetext being used in law schools and what’s the feedback so far?
Absolutely. The entire Casetext platform is completely free for students and legal educators. My colleague Katherine Anton has done a phenomenal job integrating Casetext into schools; I believe at last count we have over seventy student ambassadors and many schools incorporate Casetext into their legal research curriculum. I have had the privilege of presenting to over twenty advanced legal research classes around the country and never get tired of seeing how receptive law students are to new technology.
Three years is an eternity to a start-up. By then, we hope to see Casetext being used as the principal legal research platform by a majority of lawyers (and judges for that matter). We will extend the power of our contextual approach to all aspects of the research process. A citator that helps you prioritize citing decisions most relevant to legal and factual issues in your specific matter. A platform that can monitor your docket in real time and instantly provides valuable material to attorneys right at the moment of need. An “adversarial” version of CARA that will be able not only to bring you relevant case law, but filter depending on which side of an issue/argument it supports. These are just a few examples and I will be disappointed if three years from now we haven’t built some things that I haven’t even thought of yet. To put it simply, we want to build the best legal research platform the world has ever known.
Q. What is your day to day role in Casetext? Can you describe a typical day (if there is such a thing?)
Thinking about this question makes me realize just how obnoxiously happy I am right now. Start-ups move so quickly that “typical” days tend to be short lived, but what has been constant is that I get to work with fantastic people across all parts of the company. So in the same day I might review the output from a machine learning experiment, discuss citator design, help prioritize what to build next and (very painful) what to wait on building, debate the merits of a potential business partnership, and help pitch CARA to a law firm.
Q. How does your training and experience as a lawyer help you at Casetext?
And as I alluded to earlier, the experience of using awful technology day and night, year after year, created an eternal fire in my soul fuels the passion to build legal technology.
The VCs have the phrase “authentic founder” to describe founders who are trying to solve pain that they personally experienced. Sometimes I see folks who didn’t practice come up with ideas that look really cool on the surface but aren’t actually all that useful to a litigator. I like to think that the stuff we are doing at Casetext is all directly focused on solving the real world problems of litigators.
Q. Do you think that coding experience is beneficial to lawyers who want to start a legal tech company? What other skills do you think are most important to lawyers who want to start a legal tech company?
Coding helps. I actually met Jake by way of an introduction from Paul Lomio, then Director of the Stanford Law Library, after I taught myself just enough coding to build a prototype search engine. If all I had was the idea, I am not sure things would have happened the way they did. And of course Jake is a real coder, and this was one of the reasons I was so keen to team up with him. In the early days, we would have an idea in the morning and Jake would have it coded by the afternoon. Even now, Jake will occasionally code something to make life easier for sales or marketing. Coding is just awesome.
As for other skills you need to start a legal tech company, what I would say is choose co-founders that complement your skills and balance your temperament. Because the truth is, you pretty much need every skill at one time or another. Thick skin is another prerequisite.
Q. What was the biggest challenge for you in transitioning from law practice to your current position?
Learning about the business aspects of a start-up. Fortunately Jake and Laura have a natural acumen for it; this is what I mean when I say find co-founders that compliment you. As a law firm associate you just litigate at whoever they tell you to litigate at. Nobody at Kirkland and Ellis asked me if I thought we should merge with another firm. When I was doing legal informatics research, I just tried to do the coolest stuff I could think of. Nobody at Stanford CodeX asked me if an employee wasn’t working out. But as a co-founder you have to help vet business strategy decisions, company personnel decisions, product decisions and many others and the truth is law school doesn’t really prepare you for it.
Q. Do you miss practicing law? What have you done, if anything, to keep your law license and legal skills intact?
I kept my law license active for both New York and California. There are times when I have small flashes of nostalgia for some aspects of litigation. I like how we got to use the word “purport” all the time. Litigators are always accusing each other of purporting. I liked the intellectual energy that went into a dispositive motion and the suspense of waiting for a judge’s ruling. But the truth is, I found what I love to do and it is Casetext.
Q. If another lawyer was interested in starting a tech company, what advice would you have? What excites you most about the future of legal technology?
My advice to a lawyer interested in starting a legal tech company would be to do it, and to remember to breathe. What excites me most about the future of legal technology at the moment is probably seeing what is going to happen in the field of natural language processing. Will the algorithms get good enough to identify when Justice Scalia was being sarcastic? Or will things remain largely where they are right now, with even sophisticated tools like Siri understanding only the simplest queries? There is a tremendous amount of good work to be done either way but watching how that branch of AI plays out is very exciting to me.