As technology improves, it opens new opportunities to change the way attorneys practice. From rethinking roles, tasks, and strategies to providing assistance in ways we’ve never contemplated, each year brings new opportunities for change. In this episode, Todd Smith and Jody Sanders talk to one of the people leading the technological revolution in law, Jake Heller, co-founder and CEO of Casetext. Jake explains his background in technology and law and how it led him and others to develop Casetext to make legal research better and more efficient. He also shares his thoughts on how legal technology is reshaping the industry and where it may go in the future.
Our guest is Jake Heller, the Cofounder and CEO of Casetext. Welcome to the show, Jake.
Todd, Jody, it’s a pleasure to be here.
We’re happy to have you here. A lot of our guests probably don’t know you and they may not even know about Casetext. Let’s give them a little background on you personally, how you started your career and the first few steps that you took before you wound up at Casetext.
I’d be happy to. Before running Casetext, I was a lawyer. Before that, I was a computer programmer. In law school, I had a traditional law school and afterward, I had a traditional legal career. I had the opportunity to work in government and nonprofits. I worked as a clerk at the First Circuit Court of Appeals in Boston, which was an amazing experience. I was able to work at a great large law firm in Boston as well, Ropes & Gray.
That’s what I did leading up to starting Casetext but my passion is in building things. I have been coding and building websites since I was about nine. My dad started an internet business in our garage in Silicon Valley. Some kids play catch with their dad or fix cars or what have you but I was building websites with him at a young age. It’s been in my blood.
I remember being at the law firm doing what any young associate does, a lot of research, a lot of writing, discovery and some other menial tasks and thinking there is so much here that I am doing that I did not go to law school to do. There are many things that I know that technology will help me do. That inspired me to decide to take a slightly different path from legal practice and start a business.
I’m curious about what made you make the jump from computer programming to law school. That seems like a big transition.
It’s funny. The answer, in some ways, goes back to high school for me. In high school, I didn’t do anything else other than speech and debate and coding. I was a coder first. I got randomly dragged into a speech and debate meeting and I loved it. I loved the fact that, unlike things I was working on in code where there was almost certainly a concrete solution to a given problem, in law and policy that’s never the case. That attracted me to it. You can probably imagine that I was the most popular kid in high school being both a debater and coder. I’m just kidding. I was one of the dorkiest kids, which was saying a lot for Silicon Valley.
That part of me was attracted and I still am attracted to how difficult some of the problems are that law is trying to solve. I hope I brought some of that in when I was an appellate practitioner or clerking. Oftentimes, you’re not just thinking about the specifics of the case. Would this make business impossible in the state or in the country if the rule is going to be so indeterminant? Those questions still excite me and still interest me. That’s what pulled me into law school.
The only thing I felt was that as a litigator at a big firm, I missed that piece of creating new things and seeing that creation, besides a brief, which is a cool creation in itself. It’s a one-time use. It goes in the case and, in some ways, disappears forever. I like the permanency of building things like websites, applications and so on. At the law firm, in whatever spare time I had, I found that I’d be coding things for myself to make myself more efficient and I knew that maybe it was time to pursue this full-time. That’s what I found myself doing after the 100-hour workweek.
What were the things that you were seeing as a practicing lawyer that stuck out as inefficiencies to you and things that you could maybe do better?
There are a few things. The thing that I ended up working on full-time at Casetext is in the research and writing space. I felt that when you were using anything as a person, not as a lawyer, you’re on Google, you’re using your iPhone, everything was easy, fast and efficient. You found what you’re looking for. It felt like magic.
When you walk into the law firm, it’s as if you stopped becoming a person; now you have to go backward and step into a time machine ten years. Especially at the time, this is years ago, the tools available to me to search for information and find information, though having a comprehensive library of information, which is great, it fell short of the magical experiences you get on the iPhone or Google.
That wasn’t just true of my researching and also writing. I always felt that in the world of code, there are all these tools that make you an efficient coder, things that as you start typing, it knows when you want to say as a tab does that, there were checks to make sure you’re not making any errors. The writing tools and code are amazing. The writing tools in law are still Microsoft Word or, when I clerked, still WordPerfect. They were old technologies that are not purpose-built for what we’re doing.
What else did I do with my time? I spent time creating discovery and that field’s come a long way. Going through thousands of documents, all of which are about people’s ESPN fantasy football scores, to find the one thing that mattered in the case and all those things are painful experiences. The whole time, I was thinking my peers and I went to law school to have experiences that more closely mirrored what I was doing in my clerkship.
What a lot of appellate attorneys get to do once you reach a certain level of maturity and a lot of things that you feel a little closer to applying the gray matter between your ears to difficult problems and challenges as opposed to, “Next, next, next.” “No, no, no.” All that felt like, “Why am I being paid this much?” I feel a bit bad for the client that I’m being billed out for this much to do these tasks that software could at least help with, if not do entirely.
It sounds like you’re a big disrupter, Jake. I like to talk about how the wrong answer to any question in the law is, “We’ve always done it this way.” You’re one of those forward thinkers who because of your unique background, it sounds like you decided to attack problems from a little different angle. That’s great.
It’s interesting to me that you’ve got that tech background but relevant to our audience, we do talk about legal technology here on the show a fair amount. You’ve got the clerkship background, the big firm, Ropes & Gray, experience and now you’ve moved on to being an entrepreneur. How long did you practice at Ropes & Gray?
It was a couple of years.
It’s long enough to get a real sense, though for, “This system may not be completely broken. It’s maybe close. I’m sure there’s room for debate on that but there are definitely things that can be done to improve it.” I assume you took that insider knowledge being a lawyer and having worked as a clerk and at a firm to lay the foundation for starting your company.
That definitely played a part. To react to something that you said earlier, the funny thing about law is there is a time when doing what we’ve always done is the right answer, which is adhering to precedent. In some ways, as lawyers we do fall back on that way of thinking in other parts of our lives because we spent three years in law school and 10, 15 or whatever years of practice thinking about the precedent and doing things exactly what we’ve done before.
In that sense, maybe there’s this, in the water, a psychological thing that in our professional work, we’re constantly going back to precedent. How we do it, how we find that precedent and how we write about that precedent, that’s the stuff that should change. That concept is tough for some lawyers to have that mode switch between precedent and not precedent. It’s funny that there is that tension sometimes for some people.
I never thought about it quite that way but you’re onto something. We are locked in with what is the law and the law is sovereign. When it comes to not only the things that I know your company does but in the day-to-day routine of how to manage a law practice. We’re so stuck in, “Yes, it’s important to learn lessons from those who came before us,” but one of the things that our industry lacks, and you’re one of the exceptions to this, is innovative thinking. How do we apply this new technology to what we’ve been doing day in day out for however many years? I am intrigued by the approach that you’ve taken in implementing your experience with Casetext. How did the company get started? Tell us a little bit about the history.
The company got started years ago. I was practicing at this law firm and I had a great experience there and nothing bad at all to say about that experience. It’s a wonderful firm. To their credit, they treat our clients totally right. I had a great experience there. During building things, at one point, I goody-two-shoes in a sense. I went to the general counsel of the firm and said, “I’m thinking of building something. It’s on the side during my spare time. Would that be allowed?”
She said something along the lines of, “If you did, we would own it because you work for us.” I remember getting into this silly conversation of, “I’m doing all of this in my spare time and the weekends, late at night. What would it be like if I was in a band that played at weddings with a bunch of friends or something like that and we made $500?” They go, “We own that too.”
I was like, “As a woodworker, I donated a bench that I made to a park?” She goes, “We probably own that too.” They’re taking a hardline stance. They were just trying to make a point. She also questions, “How do you have extra time also? Do we need to give you more work?” I was definitely hitting my billing requirements but I understand where they were coming from.
That conversation led me to think, “I can’t be half in or half out. The firm’s essentially telling me as much. If I’m going to do this, I’m going to jump with both feet in and start.” I decided to leave the firm without $1 financing to the idea and without a line of code written. I said, “We’re going to do this.” The funny thing at the time was, I said, “Either this is going to be a total failure and I’ll need to come back and hopefully you’ll have me back or it will take off and we’ll IPO in three years, etc.” The reality is a little more in the middle of those two things.
Thankfully, it was a little closer to success and we are doing well. That’s how we got started. We got lucky in the early days. I applied for a program called Y Combinator, which out here in Silicon Valley is a hard to get into early-stage they call Startup Incubator. They give you a tiny bit of money in the grand scheme of things and a lot of advice from people who are good at this, the first financers behind Airbnb, Reddit and Dropbox. Also, thousands of other companies that you might use in everyday life like DoorDash or Instacart. Some of those folks were my peers as I was entering this program, Y Combinator.
That set things in motion, it helped us raise the next bit of capital that is necessary. For these businesses, you need some money in, in order to get going. Sometimes that money is millions and millions of dollars. You’re trying to acquire, for example, digital copies of all the law and hire the best AI engineers from Google and Facebook and build something that is truly extraordinary. All of that requires time and money. We were lucky to have that in the beginning.
I was lucky to hook up with two cofounders. One from my law school, from Stanford Law School a few years before me, who is an absolute genius when it comes to innovation and legal, he had already started a company in that area that did well. One who’s a friend that I knew all the way back from high school, who ended up going to Yale Law School and running her law review as I was writing my law review. It’s parallel paths on the opposite coasts. She also clerked for the District of New York. She is operationally minded, process-oriented and could help complement the product engineering side with the business side.
In some ways, you need all that. It’s like the Captain Planet rings that all come together to make Captain Planet. You need every one of those sides and more to make it a successful business. A lot of this is luck that these people were available or interested in leaving their otherwise stable and high earning professions to do something crazy with me and get it all started.
What was your role in the company as you rolled it out?
I’m the founder and CEO. As the founder and CEO, your job changes dramatically as the company goes through. There’s a great article. I forget the exact contents but the title was You’re not the CEO – you’re the F-ing Janitor. That’s true. In the early days, I was coding a website and the company was out of my apartment so I was the janitor and the chef. You do recruit great talent as a big part of what you do. You raise the capital necessary for a business to run.
You need to report to the investors the metrics, how things are going and continue to get buy-in. I did a little bit of everything. For those who are thinking about starting these kinds of companies, I wouldn’t do it any other way. I now know a little bit, enough to be dangerous, about marketing, sales, coding, engineering, product management and metrics and analytics of the business to give me enough insight to know, for example, who’s a good hire, who’s not, who’s working out.
Frankly, when you start a company before you have something, in Silicon Valley parlance, is product-market fit, which is to say, something that people like and would pay money for essentially. It takes a long time for many businesses, including ours, to get to that place where you feel confident. After you have product-market fit, you should invest in hiring out dozens of people to throw at selling and marketing the product, etc.
You, as the CEO, become the person who hires those people, manages those people and makes sure they’re all doing their jobs. Before you have a product-market fit, all you need to do is conserve money because you’re trying to do this with as small a team as possible. You’re a small scouting expedition. Everybody is everything. It’s a fun moment for the company because it’s a real struggle. Most companies never survive past that stage, but your role as CEO shifts super dramatically as you go from that first stage to that second stage.
We’ve set things up well for you to tell us now exactly what it is that Casetext does. I know there are some different things you’ve got going on. You mentioned something that stuck with me. It was this idea of getting digital copies of all of the law. To set it up here, I know that this is a legal-assisted research and writing platform. That was my layperson’s initial description but tell us from your perspective what Casetext does.
We help lawyers do great research and writing. What that means is a lot of different technologies and tools that support a lawyer in the office. On the one hand, our first and initial tool called Casetext Research is an AI-powered legal research ecosystem that looks to a lot of people the traditional old players with a lot of things that that we do that they don’t. We have all the laws, statutes, regulations, cases, etc. for the federal and state.
It will have great search built on top of it and we believe better search on top of it. It’s like the search where you ask a question and you get a direct answer from case law or write the statement that you wish would appear on your brief. Have you ever had one of those moments where you’re like, “I know there’s a case for this proposition but I can’t quite find it where I’m looking?” Write in that proposition of law into the search bar in full sentence and it will find you exactly that case you’re looking for.
On the one hand, we are a traditional legal research tool with some superpowers built in. On the other hand, there are a number of other things that we do as part of trying to solve the whole problem for a lawyer. One example of this is our second big product called Compose. The way Compose works is either within Word or within our online editing environment. For example, I’m working on a Fair Labor Standards Act case in the Fifth Circuit Court of Appeals or in the Eastern District of Texas and I’m trying to move for summary judgment our protect summary judgment on appeal or what have you.
You give it a bit of context. Based on that context, it outputs for you all the issues that might come up on a summary judgment motion for a Fair Labor Standards Act case. Within each of those issues that come up, what are all the standards of law that a judge might apply to determine which side ultimately prevails on that issue? Within each standard of law, what is case law from your district in your jurisdiction that you can use to support that proposition?
With all of that, with one click you can add it to a draft brief. That’s an example of the thing that we do to go beyond search and find a case. With this technology, we’re seeing with the now many hundreds of firms that have adopted it, you can get out the 2nd year and 4th year associates inequality. The first draft of a legal argument section of the brief and all the standards and the arguments, etc. within maybe 20 to 30 minutes and skip over the, “What is the three-part test for this thing?”
What is the exception to the third part of the third test? All that stuff, you may have to dig through case law or maybe find good treatises. It’s all there and it’s all added to a draft of your brief with a click. The last thing that we do and this is new for us so you won’t even see it that much advertised on our website, etc., is to enable an attorney to do the advanced searching you do on Casetext against the law against whatever set of documents you might want to point that search engine at.
We’ve seen, for example, in the context of appeals, the appellate records can get sometimes meaty. You have deposition transcripts, trial transcripts, the key evidence turned over, the drafts of all the briefs that were filed, etc. All of which might be the things in the appellate record and they can be 10,000-plus pages. We’ve seen a lot of lawyers who’ve been given access to the beta version of this new technology, the ability to drag and drop and upload all that into this system.
What you get in return is that same super-powerful searching where you can ask questions. An example of contracts, you’ve seen people upload a bunch of contracts and you can ask a question against thousands of contracts, “Would the company be liable if a terrorist attack disrupts the supply chain?” It’ll point you right to the force majeure clauses of those contracts. Even though this is a higher level of understanding of language that without even mentioning terrorist attacks or supply chain, it gets it. Those are the technologies that we’re working on. At a high level, it’s about helping attorneys do what I know appellate attorneys do a ton of.
According to a study from the ABA, 1 out of every 3 hours of an associate’s time is doing online legal research. It’s a massive amount of time being spent there and the challenge that we have to the status quo, is that all necessary? Is that the best use of an associate’s time? Is that digging for documents? Is that what we all went to law school for? There’s certainly a piece of the legal research process that is a learning process, necessary and good.
We don’t want to eliminate all of that. If you’re trying to find the general three-part test that everybody knows and you need to find a KeyCite for it or if you find yourself in that position of going, “Next, next. No, no, no. Yes, this includes my words but no, it’s not what I meant.” If you do that for hours, that’s the stuff that we’re trying to eliminate from the process.
You take what I would call the Funneling Process and cut the top part of the funnel off so when you’re getting into it, you’re getting the real meat of what it is that you need.
That’s exactly the idea. Whether that be through amazing search or technologies like Compose. If we think about our regular lives, in my personal experience if I ask the question to Google I get an answer in a box at the top. At this point, we’re not all the way there yet in Casetext. We’re going to be there soon, but at this point you don’t even have to dive through 20 or 30 different web pages. You get your answer.
If it’s a health problem, it’s almost always telling you the answer. It’s like, “I was right.” In general, what is the showtime for this movie? What Chinese restaurants are open on New Year’s or whatever in my area? You get those answers directly to you. You don’t even have to go digging anymore. That couldn’t be further from most people’s experiences, especially using some fancy products, and we’re trying to end that. Imagine what your life would look like if Google was like legal research. You’re spending hours trying to find answers to sometimes basic questions. It would be difficult.
What’s been the response in the legal market to this? I know you have signed up a fair number of larger firms. This is the tool that to me and my former practice as a solo that I would have liked because it would have enabled me to leverage myself better with the tools that you have to offer. It looks like there’s a lot of AI algorithmic application here that to get to those nuggets of exactly what it is you’re looking for that would help someone like that save a lot of time in their practice and therefore be more profitable. Has there been skepticism or resistance expressed to those that you talk to about these products?
Absolutely. The response has been phenomenal and we’re super delighted that we are where we are. At this point, 8,500 law firms are subscribers to the system. That includes 40 in the 100. Also, it includes many thousands of people at 2, 5, 10 or 15 person firms. They made those kinds of firms as well. The firms that have joined us in using this future-oriented technology are doing it for those exact reasons. If you’re a smaller firm and if you are up against a well-financed defendant, for example, you’re on the plaintiff’s side, they can throw money and people at the problem but you only get yourself. Technology is the great equalizer.
We’re able to make our products cheaper than the alternatives by a good substantial margin because we apply some of the same machine-learning techniques that make the product better internally. For example, we will compare our version of Shepherd’s or KeyCite, which is called SmartCite in Casetext. We put that toe to toe with those and we analyze our discrepancies.
They’re few but what we don’t have is thousands and thousands of editors behind that system. We have smart computers and a few people to get exactly the same results. That’s why we are able to charge lower prices for good, if not better quality stuff. On the one hand, you see the lower price point, the ability to leverage your time better. That being an important piece.
There’s another fundamental piece that you talk to people on the ground, which is a little mushier but was a company founded by lawyers who know how to code. That’s not me, it’s a number of people at the company. What you get by that is not great technologists but people who have a deep respect for the law and legal practice. Not like a bunch of MBAs behind the scenes trying to tweak things here or there.
What is the output? We’ve heard from our customers that it’s easy to use, intuitive, built by litigators for litigators and it’s something that a lot of the intangibles that attract people to modern smartphones, for example, compared to their earlier antecedents. A lot of those things matter when it comes to making a product that people like. Are there skeptics? Absolutely. The skeptics come in a few buckets. The first is why do I want to be more efficient, I’m billing by the hour?
Truthfully, those people are not good clients for us. That’s not the way that we view the world. It’s not the way that most of their clients view the world. We would rather align ourselves with the people who know and understand that with more efficiency means you can spend more of your time doing the most important stuff. Maybe even bill the same amount of time, frankly, but it’s not going to be trying to find that Darrin case or trying to find that three-part test. Somewhere in the appellate record, I know it but I don’t know exactly what page it’s on. It says something about cost.
Using AI technology means that if in a deposition, the deponent says, “This cost too much.” You don’t remember exactly the words that he used. You can search for the concept of the price being too high. Even though there are no keyword matches, it finds you the right place. It’s those kinds of things. We talk to people who have been in exactly that circumstance where they spent hours searching for the wrong word that was billed to the client. They’re spending hours searching for a word difference. They would have saved themselves many hours. That stuff needs that and people who want to bill on that time, not good clients for us. I was a skeptic myself about a lot of modern artificial intelligence technologies. In reality, this stuff is only getting good recently. I’m talking about the last couple of years.
In the earlier versions of our product, it did not claim to use machine learning artificial intelligence because we didn’t because the technology wasn’t good. We did a lot of things around automation and making your lives easier through smart algorithms. The technology has gotten good. One of the things that we could show you are in the labs. We start writing a sentence and it will write the next paragraph for you with the correct citations to law. The standard on the motion for summary judgment is. It’s like, “There must be a material issue getting the facts.” It will fill out the next paragraph for you. It’s getting scary good.
A lot of people correctly are like, “I feel a little uncomfortable with what that means for their practice, what it means for their role.” The products like Compose give you all the issues and language. People feel a little bit like they want that ownership over the process. What we found with most folks, though, is that they start to view this as a support rather than doing the job for them.
As if you’re a solo but you’re able to hire a 2nd, 3rd or 4th associate to give you that first draft and the efficiencies that come by having that support under you. As opposed to, “This is like the Terminator, Nexus, Skynet or whatever.” Some of the stuff that we’re working on is not public yet like the thing with the paragraph. Even for us as attorneys inside a company, I can’t believe they can do this now. It’s hard to even fathom. You assume it’s going to be wrong and bad but it’s not.
Some of the reactions there are natural. As with any new technology, at first people are like, “I don’t know where this fits in. I don’t know how to use this.” Eventually, it’s viewed correctly as I’m a professional. There’s a lot of stuff for me to do still, even if I’m getting support from below from the technology to help me with this.
Have you seen either adoption or awareness and maybe pressure from the client side to adopt these types of machine learning efficiencies, best practices on the firm side?
1,000%. What you’re starting to see from a lot of clients is in their request for proposals for the different law firms they are shopping with, they’ll include how you are using machine learning artificial intelligence technology to be more efficient? I spend my time doing legal research and writing, it’s not about that. It’s about reviewing thousands of contracts from companies like Cura on the cutting edge of this. It’s about document review. Companies like DISCO and Everlaw are on the cutting edge of this. There are a lot of companies addressing both the transactional and the litigation sides.
Litigation will matter more for folks probably reading this one. They are making big steps and clients do care generally about efficiency, always have and always will, or will always be asking for ways that we as lawyers can serve them better, more efficiently and effectively. You’re starting to see that come into explicit requests for exactly how you are doing this new practice.
On the plaintiff’s side, it’s a different story. I know a lot of folks both reading this and that use our products are on the plaintiff side. They are the clients who are not sophisticated buyers of legal services. Your incentives are aligned with being as efficient while winning. You see that as well there. Whether you’re a 4 or you’re a 5-person personal injury shop. They’re both sides of these things, both from the biggest companies to representing somebody who got injured in a medical malpractice case. We’re seeing that interest across the board.
Have you seen any interest from the judicial side of things for this?
Yeah. Some of our technologies that I haven’t talked about yet, there are some small things but important things like drag and drop in a brief from opposing counsel and see all the cases that they are relevant but they might have missed. It was one of our first products. Judges love that product because they can pull in the digital copies or the scanned copies of the briefs submitted to them and see, we didn’t talk about this a little bit.
Cynically, what were the things that you’re hiding for me? There are these great blog posts that are written by judges at the state and federal level about using this technology and the fun things they are finding while using it. You have seen a lot of clerks and judges use our searching technology that makes searching super easy. All those things are being adopted. We made one mistake early on, which is we made it free for judges. At the federal level, we got in trouble because they’re like, “Are you trying to bribe judges?” We’re like, “No.”
It’s going to be a longer process where we have to sell to judges and the court system. We thought we were doing a public service. They were like, “We’re not going to gain any from judges anyway. It doesn’t make this free.” They’re like, “What if you end up in litigation before that judge? It’s easier if your product is free. You can have that.” You have to be thinking about these kinds of things. It’s funny as lawyers as long as you didn’t cross the line.
I hear you talking about the different uses of the products and one thing that I thought of was whether there was any room for a customized application along the lines of, “We’ve got our firm’s brief bank. This is the way that we present an argument. These are the cases that we like to cite.” Using your example of summary judgment, I was like, “How many times have I gone back and looked at a summary judgment motion I wrote years ago to get the law that I want, get it stated and cited a certain way?” Is there anything like that on the horizon with Casetext as far as being able to use your own work product and apply it?
That’s exactly where things are headed. We’re starting with something that looks a little more in the post product that helps you see all those standards of logs. We’re starting there from something that looks a little bit more like a treatise or a practice guide, in a sense, which is to say that it has pre-programmed in the general standards and a good way of stating them but it may not be your way.
We’re starting there in part because we know that a lot of people who are completely switching from some of the older products to our product miss that. They missed having the comfort and the safety net of the practice guide that spells out whether it’s O’Connor’s or Rudder’s in California or Wright & Miller on the federal level. They spell out, “Here are all the different issues that might come up. Here are all the standards, etc.” It plays that role to some degree, but 1,000% the question we get asked a lot is, “How do I take this to the next level and make it more about me?”
This is still in the lab. There’s not a product that we have available yet, but the products I mentioned earlier where you start typing the standard piece of it as it fills up the rest. The way that works behind the scenes is you feed it in dozens, hundreds or thousands of examples and it starts to pick up on your language, essentially. That’s why it is able to spit it back out to you. The version that we’re working on that’s more general, we feed it in all the law.
It’d be an easily imaginable world where you feed it in all of your firm’s briefs and or all of your briefs and it uses that to go off of. There’s still a danger there. One of the reasons why tools like Compose and also frankly, treatises, more traditional sources of treatises and practice guides have a place is there’s a group of people behind the scenes every day in the case of Compose. It gives treatises maybe every year making sure all the laws are good and and if they have changed.
The case that you are citing are not going to be overturned on some other grounds. They are a little bit iffy when they run it through Shepherd’s. There’s always going to be a place for those resources, but most attorneys do exactly what you said, Todd. Look up what you wrote years ago and use that. We want to accommodate that and make that easier, better and faster.
I know that there have been conversations nationwide about adding technical competence to ethical standards governing attorneys. Have you seen that movement? It seems like you have a nice synergy here with your practice to help with that adoption.
We have seen that movement. It’s definitely been more widely adopted across multiple states. The reality is that we are decades away from technologies like ours on the far cutting edge to be a requirement for technical competence because many attorneys are still, “Do I use email? Do I use online legal research at all?” Truly to be competent at your job, you don’t need to. We’ve been practicing law for many hundreds of years before there was AI-assisted review and discovery and these more advanced AI technologies for research and writing.
There’s one reality in the world. The world’s getting more and more complex. There is infinitely and exponentially more data nowadays even amongst court cases than 10 years, 15 years or 20 years ago. In some ways, you do need these technologies to keep up but that you can be a fully competent attorney without having to adopt technologies that are still there.
AI assisted review on discovery and AI contract review on contract review stuff. That’s the reality. The reason you should do it is not that you’re required to do it, it’s because it will make you a better lawyer. To your point, there’s a broader conversation. That is pushing people in a new direction from where they were before. That is opening people’s eyes to, “We got to take this more seriously. We got to look into this. Our clients are asking that for us.” I feel like I’m working 90 hours a week getting nothing done. We’re not getting enough done. We need to choose the way we’re doing things.
I remember when not that long ago, we moved into electronic filing and moved into electronic appellate records where everything’s PDF searchable. What a difference that simple change made, much less having machine learning AI technology that can then take that to the next level. It sounds great to me.
It all happens in stages. Honestly, you need the world to get to that place where you have electronic digital records that are PDF searchable for machine learning processes like ours to even exist. The product that’s still beta but the one where people are uploading their appellate records and doing searches against them, the user experience will be terrible. It’s like, “Take my 10,000 pages and scan it to get into the system.” Before scanners, tough luck. It’s this progression to the future that requires simple stuff like digitizing, making it searchable first so machines can read it, access it and use it. Now that opens up a whole world of opportunities for the future.
We talked about how AI has taken off in the last few years. Where do you see it going aside from these future projects that Casetext has going on? Where are we going to go in the not too distant future with AI machine learning and how they’re going to affect our practices?
What is AI even? The way I look at artificial intelligence is it’s taking certain aspects of human intelligence and making it so machines can do it to some degree that mimics human intelligence. Outside of some of the stuff that we were talking about so far, there’s speech recognition. It’s a miracle that you all can understand what noises coming out of my mouth and turn that into words. That’s a huge aspect of human intelligence. Things like Siri and Alexa are okay at. To be honest, that’s where we are now.
There’s computer vision, which is taking these beams of light and turning them into the images we see in front of us. It’s amazing. It’s another aspect of human intelligence. Robotics, in the field of artificial intelligence, is our interaction with the physical world, not bumping into things, walking over difficult terrain and knowing how to balance. All these artificial intelligence and the fields that matter most for law are probably natural language processing, understanding and pattern recognition, otherwise known as machine learning. All those things are the technologies that are getting dramatically better every single day, and you’re seeing this across the board whether it be cars that drive themselves or you talk to your phone, it gets what you mean 90% of your time. It knows to call my wife and not the police or whatever.
All these things are getting dramatically better. We’re seeing, for technical reasons and for hardware reasons, their rate of change is accelerating dramatically. You’re starting to be able to train a machine learning process or a natural language program around many billions of documents and words. Also, to run a system where it’s learning and teaching itself over and over again in what they call parameters and the trillions of parameters. It’s starting to feel like we’re getting to a place where it’s like the neurons in our brain.
What does that mean in terms of legal practice? One of my cofounders, Pablo, the guy who is a total legal innovation genius, his vision for all these things and he jokes is called he calls it Law2-D2. Imagine that it’s like the robot that would follow you around more virtual than physical. It is doing a lot of the rote work for you. It could be everything. It could be drafting the first draft of a document. It could be knowing what things you need to search for before you even know to search for them and surfacing them for you on your desktop and are in your email every morning.
It could mean finding the document of the many millions and you discovering and servicing that for you parsing through thousands of contracts and pointing out the clauses that might be problematic. All those things are possible in the near future. What that means for us as a practice is liberating. Google has accelerated the pace of our knowledge and understanding dramatically by making it so effortless to look something up and put us in a position where we aren’t spending a lot of time searching for information but using information. The legal profession as a whole can move a lot more in that direction. It’s an information-intensive industry.
A lot of what we do is researching, writing, parsing through documents, thinking through what’s analogous and what isn’t and there’s so much more that we could be doing all the time. Time is our most precious resource not because of the billable hour but maybe there are only so many hours we are alive for in practice. What practice might look like in 5, 10 or 15 years is you are spending 100% of your time grappling with the most difficult questions like, “Is this case truly analogous to the cases of the past?” Those have all been surfaced for me, the facts that have been summarized. There are key statements that have been pulled out for me and now as an attorney, I’m going to sit and think. What a luxury to sit and think for a little bit and think through. Is this case relevant to me? How can I make the best case for my client? This is a correct analogy to the past.
Of the 400 pages of deposition transcripts that came out of this one person over the course of a few days of depositions, what is that one page that is going to be the thing that I’m going to hammer in a trial or bring up again on appeal as evidence that there’s a plain error or whatever? To spend all of your time also working with clients and being a strategic adviser for them, which is what they want.
They don’t want you to spend all your time in a truffle pig rooting for the morsels of good stuff in your house of documents. It’s what the future looks like. It’s bright. What you see time and time again, if ever another technological revolution came through based on other industries, you don’t have fewer people doing it. They’re not making less money. They’re making more and they’re doing more things.
They’re doing more impactful things or they’re trying new business models and alternative fee arrangements or fixed fee arrangements. Potentially more contingency work, even on the defense side. All those things will open up and be available once you become more efficient, effective and intelligent about the way you interact with the work. It’s an exciting future for the profession. I think we will look back on this episode and be like, “How barbaric.” You spent so many years training and working and you’re not spending all of your time doing the stuff that you trained and prepared for.
What do you say to the lawyer who is worried that AI is going to replace them or replace their industry?
The short answer is it might replace the things you’re doing but I can guarantee that most people, not everyone, but the vast majority of people have a lot more to add than filling out forms, finding cases or searching through discovery to find that one document. There’s a lot more to add than that. That’s what clients are looking for. Those folks that are most concerned will feel, in many ways, the most relieved later to find that they’re not doing those things anymore. They’re doing a whole new category of things or new business models.
Talk about automated industries. Not so long ago, they came up with this technology, the automated teller machine. Imagine being a bank teller and all of a sudden there’s an automated teller machine nailed on or screwed on the side of my building. That would be terrifying. There it goes. There goes my job. There are five times more bank tellers than there used to before automatic bank teller machines or ATMs. The reason for that is a lot of these banks realized that with the advent of the ATM, they can open up many more branches and be much closer to the customer. It opened up a whole new business model that nobody’s even thinking about before the ATM. ATM is not like a fancy guy technology. It’s simple stuff. They’re able to go through this revolution earlier and see the effect.
Where I live, there’s a Bank of America as much as there is Starbucks. I can go to one in any corner. Bank of America is probably not as prevalent where you are but it’s probably the same story with different brands than anything else. It opened up the possibility for ATMs to be in communities that never make sense to open a bank branch because you had to staff it with 13 or 15 people and having fifteen people at that corner of nowhere else would not make any sense from an economic perspective.
If you had one person at that ATM machine and that happened, it makes sense because of the customer. That would be more accessible. I don’t know exactly what it will look like for law, but we’re almost certainly going in the direction of bank tellers. There will be many more opportunities for lawyers to create new different things, change the business model, change the way you charge and practice that will be more empowering, better for the customer and better for the lawyer. That’s the history of technology.
It sounds like there’s a big spot in there to close the access to justice gap that exists.
A lot of people talk about this. The statistic that I hear a lot is 80% of legal needs in America are people who’d never see a lawyer. Part of that isn’t necessarily a price issue. Some of it is a lack of awareness, knowledge, information or desire to get a lawyer and try to handle things without them. Say it’s 50%, 40% or whatever.
What that implies is there are many hundreds of thousands of people who have a need that can be helped by a lawyer. It probably does come down to some basics about, “If I go to a lawyer for this.” The average medical malpractice case costs about $100,000 in legal fees. If you don’t have $100,000 ready to put out if they’re not taking contingency as a plaintiff or if you have to defend that and for whatever reason, you’re not fully covered, the insurance bill goes up. Either way, it’s $100,000 out of pocket.
You guys know better than I do what it would cost to take something on appeal. Say you lose but you’re not somebody from means but you think there might be a legal issue wrong here. Maybe too bad. I’m sorry. You cannot have this case heard on appeal because you cannot afford it. One business model change here might be around being able to offer the same or better services for cheaper. It’s as simple as that.
I’m sure that there are those who would say that AI and machine learning are not all good. There are certain aspects of it that I’m thinking more about from the perspective of the lawyer-consumer. It’s all the rage that even we know about it. What are the things that lawyers should be on the lookout for and maybe a little wary of?
First things first since I’m a little self-interested because it’s something we do, try products that have free trials. If they make you buy before you try, major red flags. Despite the major advancements in the general field of artificial intelligence everywhere including law but also in finance, etc. to this day, there are a lot of companies out there that say they have amazing AI technology and it’s going to change your life, etc. They may sound exactly like me and look like me but they don’t offer a free trial for the product for a reason. That’s one thing to watch out for.
The second thing to watch out for is not to blindly trust it. Some of the old, even for legal research in the field we feel good and competent in old trust and true techniques like Boolean searching, etc., in the legal research space or similar techniques that hand review things. AI is good but it’s not perfect. We will never claim that it is or will be, frankly.
There is always going to be that place for you to dig in, hand review, do a different search, open up a different document or view on things, print out things and start reading them. Oftentimes, I’ll be like, “I don’t need to do that. Everything’s fine.” This is the 1 in 10 chance that this technology went off the rails. Even for some of our newer technologies, not yet released, the technology where you start writing and it completes your sentence or paragraph.
The reason why it’s not yet released is 9 times out of 10 that’s exactly right and 1 time out of 10 it will flip the words defendant and plaintiff in the most important place. That’s why we can’t release it to the world yet. We’re cautious. Having practicing lawyers, we’re cautious about the things that we release and what to put our name behind.
It’s highly encouraged that people double-check that work. Not unlike how you might double-check the work of, say, a first-year associate. That person’s smart, hard-working but let’s read this carefully. There are a few things to watch out for there. There are still some of those snake oil salesmen out there. That will always be the case in technology and to be totally clear, we as a company aren’t perfect. We are doing our best.
Our technologies sometimes make mistakes. We are fast to admit that and to fix it. Being practicing lawyers, we know the importance of getting every single thing exactly right always. We aim for that level of perfection but once in a blue moon we’ll hear, “This was reviewed by the Texas Supreme Court and you didn’t have that review grant petition.”
I practice in California and Boston, so I don’t know how it all works in Texas. Their stuff is the only solution there and this is, by the way, true with Westlaw, LexisNexis, Bloomberg and Relativity. Every technology provider, new and old, is to do some of your own work and always double-check. Be the human in the system. You’re going to save a ton of time still using new technologies because you’re only doing a checking role as opposed to all the upfront work, but we recommend you do that check.
It is the time of year when law firms do have new lawyers starting with the graduates and bar takers who are perhaps waiting on their bar results. They’re entering the practice of law at a far different time than when I did back in the mid-‘90s and even Jody in the early 2000s or mid-2000s. Considering where things have gone in the last few years, what’s the biggest piece of advice you would give to someone who is now standing on the precipice of beginning their law practice? From your perspective, what should be at the forefront of their mind going forward?
This may sound funny coming from a tech guy but none of it would be about technology. The technology, in the grand scheme of things, is a relatively small piece. Whether because of COVID working from home or in the office, whether you are using the most advanced newest AI technology or doing things the old way or whatever, there are certain things that serve all lawyers, especially new lawyers, well. Have extreme attention to detail, put in the work to stand behind the statements you’re going to make and everything you write, whether it be to the partner or to the court.
Think introspectively about the work you want to be doing because you’re going to set yourself up now for hopefully a long career. Don’t be shy to ask for mentorship and guidance, if you’re joining a firm, for example, as opposed to starting your own, from the more senior partners around you. What do they enjoy about their work? Where do they draw joy from it? See if you can find that same joy that they found or find other areas of work within the firm that might be interesting to you.
When I started at the law firm, they had a system where they would randomly assign at work to be fair about who gets how many billable hours people get. I’ve always been a little bit of a rule-breaker but there was this one partner there, Doug, who was doing, in my opinion, what seemed to be the most interesting work. On day three of my job I walked into his office and said, “Can I work with you? I want to talk with you about the work you’re doing.”
He told me what he did and that confirms it. It’s super interesting. I was like, “Can I work with you on this stuff?” To my surprise, not a single other of my 100 other associates as far as I’m aware, even attempted something similar. It was whatever fell into their lap. I’m like, “I’m a privacy lawyer now.” I’m like, “Do you like privacy? Are you interested in privacy?” No, but that’s what I’m doing. There are so many opportunities to take control of your own fate as a young lawyer. Whether you’re at a 5-person, 20-person, 1,000-person firm or starting something on your own, try to find that thing that’s going to keep you passionate and excited about the work that you do because your first few years are going to be the hardest, frankly.
At least for the first few years, hopefully for the next 50. Not enough young attorneys. In my experience talking to a lot of young attorneys, they feel like, “This is handed to me. That’s what I have to do. It’s the way it has to be.” They don’t ask themselves a question enough. What do I want to do? Where can I make the biggest impact? What will inspire me? What will bring me joy? With the appellate practice and my research and writing, I’m assuming all three of us are, am I a trial guy or girl. Am I a person who’s going to enjoy doing contracts, helping businesses build and negotiating things? Be introspective about that, that would be my advice.
As we’re reaching the end of our time and wrapping up, where can people find you online? Where can they find you? Where can they find Casetext if they want to know more or follow what you all are doing?
We’re Casetext.com or @Casetext on Twitter, Casetext on LinkedIn and Facebook too. I’m super accessible. If anyone wants to reach out to me personally, I’m Jake@Casetext.com. You can also reach out to me via LinkedIn, Jake Heller, and I love to talk to people. I have a pay-it-forward thing. If somebody has a question, wants some help, wants help finding a job, a clerkship or whatever, what my work experience is like and all those things, I’m happy to do it. Life’s busy so it may take a few weeks but I’m willing to listen and talk.
Thank you for talking with us. This has been interesting. It’s amazing to think about where we sit versus where we started and where we’re going with all this stuff.
It’s going to be wild in the next 20 to 30 years. That’s based on what we’re seeing in technology and everything else. It’s a brave new world.
We sure appreciate your time, Jake. Thanks for being with us.
Thanks for having me.
Important Links:
- Casetext
- Y Combinator
- You’re not the CEO – you’re the F-ing Janitor – Article
- Casetext Research
- Compose
- @Casetext – Twitter
- Casetext – LinkedIn
- Facebook – Casetext
- Jake@Casetext.com
- Jake Heller – LinkedIn
A special thanks to our sponsors:
About Jake Heller
Jake is cofounder and CEO of Casetext. Jake was named to Forbes 30 Under 30 and the Fastcase 50 for legal innovation for his work at Casetext. Before starting Casetext, Jake was a litigator at Ropes & Gray. He graduated from Stanford Law School, where he was president of Stanford Law Review, and clerked for Judge Michael Boudin on the U.S. Court of Appeals for the First Circuit. He’s a Silicon Valley native and has been programming since childhood.
Love the show? Subscribe, rate, review, and share!
Join the Texas Appellate Law Podcast Community today:
