Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

Podcast: Data Science for Judicial Analytics

How AI is Transforming Litigation Risk Assessment and Prediction

Using Data Science for Judicial Analytics (Dan Rabinowitz, Founder, Pre-Dicta)

Chad Main, an attorney and founder of Percipient, combines a strong background in litigation with a passion for technology. Practicing in Los Angeles and Chicago, Chad established Percipient to leverage innovative software, enhancing the quality and efficiency of legal work. His firm focuses on reducing time and costs for legal and compliance teams, ensuring effective results. Chad also hosts the Technically Legal Podcast, where he explores the intersection of technology and law, demonstrating his commitment to integrating tech solutions into legal practice.

Chad’s unique blend of legal expertise and technological innovation makes him an ideal interlocutor for Dan Rabinowitz, founder of Pre/Dicta, a firm that predicts litigation outcomes using AI. With attorneys at John Quinn’s firm already utilizing Pre/Dicta’s services, Chad’s insights into leveraging technology for legal efficiency and his practical experience with data science in the legal field will contribute to a meaningful discussion on the transformative impact of AI in legal practice.

Podcast Summary

Dan Rabinowitz, a former Big Law lawyer turned tech entrepreneur, shares insights on his company Pre/Dicta and its judicial analytics app. Pre/Dicta uses data science to predict judge rulings by analyzing biographical and demographic data, incorporating both publicly available information and proprietary data to create detailed profiles for each federal judge. This innovative approach offers lawyers a powerful tool for predicting outcomes of motions to dismiss and other legal proceedings, enhancing strategic decision-making in litigation.

The conversation delves into the broader implications of leveraging data analytics in the legal field. Rabinowitz highlights how Pre/Dicta’s platform enriches legal data by considering the identities of parties and attorneys, identifying patterns that inform judicial predictions. The discussion also covers the ethical considerations and potential limitations of current data analytics, emphasizing the importance of understanding classifications and categories in legal data analysis. This forward-looking approach is set to transform the legal landscape, providing more accurate and strategic insights for legal practitioners.

Technically Legal Podcast Link: https://www.tlpodcast.com/using-data-science-for-judicial-analytics-dan-rabinowitz-founder-pre-dicta/ 

  • Legal tech, judicial analytics, and entrepreneurship with former Big Law lawyer turned tech entrepreneur Dan Rabinowitz 0:10
    • Dan Rabinowitz, former Big Law lawyer turned tech entrepreneur, discusses his company Predictive and its judicial analytics app.
  • Creating bottled cocktails for home consumption, with a focus on legal background and experience. 1:36
    • Speaker 2 has a passion for cocktails and wants to bring the craft bar experience home with their company.
    • Speaker 2 is transitioning from direct-to-consumer sales to larger production and distribution channels.
    • Speaker 2 worked at Sibley & Dolan, then became a Supreme Court lawyer, handling government contracts and fraud cases.
    • Speaker 2 investigated Edward Snowden’s leak as a Booz Allen contractor, gaining experience in internal investigations and systems work.
  • Leveraging data analytics in legal field. 7:08
    • Speaker 2, a General Counsel at a data analytics company, discussed leveraging commercial technologies in the legal sphere.
    • Speaker 2 explained how a company used technology for an investigation, sparking the idea of applying it to legal issues.
    • Speaker 2 had an idea to create a company that analyzes judges and courts using data.
    • Speaker 2 considered returning to law after the US shutdown, but instead pursued the predictor idea.
  • Predicting judge rulings using data science and publicly available information. 12:15
    • Speaker 2 brought cutting-edge experience and capital to the startup, while Speaker 1 bootstrapped the business.
    • Speaker 2 was impressed by the potential for forward-looking data analysis in the industry, leading to the creation of the company.
    • Predictive Law uses data science to predict judge rulings by analyzing biographical and demographic data.
    • The company’s model incorporates publicly available information and proprietary data to create a unique biographical profile for each federal judge.
  • Using data analytics to classify legal cases. 17:17
    • Speaker 1 discusses using data analytics to classify individuals and organizations based on their characteristics, rather than looking at them individually.
    • Speaker 2 explains the limitations of current data analytics, including the inability to get down to a super granular level and the potential for inaccurate or incomplete information.
  • Using data analysis to predict judges’ rulings based on past decisions and demographic information. 19:53
    • Speaker 2 highlights the importance of understanding classifications and categories in analyzing legal data.
    • Speaker 2’s platform enriches one-sentence orders by accounting for parties’ and attorneys’ identities, and finding patterns within the data.
    • Speaker 2: Judges’ biographical data + past rulings predict future preferences.
    • Speaker 1: Combining biography and rulings yields highly accurate predictions.
  • Leveraging data to predict court outcomes. 24:51
    • Speaker 2: Rick Merrill’s work on judicial analytics was a pioneering effort, creating technology and interface to surface state court data.
    • Speaker 2: Catalytics acquired Gavin Linux’s state court data, leveraging Rick’s work to forecast judicial outcomes, providing a more advanced service.
    • Speaker 2: Focus on predicting state court judges’ decisions, starting with motions to dismiss.
    • Speaker 1: Important information can be gleaned from motion to dismiss, including judge’s leanings on summary judgment.
  • Predictive analytics for legal cases. 30:28
    • Speaker 1: Judge’s rulings can be predicted based on similar cases.
    • Speaker 2: Tailored statistics for case-specific judges will be provided.
    • Analyzes judges’ jurisdiction scores and individual likelihood of granting motions to dismiss.
    • Predicts venue transfer success in Central District of California based on judges’ past decisions.
    • Predictive analytics platform charges per case, no user base or time-based system.

SPEAKERS

Chad Main, Dan Rabinowitz

 

Chad Main  00:10

I’m Chad main, the founder of legal services company percipient. And this is technically legal, a podcast about legal technology, legal innovation, and the impact tech is having on today’s show, I have a conversation with Dan Rabinowitz. He’s the attorney and the founder of predictive that’s a company that uses data science for traditional analytics. My conversation today is with a former Big law lawyer turn tech entrepreneur Dan Rabinowitz, after stints with law firms, the DOJ, and time as general counsel, Dan’s going to tell us all that led to the founding of predictive predict is an app that uses data science to tackle judicial analytics. But unlike a lot of software out there predicted doesn’t just look at the judges opinion and track record. It looks at other factors influences the judge’s decision, like a judge’s net worth of judges political affiliation, their educational work experience, along with other biographical data points, you may read about predicted recently in the legal tech press, because it just bought Gavin Linux. That’s another judicial analytics company founded by Rick Merrill, who has also been a guest on the show way back in 2018. predictor was originally focused on federal courts, but by joining forces with Gavin lithics, the company got a trove of info about state court judges and opinions. But Dan’s not just all about judicial analytics. Before we get into predictor, tell me a little bit about deco cocktails.

 

Dan Rabinowitz  01:34

Oh. So I have a number of varied interests, if you will. And one of them is cocktails, alcohol, I’m with you. There you go. And for many years, I’ve been doing cocktails, I’m making it for my friends, we have like a weekly get together where I do cocktails, and just have a good time. And my partner who also loves cocktails, but cannot make cocktails, always have this dream of producing ready to drink or ready to sip as we like to refer to ours. And he approached me, he’s in real estate, I’m a lawyer, and he called me and he called me up. He’s like, Hey, I have this great idea. And I’m like, we don’t really have much in alignment here. Like, no, we both love drinking. And so we started this company. And we produce these really beautifully done, I think, bottled cocktails, were actually in the process of revamping our formulation, and our and our bottling and marketing and labeling, I’m actually going out to LA, three or four weeks from now to talk with a new manufacturer. And essentially, what we’re looking to do is recreate the craft bar scene at home, so you can’t get out or you’re not interested in getting out. But we want to produce a cocktail of that quality, that complexity in a bottle that also evokes that like great cocktail experience where you’re like you’re in a bar, it’s like, you know, beautiful mahogany, great bartender, and so on. So that’s really what we’re looking to bring to the home consumer

 

Chad Main  03:03

work we buy it now is it just in the DMV there in the DC area.

 

Dan Rabinowitz  03:07

So right now, technically, it’s direct to consumer, we’ve sold out our first run, and now we are transitioning to a larger producer. And then we hope to get through both DTC direct to consumer as well as to get into normal distribution channels. So where are you located? Chicago. Okay. So hopefully, we will be there at some point in time. I can’t say that’s our first market. But at the very least, you’ll be able to order online probably in the next three to four months. Well,

 

Chad Main  03:35

I digress. Because this is technically legal. It’s not technically liquor. But when and why do you want to become a lawyer?

 

Dan Rabinowitz  03:40

When in life I want to become a lawyer, I think pretty early on, I want to say, probably a stupid decision to make so early and light yourself up for that. But probably pretty early on I I thought I wanted to be a lawyer, there was this guy I knew who was was a litigator, and you know, he was always sort of model for me, and always wanted to do that. Whether or not that’s the best thing to decide when you’re looking at old is another conversation. I

 

Chad Main  04:06

guess. You come from a family of lawyers. No,

 

Dan Rabinowitz  04:08

not at all. My dad is a truck driver. My mother worked for Social Security for many years. i My grand father was a rabbi. So not really, in that vein. Yeah. Not a

 

Chad Main  04:22

lot of lawyers, their lawyers. You go to Georgetown, did you go to big law, right. You go to Sibley. Yeah.

 

Dan Rabinowitz  04:27

I went to Sibley did big law for five years or so to trial and appellate work, some product stuffs and white collar and then suddenly has actually one of the premier Supreme Court practices. So we’ve got some do some work around that way. I like to refer to it as I was at the bottom of the list, which meant actually wrote the brief. Went to DOJ also gets to do trial and appellate work, mainly government contracts, but also we actually had a supreme court brief as well as I was involved in

 

Chad Main  04:55

what type of stuff we do the DOJ what were the issues, so

 

Dan Rabinowitz  04:58

we’re mainly focus sent government contracts. So that was a lot of stuff having to do with the FAR, which controls government contracts. We also did some fraud because we have the opportunity to file fraud counterclaims. So we had a couple of trials that fraud was involved. And then also, we handled certain cases in front of the Federal Circuit mainly having to do with federal employees. Benefits. We also handle those, the Federal Circuit level, but that was probably the bulk of what we did. We I also did, I also worked for a period of time on a series of litigation having to do with spent nuclear fuel. So essentially, the trash that’s produced in the production of nuclear energy. So the government of the early 80s signed contracts with all the nuclear power producing companies that they agreed to take all their trash now before they signed those they did not identify where they were, they didn’t really identify where they’re going to put all those nonetheless, all the utilities paid them. And then the government when they went to go ahead and collect it, needless to say, none of the states either and disabilities really interested in having nuclear waste in their backyard. So that’s where like Yucca Mountain came in where they attempted to put it. And until today, they still have not found the location to go ahead and put it in or to deal with the spent nuclear fuel. And as a consequence, the utility Sue more or less every five years or so to get damages, because they’ve had to house the nuclear waste on their own sites. And that has a cost associated with it. So

 

Chad Main  06:25

you’re in house, ultimately, Booz Allen, I assume it’s the government connection there. Yeah. That background. That’s how you get in there.

 

Dan Rabinowitz  06:31

Yeah, government background, I ended up doing internal investigations. The first major investigation I did was the the Edward Snowden investigation, he was a Booz Allen contractor. Oh, yes, fairly relevant. In light of the fact of, again, another major leak of government confidential or super secret information, mainly focused on, as I said, internal investigations. But also there was exposed a bit to systems and corporate systems, and also the ability aside from what my previous experience had been, in terms of systems mainly, like eDiscovery, and the like. So I spent a couple of years there, and then went to a small data analytics company, where I then had the opportunity to be exposed to a number of data analytics platforms that they operated mainly in the intelligence community. Some of you may have heard of them, I think the most well known is Palantir, other platforms as well that they service and sort of there was the kernel, that there must be a better way, if there’s a better way to do you know, intelligence, if there’s a better way to look at those sorts of problems and issues, there has to be a better way to operate in the law. And more particularly, you know, in the law, we spend all this time on factory search, brief writing and crafting arguments, you know, moving that working through all the nuances of it. But as we know, today, let’s just take a pretty straightforward example that’s in the news, you can have the exact same facts, the exact same law and a judge in Washington state rules, one, and a judge in Texas rules the other way. So how do you account for that variation? So you’ve advised your client look and hear the facts we’ve spent, you know, months if not years, going through documents, interviewing people internally, looking at the law, that’s, that’s relevant here, all their legal issues? But then how do you account for the person that ultimately is going to make the decision whether or not your argument is compelling? Whether or not your argument carries the day? And again, it isn’t so much has to do with the particular argument. You know, everyone makes great arguments, everyone has done, you know, phenomenal research, everyone has written the best brief that they possibly can. But how do you assess what that actually will mean, in the real world? That is the judge? How is she going to be affected by who the lawyers are, who the parties are, and so on? How do you account for that? And therefore, how do you advise your clients? And how do you assess strategy? And like,

 

Chad Main  08:55

you’re at the data analytics company? Yes. And you’re thinking about this, you’re thinking about this, what was your role at the data analytics company?

 

Dan Rabinowitz  09:01

I was the General Counsel. But I guess this was sort of where the colonel was where I was brought into the general counsel, but also to try to understand how they might leverage the technologies that they had access to, in the intelligence side mainly, and how they might leverage those and implement those in the legal space. So the idea, if you will, that you could take commercially available technology, or commercial approaches to technology and apply those in the legal sphere.

 

Chad Main  09:26

But how did they wonder if they’re an intelligence company? How did they want to employ that in the legal sphere?

 

Dan Rabinowitz  09:31

So they actually didn’t know they just had an idea. They had worked on a particular investigation, which I can’t get into right. And they had used and leverage some of the technology that they had access to for a particular investigation. So they had the idea that hey, you know, we have all these platforms that we operate on, is there any way that we can expand our business lines so not only include the federal government and you know, others banks and the like? Can we also embrace the legal community? Is there anything that we can bring to the table for them?

 

Chad Main  10:02

So then fast forward, your wheels are turning this is where the idea for predictor comes into play, like, what is this? At what time? Are you starting to think? Well, maybe this is a company I can create. Yeah.

 

Dan Rabinowitz  10:12

So at that point in time, I understand and appreciate that big data has a number of use cases. You know, at that point in time, you know, I’m working there, I eventually transitioned to a healthcare company where I’m able to leverage both my analytic background, as well as my legal background, because we’re creating a anti fraud technology that has both a technology component to it, but also there was a government interface component. So with DOJ, we were part of Anthem healthcare. We were an entity that was created, it was called WellPoint, military care to bid on the second largest government contract and you listen to the largest government contract because of the United States and all its awesomeness. is health care. Yeah, it isn’t, you know, guns or you know, artillery, it’s a healthcare. So we were tasked with creating a bid and proposal for that. And there were two or three other competitors. And ultimately, we did not win that contract. So eventually, the unit was spun down. And I was offered, you know, it was closed down. And I had to exit that What year was this? I’d have to check. But I want to say 2016, maybe

 

Chad Main  11:18

what’s the official creative predictor?

 

Dan Rabinowitz  11:21

Maybe it was 2018. Because I think that our official creation date was 2019 of predictor, the

 

Chad Main  11:26

US shutdown, you get some time off, you’re thinking about it. It’s during that timeframe, you say, Hey, I’m going to pursue this. Yeah. Taking data and applying it to legal analysis, specifically, of judges and courts and stuff.

 

Dan Rabinowitz  11:39

At that point in time, I was trying to consider like, do I go back to the law? Do I go back to a firm Do I go back to in house or the like, and I started thinking about this particular idea about analyzing judges, I live in DC, and you know, every other person that you talk to is a lawyer. And I was like, okay, maybe I could do something different here. Maybe I can look up, take my experience, and go in a different direction. And my partner who has a deep experience in the financial industry, he’s one of the earliest ETFs. And he had been very successful in that. And now, he sort of was looking for other opportunities. And I had reached out to him, frankly, just as a sounding board for whether or not this made sense, how I could potentially go ahead and raise capital in order to create the company and then move forward. And after I met with him a couple of times, he actually said to me, Hey, put together a proposal, I think this is something that I might be interested in joining. And that’s ultimately how we linked up. So he was bringing the capital, he was bringing some of the, you know, sort of cutting edge experience working in industries that were sort of emerging. And from there, you know, we had sufficient capital to go ahead

 

Chad Main  12:53

and did you raise money? Or was it your business partner had enough sufficient capital? It’s just my business partner and myself, quote, unquote bootstrapped coding correct.

 

Dan Rabinowitz  13:01

We are a bootstrap, we have not had to raise any additional capital throughout this process. Although it has been, as you would imagine, you know, incredibly time intensive as well, as from from a financial perspective, we’ve been able to, to just do this, between the two of us,

 

Chad Main  13:17

and what was it about entrepreneurship that tilted the scales that way versus going back to the

 

Dan Rabinowitz  13:24

law? You know, at the end of the day, I met a lot of great lawyers, I thought that I had a unique proposition for the industry. So rather than just being another, I would like to think, although I can’t promise it, another great lawyer out there, let’s bring something new that hasn’t been done before. Let’s come into the space of analyzing judges. And let’s come into the space, not simply through like backwards looking. Statistics, just gather data. Let’s actually and this is one of the things that really was impressed upon me when I was updating analytics firm. And they were working for the intelligence community that like you could use data, not simply backwards looking, but forward looking. So in that instance, as you would imagine a lot of their work or a lot of what the federal government does, without speaking to, you know, the particulars. They were trying to determine what people will do, who is going to be an issue? Where are the issues going to arise? Who are the actors and the entities that we need to worry about? What are their linkages? What can we glean from what they’ve done in the past, to where they’re going in the future. And that’s really where the value add is the value add is not simply, you know, surfacing statistics. It’s understanding the value of those statistics, and then using big data, machine learning and AI in order to take it to the next level and understand well, what can we predict future behavior with using all that information?

 

Chad Main  14:47

When we come back, Dan gets into the nitty gritty about predictor tells us about how it works and what’s on the roadmap. I’m Chad, Maine and you’re listening to technically legal. Let’s get back to my conversation with predictive Under Dan Rabinowitz, you run into somebody on the street or the coffee shop and ask what you do. I’m a lawyer to entrepreneur, I started a company called Predictive What do you tell them?

 

Dan Rabinowitz  15:07

We do traditional analytics, that enables us using both historical data, as well as biographical and demographic data, in order to predict how a judge will rule. On your case, the specifics of your case, what will influence them, the lawyers, the parties, the fact that that’s an individual suing a large company, represented by a big, firm, small, firm, plaintiffs for publicly traded company, how that will affect the decision maker, just like all of us are affected by any number of factors as, as we people, and we make split second decisions about those, how does that transfer and relate to judges. And once you understand that, it becomes this exercise in predicting what they will do next.

 

Chad Main  15:51

And specifically, on your website, you say, we use data science to predict whether cases survive a motion dismiss by identify factors that influence such a decision, including the judges net worth, political affiliation, educational work experience, and other biographical data? How and where do you get that information?

 

Dan Rabinowitz  16:08

So it’s a variety of different sources, I can say with absolute faith that we do not hire private investigators ahead and surveil federal judges or state court judges. Instead, much as you would imagine, with regard to you know, us as civilians, if you will, there’s a lot of information that’s publicly available. And we call from those sources. And we create our own proprietary database that includes information, biographical historical career, about federal judges that then we incorporate into our model. So every federal judge right now, and we can talk about the Catalytics acquisition and a bit, every federal judge has a unique biographical profile that’s created from those multiple data sources. Okay,

 

Chad Main  16:53

so let’s back up a little bit. Let’s talk about the nuts and bolts, it’s 2019. Get your business partner said get some money. What’s the first thing you do? How do you develop this? How do you create the model? How do you create the technology?

 

Dan Rabinowitz  17:03

Yeah, so I had the idea that if we could marry up if you will pass performance, with the human being, we could go ahead and predict how that human being rule based upon the case specifics.

 

Chad Main  17:17

So would you say past behavior, you you talk about, okay, we’ve got a motion dismiss based on forum, non conveniens, or whatever it is, and they’ve ruled and X percent of the time, what are you looking at?

 

Dan Rabinowitz  17:27

So let’s actually take it to a higher level, we don’t want to necessarily look at the individual, because when you’re operating under the concepts of big data, or data analytics, what you’re really trying to do is to take the individual and divide them into multiple pieces. So for example, I am Dan Rabinowitz. I live outside of DC, I’m a lawyer, I live on a particular street, I have a particular net worth my neighbors or so and so, and all those factors. I have a postgraduate degree, you know, I’m married, and all these things are now Dan Rabinowitz. So essentially, Dan Rabinowitz is a huge spreadsheet of a variety of characteristics. So those characteristics then enable us to see the person and put them in buckets in other buckets. So what we want to understand is not necessarily who the lawyer is, specifically, or who the plaintiff is, specifically, we want to understand who they are from a classification perspective. So is it a corporation? Is it a publicly held corporation? Is it on the s&p? Is it in the financial industry? So now what we can do is we can take something that is simply text, right, JPMorgan Chase, that’s just text. But we want to then take that and compare JP Morgan, because JP Morgan is a lot like Wells Fargo. It’s a lot like other financial operators. So the benefit of that is, is now we’re not looking at just cases with JP Morgan Chase. We’re looking now at the universe of all cases that involve financial institutions that are publicly traded. So that opens the aperture to now include a whole lot of cases that have been litigated across the country, maybe in front of that particular judge. So that’s really what the way that we want to approach this is sort of back out of that. And this is one of the the limitations of a lot of the current data analytics, where if you want to get down super granular, you want to say, Well, how is Dan? Or how is his law firm at winning? The problem is, and I experienced this, and I’m pretty sure this is not unique. But when we had a case, there would be the firm wide email that went out. Right email would say, Okay, we’re in front of judge Smith, has been in front of judge Smith, our client really wants to know and inevitably you’d have someone say, she was really tough on us. Now, what they don’t say is, is that they were in a USA in a criminal case. Right? Right. And that has no relationship to the contract dispute between two major corporations. So the way to get to that is by understanding those classifications and categories. And that enables you to sort of pull back and capture a lot more data and then to understand patterns and trends. So what we do when we’re looking at historic data, essentially when it comes to motion to dismiss, so judges write opinions and fewer than 2% of all their motions to dismiss. So if you’re just deciding, hey, I want to see what this judge is going to do. Let me look at their opinions. You’re ignoring 90% of what they actually rule. So then you say, Okay, fine. So let me look at the decision. So those decisions are those one sentence orders, right, I hereby grant I hereby deny. And there are a number of platforms that provide that backwards looking historic data, and they say, you know, 67% of the time, judge Smith dismisses cases and then you can drill down one level and you can say, well, 72% of securities cases judge sniffed grants, those motions to dismiss. But if those 72% If all those securities cases that she heard, were individual suing their brokers, and you’re representing to go back to the case we were talking about before Wells Fargo, and they’re suing Chase, most of those have no relevance. Right. So how do you then account for those cases? So what we do is we look at those decisions that are just the grandson denials. And we enrich that data by understanding who the parties are, who the attorneys are, who represents who was an individual suing a large corporation? Was it a regional firm that was representing the individual? Was it a plaintiff side firm, a normal one? Was it an amyloid 100 on the other side, so now we’ve taken those one sentence orders. And we now have much, much more information about those one sentence orders. It was when an individual was suing a large corporation, represented by a solo practitioner, on the other side was an amla, 100 firm. And then the beauty of our system is it then starts looking for patterns within that information. So those yeas and nays those grandson denials are now turned into a much richer data set. And that’s how we’re able to then to your question about, like, how do we get like a, you know, the particular type of motion or the particular type of law, frankly, and I didn’t from a lawyer perspective, I find this somewhat, I don’t wanna say frustrating, but certainly disappointing that we don’t care about what the precedent is, I don’t care. If Twombly came out and 12 B six motions are treated differently than than they used to be before, or it’s forum nine TV, whatever the reason, or the rationale, or the argument that the lawyer articulated as to why the motion to dismiss should be granted. Instead, it’s these non obvious patterns that are much more telling if you want to, to arrive at a prediction. That’s

 

Chad Main  22:41

interesting. I’m oversimplifying here, but I take it you take a particular judge particular motion and you got attributes. And you’re just like filling in the blanks? Like, is it a big company? Yes. Is it you know, a foreign defendant, no attributes like that? And then that’s what the analysis is conducted on. So that’s

 

Dan Rabinowitz  23:00

the first half. Yes, that’s exactly what the first step the second half is. Now, we want to understand the human being. So now let’s look at the demographic and characteristics of that particular judge. Where’d she go to undergrad? Where’d she go to law school? Where did she practice? Was she a state court judge, before she was elevated to the federal bench net worse, you know, geography, all these other pieces that go into making us human? That’s the other piece that we need to combine with that historic data?

 

Chad Main  23:29

Because a lot of this information is public record, because I got to disclose a lot of it. Right? Correct, especially the financial stuff, except for the Supreme Court justices. But that may change.

 

Dan Rabinowitz  23:40

Perhaps, perhaps, but it’s really the combination of the biography and what they’ve done in the past understanding that you have to look beyond the simple yeas and nays, that’s what enables us to get to that highly accurate prediction, the way I like to think of it, in some regard, the way that Google is able to predict whether or not we’re gonna be interested in a trip, or we’re gonna go somewhere or buy something, it’s probably not because they’re listening to us, although that’s always a possibility. But you know, how, like, you’re thinking about something and then the app pops up. The reason they’re able to do that is because they have seen our past buying habits, right. They’ve looked at our past buying habits, and then they know information about us personal, right personal data. So it’s that combination of those two pieces of what we’ve done in the past as it relates to buying a TV or whatever it might be, or a judge deciding a particular motion, as well as understanding the the biographical features and how they live their life, what street they live on, you know who their neighbors are, etc. That’s how Google is able to make many billions of dollars business out of their advertising. And it’s sort of the same concept that we’re using to predict judges, preferences and rulings in the future. Predictive

 

Chad Main  24:51

starts is focused on federal court. Fast forward to this year, he acquired Catalytics analytics. Rick Merrill started in a few years ago had a good run, then kind of shut the doors and then like a phoenix, you guys come together, and you combined forces. So tell me about that. What interested you? How’d you guys get together? How did it all start?

 

Dan Rabinowitz  25:13

So Rick obviously was a pioneer in the judicial analytic space. I mean, he was one of the early ones to understand the value of looking at what a judge has done in the past. And in particular, you know, he attacked the challenge of state court data, which is highly complex and very, very difficult. It’s very difficult not to correct every state and then every jurisdiction within a state has its own, you know, court record, system of record for the courts, how to get that data, where that data resides, and how to process it. And all the challenges that come with it, Rick, created great technology, as well as a great interface in order to surface that and allow attorneys access to that information. Rick, being someone in the space, when I launched, Rick actually reached out to me just to say, hey, great idea. This is a space that’s really ripe for innovation. And then he I think it was a couple days afterwards, he, you know, had to shut down Gavin Linux, and in the early conversations with him, you know, I have been, you know, while we initially focused on federal data that were federal courts and federal judges, while that’s a huge amount of information, it’s over 700 judges. At the same time, obviously, for us, we always had an interest in moving into state courts. Nonetheless, in terms of starting a company, you first tried to crack the first level. And in my discussions with Rick, you know, I saw this as a terrific opportunity for us to obtain the state court data that he had worked so long on the scraping, processing, and then pulling it all together. And of course, he was interested in while Catalytics didn’t work out in the end, but the idea that all that work, that he done all that heavy lifting, could then be used for an entirely different application, you know, what Catalytics did was essentially, and this is a way I think that Rick articulated to me, which I find very compelling, which is Catalytics was essentially an almanac, right? It told you what the weather patterns had done. And then you know, what you might anticipate from them, what we do is forecast, we tell you two weeks out whether or not it’s going to rain. Now, the way we’re able to do that is using the data, such as the data that Rick and Gavin Linux I collected. But this is, you know, two or three steps beyond where Gavin lyrics was. So the idea that we could leverage all that work, and then not only just simply reproduce it, that is just turned back on Catalytics. But take that data and really using it in a more meaningful way. Because, of course, to go back to the earlier discussion about how, you know, if you’re looking at, you know, 72% of cases, our judge grants those when it comes to securities cases, but without understanding the underlying issues as it relates to those again, was it Wells Fargo? Or was it individuals suing their brokers that really can’t help you when you’re looking to predict, right, if you’re interested, if you will, an academic exercise to like what the judge has done in the past? Great. But if you want to answer the question, which every lawyer wants to know, which is what is the judge going to do with their case, the way to get there is to take data like Rick hat, and then work through it the way that we have create the predictive models and so on. And then you can arrive at the answer to that question, what the judge will do with your case with the motion that you filed?

 

Chad Main  28:29

Is it a matter of normalizing the data because as Rick was doing for state court and gavel addicts was clicking different data points? Oh, there was overlapping, I’m sure. But you’re collecting some different stuff. I mean, like, you know, where judgment to school and how much the judge is worth? Has there been some growing pains trying to normalize that information. So

 

Dan Rabinowitz  28:46

I would actually say less growing pains and more opportunity for us, because we had a slightly different focus than Rick and Catalytics. And Rick at his particular focus, one of the things that we’ve been really excited to learn is during this integration process, we’ve created a much richer dataset than either of us had beforehand, right number of things that we’re contemplating doing in the future, but that has really enabled us to look in a very different direction. Because of the two datasets being natively different or having different foci. You notice that they’re looking to do something very different. The combination really produces, if you will, a beautiful stepchild.

 

Chad Main  29:22

Well, though it’s a child, right, it’s not a stepchild, you guys are married, it’s a child, and you guys aren’t remarried to new kid. So what’s on the roadmap, which the initial plan what a short term goal is long term goals? Sure. So

 

Dan Rabinowitz  29:34

we have both near term and longer term. So in the near term, you know, the major driving force in purchasing and acquiring Catalytics and its data is to take our same philosophy that we’ve applied to federal judges and apply that to state courts. So to provide predictions now, our limitation right now is that we’re down this way limitation but our our first, you know, sort of air area that we’ve decided to tackle is motions to dismiss. And I’ll come back to that in a moment. But the idea number one on our product timeline is to move beyond motion citizeness, whether that be summary judgment or expert motions, but to start looking at right, I

 

Chad Main  30:14

just want to say that because you can kind of get some pretty important information from that motion dismiss. I mean, a judge is less likely to dismiss the case, you know, he’s gonna give a few times, right. But there’s going to be some good indications there. Like how that Judge leans come summary judgment time and how to apply that info. Exactly.

 

Dan Rabinowitz  30:28

So there are other opportunities for us to push this out well beyond the motion to dismiss. So that’s one sort of on a timeline. The second piece, and this goes back to the more robust data set that that we’ve been able to create by the integration of the gavel attics with ours. And that is to the extent that we can’t necessarily immediately get to a prediction, we are able to offer more tailored statistics meaning, again, to go back to this, you know, I don’t mean to harp on it, but about the Wells Fargo versus JPMorgan Chase. So let’s say we haven’t reached the point where we can provide a prediction about summary judgment. But at least if you’re looking at the statistics for summary judgment, you want to know what the judge does with similar financial institutions in summary judgment, so that information we are going to be able to provide also in the near term, so in other words, these more case specific case, tailored statistics is also an area that we are going to go ahead and roll out now, I should back up because I think one of the areas that I didn’t get to touch on urine, it’s really important point or really important capability that we have is we are not only able to do this after a judge is assigned, so of course being judged focused, we look at the job, we look at the particulars of the case, and then we can determine how the judge will go ahead and rule on the motion. So we are QL agnostic about the facts of the law. Now, before a case is filed, there are three unknowns or complaint, you don’t know the facts, you don’t know the law. And you also don’t know who the judges. Right. So if you’re filing a case, you’re sort of left to that anecdotal, you know, hey, I practice in front of this jurisdiction, or that jurisdiction is more favorable to plaintiffs are more favorable to defendants, or whatever it might be, Oh, of course, all that. In most instances, the data actually doesn’t bearing that out. But like, we like to think that we know how all this awkward, but the reality is, we don’t. So we’re able to do is, of course, the facts in the law are less important to us. Well, we want to know who the parties are and who the attorneys are. And then of course, in terms of the judges, so we do is we look at our jurisdiction, and we say, Okay, we’ll analyze every judge within the jurisdiction, and then we’ll give you an overall jurisdiction score, as well as the each individual judge because ultimately, it will be assigned to a judge. So then what it enables both cleaness, to do as well as others, and I’ll get to a very interesting use case that we just had with with a client, it enables them to say, Hey, should I file this jurisdiction or that jurisdiction because this jurisdiction, I have a much better chance of surviving a motion to dismiss. And of course, I’m plaintiff side, the major gate are the major hurdle to get through in order to talk about settlement and where to get to anything else to survive a motion to dismiss. So we can do that analysis even before cases filed. Now, that’s one application in in terms of, you know, on the plaintiff side, now, on the defense side, we actually had this client approached us and they had a case in the Southern District of Florida, and they were in front of a judge already. And they had filed a motion to dismiss. And we looked at the judge, and we said, hey, the judge is not going to grant your motion to dismiss this to kind of go to discovery, it’s gonna cost you millions of dollars and discovery fees. And then, of course, in terms of sell. But they said we had an opportunity to transfer venue, we can transfer to the Central District of California, our chances better. And we looked at it, and we said, in fact, your branches are much better, California, so you could go ahead, and we could run that analysis, even though the case hasn’t been filed in the central district. But we can go ahead and also run the same predictive analytics, and understand each judge in the central district as well as the overall likelihood in that particular jurors. In order to arrive at a prediction that you can safely if you will rely upon to make that decision. Should we file transfer motion or not? How

 

Chad Main  34:09

do organizations subscribe to the product? Is it subscription based? Is it project based? Is it ad hoc? How does it work? So

 

Dan Rabinowitz  34:17

the way it works is, from our perspective, it’s a one and done for each case, right? So you run our analysis for a case, and the only information that you need to provide our system and this is a critical point. And I picked up from the time that I was practicing law, as a lawyer, you don’t have enough time to do whatever you already have to do. So overlaying an additional piece of technology and an area that there’s frankly no way for you to effectively analyze that is the judge. You’re not a data scientist, you’re not a statistician. So we didn’t want to give you as more numbers, more stuff for you to go ahead and parse, analyze and try to determine its relevance. So the only information that you need to feed to the system is the case then we don’t require you to upload your briefs. We don’t require Are you to upload any facts, it’s simply the case number and then we provide the prediction. And once you input that case number, we go ahead, provide the predicted prediction, and then you understand how the judge is going to reward your case. So it is not a user base. It’s not a time based system. It’s simply it’s charged per case. And our pricing model is simply an annual subscription. With we have three different tiers based on the number of cases that you ultimately subscribe to. Dan,

 

Chad Main  35:27

I appreciate your time. If people want to learn more about predictive they want to subscribe. They want to reach out to you where you want to send them.

 

Dan Rabinowitz  35:34

Just the website, pre hyphen dicta.com.

 

Chad Main  35:49

Okay, that’s a wrap of today’s episode, as always, really appreciate you listening. If you want to subscribe, you can find us on most major podcast platforms like Apple, Spotify, Google, Stitcher, etc. Also, if you like us enough, I hope you leave us a favorable review. Thanks again for listening and until next time, this has been technically legal

 

LinkedIn
Facebook
Twitter
Reddit
X
WhatsApp
Email
Print

Leave a Reply

Your email address will not be published. Required fields are marked *