← Emerging Litigation Podcast
Loading summary
Transcript39 lines
- [00:04]
A
Welcome to the Emerging Litigation Podcast. I'm your host, Tom Hagee. I devote this podcast to topics important to legal professionals and anyone interested in law like me. I'm also founder of Critical Legal Content, which allows boutique firms to put a real life legal publisher behind their blogs, white papers and other media. Click the link in our show notes to learn more. And now, here's today's episode. Artificial intelligence. You've heard of it. Surveys show that adoption among law firms has jumped quite a bit, from about 11% in 2023 to 30% last year. You know, that's almost triple now. According to some surveys, more than half of attorneys use AI tools, whether they tell you or not, for tasks like document review, legal research and summarization. You know, the big legal research companies are using it with some success. Large firms are leading the way, but smaller practices are catching up. I think they have the most to gain, frankly, because it gives them a lot of juice that they may not otherwise have to compete with the big firms. Some firms report efficiency gains are considerable, saving lawyers weeks of work each year. But I don't think they're taking that time off. Bloomberg Law reports. Well, enthusiasm is growing. Daily use is still limited, and they say only 21% of attorneys use generative AI about once a day. Many firms remain cautious because, you know, they hear about accuracy issues, they have ethical concerns, and then they hear about some screw ups, which are, you know, those are happening. Still, practice groups like Ediscovery are seeing clear benefits because you can use AI to sift through massive data sets. Massive, massive data sets faster than ever. I said massive twice because I don't know I'm seeing it. Well, I'll talk about that in a minute, but I'm seeing it just to be a great tool for drafting anyway. Nexus adds that about half of professionals save one to two hours per day with generative AI see these surveys. About a third save three to four hours a day. So that would be a strong return on investment for firms that implement the tools effectively and appropriately and mindfully. I would say that's how you have to approach them. And according to Thomson Reuters Future of Professionals report, professionals expect to free up 240 hours a year, unlocking an average annual value of $19,000 per attorney. Well, that's what they say in the U.S. that translates into, I like how they do this. We can really blow this up in the U.S. that translates to a $32 billion combined annual impact for the legal and tax sectors. Who will get some of that? Firms with a Clear AI strategy are twice as likely to see revenue growth. It says three and a half times more likely to experience critical benefits than those without a plan. I don't know, those are a lot of numbers. But look, it's gaining traction. It's here. Learn to use it responsibly. You know what I mean? That's, you know, drive responsibly, litigate responsibly, use AI responsibly. Responsible firms are seeing tremendous results. They're cutting contract review, for example, from hours to minutes. They're automating compliance checks and summarizing depositions in minutes instead of hours. I will use it to give a summary of a hundred page court opinion, something that I used to spend hours a day on, you know, as a young journalist in law. And man, man, that's all I'm going to say. Especially if you're pointing it at something, you know, I use Adobe or I use Copilot. The benefit is, is that you can ask it to show you exactly where it got its information so you know, it's not just making stuff up. You're not pointing it at the wild west of the interwebs. In high volume litigation, these tools have reduced associate time from 16 hours down to 3 to 4 minutes, 16 hours down to 3 to 4 Minutes, delivering productivity gains of more than 100 times. That's a lot of math. It's just a huge time saver for some of these tasks, which frees you up to do, you know, some real, real thinking. You know, there are examples of people who have completely blown it, of course. One law firm was sanctioned after filing a motion with eight fake cases that were generated by an AI tool. Let that be a lesson to you. In California, an attorney was fined $10,000 for briefs full of fabricated citations. In one instance that came out in the news recently, there were even unverified AI citations that made their way into a published court opinion before being discovered and removed. So it's powerful. So what I'm seeing is AI is most powerful when you point it at structured trusted data. After that, it's all about the prompts, how you instruct and direct the tool. After that, you gotta, it's gotta be reviewed by lawyers as if it was originally drafted by a junior professional. Hey, those are some quick tips, and those are for free. Speaking of using it responsibly and strategically, today's guest is Adam Massaro, an intellectual property and commercial disputes litigator. He's a partner at Reed Smith based in Denver. He joins us to discuss strategies for Using AI in trial prep and courtroom presentations. His team recently secured a defense verdict in a neurological research case and a permanent injunction in a trade secrets dispute. We're going to talk about how AI tools supported expert witness cross examination, streamlined evidence and shaped trial visuals. Adam also shares practical tips for preparing witnesses, structuring pretrial briefs and deciding between jury and non jury trials. He earned his J.D. from the University of Denver Sturm College of Law and his BA from Pennsylvania State University. Now here's my interview with Adam Massaro of Reed Smith. I hope you enjoy it. Now, it's just the sports. I don't follow. He should be joining at any moment. That's encouraging. Whenever I have a guest, it's encouraged. Adam Massaro, thank you very much for joining me today.
- [06:05]
B
Pleasure to be here.
- [06:07]
A
Let's dive in. The first question now, you've had a string of trial wins recently, including a defense verdict and a high stakes neurological research case. Can you walk us through how AI factored into your trial prep and strategy in that case? Maybe you can tell us a little bit about the case too.
- [06:23]
B
Yeah, absolutely. This was an interesting case we tried earlier this year and had a lot of complexities dealing with complex damages models and a few other issues that we had to overcome on the defense side. And it was an interesting case between two feuding neurologists over the breakup of a partnership. So a lot of emotions, a lot of factors that we had to deal with amongst the parties. But we knew the parties well because we had actually won another case for the same client a prior year. But this dispute surrounded over a research company that they were breaking up. And, and we knew going into trial number two, a couple of variables. One, both sides obviously had refined their approach since case number one. And then separately we really thought that, you know, we had to focus on experts because it was a very expert driven case. And the damages model was a future lost profits model that we thought was just right for ATTCK here in this case. And so how we used AI was a couple ways in my experience. Number one, it's important to have a lot of critical data and information that the LLM can learn from. So what we did in this particular case was we had the expert reports which we were able to run through our LLM model at the firm and start to get a sense of what the experts opinions were going to be. Both sides, because we had our expert as well. So it was easier to train because we had a healthy amount of information. And then we started working on ways to develop cross Examination of the other side using the LL model that the we were through our proprietary software.
- [07:52]
A
Okay. And from what I understand about training them, I do very basic level use of AI. But I have found that if you give us something structured to work from and not necessarily tell it to go out to the World Wide web and look at our structured data that I'm giving it, that's, that's a real effective use of AI.
- [08:10]
B
You're right on point. And the key is the prompt. It's something where we've evolved it over time that the prompt and how you think about it and you don't. I think your default is, oh, let's treat it like a Google search or a Boolean search we do in Westlaw or all the things you've been trained on for keyword searching. And that's completely wrong. You truly have to think of it like a advisor in the sense that it has a brain in that respect. So you got to think of it as teach it like a child, teach it in the sense that it's going to be responsive to you, not just executional. So the prompt is key, laying out a clear directive. And a lot of times too, you know, we will tell the AI system we're using, assume you are that expert, create that Persona, create that avatar, and then I'm going to examine you off of that piece. But you've got to take it to that next level because if you're just doing remedial searches or remedial asks, I think you get back remedial outputs. But the prompt really is the driver.
- [09:06]
A
Yeah, that's interesting to tell it to assume a Persona. Now you've developed these AI driven litigation forensic tools. So what specific capabilities do these things offer and how do they assist in cross examining opposing damages? Experts, for example.
- [09:22]
B
Yeah, so on that point, in particular experts. In my view, one of the best ways to cross an expert is putting them into a very challenging hypothetical. Because they have to answer hypotheticals, they can't punt the way a lay witness is. So developing a well crafted set of hypotheticals is very, very powerful against an expert and it allows you to take them down a path before they really see where you're going. So in this particular case I'm speaking of preparing for trial, I knew I wanted some hypotheticals, ideally ones that I hadn't even thought of to catch this expert off guard. And so we used our AI system to help us develop prompts and hypotheticals to hit the other side with. And ultimately it proved very valuable because I think some of the hypotheticals we used at trial were ones that I hadn't thought of. And then once AI gave me feedback, it gave me a path that the exterior wasn't expecting. And in that respect, I think it caught them off guard in different ways. And so, you know, to me the hardest part is a trial. You've been involved in the case a year and a half, two years, you're going to be narrow minded. It's just the nature of being involved in a case that long. So having a true objective party in this situation, AI give you ideas, pushback sort of helps you refine that skill or also figure out, you know, what arguments you need to drop on the cutting room floor just because they're not good. But you know, we all love our arguments and it's hard to get rid of them unless you have some objective feedback.
- [10:45]
A
Yeah, I do find it helpful that way. Now you, you've taken these cases to bench trials and to jury trials. So what, how do you decide which, which way to go? And does you use AI in helping you make that decision?
- [10:59]
B
Yeah, there's a few variables. Now I will say what's interesting is we've spent so many years trying to crack the nut on jury trials using juries, experts, consultants, which are fantastic. Also wildly expensive and it can be cost prohibitive on even some of the largest cases because the cost is just out of this world. And especially over the last several years. Inflation affects all things, including even jury pool and various exercises. What's interesting is we had all this great data on jury selection exercises, fine, but not much on judges, which is somewhat interesting because a fair amount of cases with jury waivers or you make a deliberate choice. So using AI for judges has been really interesting. And an example I can think of recently actually was we had a receivership hearing that was a contested two day evidentiary hearing dealing with complex issues. It was effectively a mini trial. And we in that particular case used AI to come up with some predictive outcomes about how this judge was going to rule. Because again, you have a lot of paper, you've got all the pleadings, you got all the filings, affidavits, things like that. And so using AI to predict judges, I think is actually arguably probably easier to do than a jury pool because we have tons of information on how judges rule. We have thousands upon thousands of cases. So I think the predictive outcomes for judge trials is probably going to be stronger and faster before we get to the jury piece on large language models.
- [12:25]
A
Yeah, it'd be interesting to see judges Respond to your prediction of how they're going to, how they're going to rule. And I know there are services dedicated to that. Right. That have all the opinions the judges has ever written. Right. I don't know if you're accessing a database like that or you're going off on your own experience.
- [12:41]
B
A couple variables. One, back to treating AI like an advisor and not just a executional piece. You need to tell AI what is the veneer of your judge in this case? Background, experience, things like that. Obviously, especially at the federal level, very impactful because we've got great data sets on judges just historically, from the amount of information we have. And obviously we have, you know, the public records are far easier to obtain at the federal level. So I think part of it is setting up that veneer for that judge and then finding analogous judges that have ruled more. So, prime example, everyone's like, oh, I just got a brand new judge in this case. He was just appointed, or she was just appointed to the bench. We know nothing about them. Well, now that's not really true. We do know a lot about them because we figure out individuals that have that same background, public defender, da, prosecutor, whatever the background is, find that judge in the greater population and start running multiple scenarios off of what those judges have done. And that will actually know that even with a new judge, we probably have some level of predictive analysis, you know, far more than sort of the hypotheticals we've thrown out, you know, gut reactions over the years.
- [13:51]
A
Yeah, yeah, I can't, I can't imagine. I mean, going back to the old days in the 50s and 60s, you had to just ask people, you know, you had to network and do your research that way. So now back to the neurological research case. So narrowing trial issues through summary judgment. So how did AI help you identify weakness aspects of the case and how did that affect your briefing strategy?
- [14:14]
B
So a couple things. We've definitely had a lot of success with refining arguments. So first thing I like to do with AI, when I get the draft together, whether it's an opening brief or opposition brief that has been put together by my team, if I run it through AI and AI can summarize my arguments, number one, I want to see if AI is actually capturing what I'm saying, because if it's not capturing what I'm saying, that tells me we have an internal communication problem. So first sort of threshold test is how can you use AI to make sure that what you're communicating is actually effective? And so a third party objective individual in this case, AI knows what you're saying, so that's threshold test. Next then is I start to test a little bit of what are the strengths and weaknesses, where are the underlying flanks I haven't addressed? And I press AI for strengths, weaknesses, things like that. In my view, simple logic is far stronger. I think there's a bunch of concerns with hallucinations when it comes to case law research, so I don't play in that pool. However, I do believe that simple, thoughtful logic, which AI is trained on for years, and in that case, you know, decades and centuries, because we've been applying logic far along than we apply a case law, I really test it using logic because if logic wins the day, I can do that far more simplistically than I can with some case that may or may not actually say what it says. So a big part is, can I refine the arguments and get better logical outcomes? Because if it's logical, then a judge is going to understand it. If a judge understands my argument and I have a simpler path to victory, I've already increased my chances of success.
- [15:47]
A
All right, now, in the Blue Sky Restoration case, you got both a permanent injunction and a monetary award. So what role did AI and digital forensics play in building your trade secret misappropriation argument there?
- [16:02]
B
You know, I think that's an interesting point because I think you're going to see forensics and AI come together here quite soon. So on the forensic piece, of course, on these trade secret cases, we're getting access to the former employees, computer, their devices, things like that, Internet search history. That's fantastic, that's great. And in that particular case we had a really well laid out strategy because we knew a lot before we ever filed. And as we developed the case, we kept getting stronger, irrefutable, concrete forensic evidence. So that the key in that is always having a good vendor and knowing what you're looking at and knowing where to find the pockets of information, Internet search history, things like that. So that was the key driver there, I'd say on the AI piece for that particular case, the help was getting ready for trial and synthesizing it, simplifying the process itself when it comes to developing sort of the pre trial tasks we all think of and starting to limit the amount of additional extra hours we have to spend because we can do more with AI to get ready for trial. And so I think that was some of the early drivers there, which it's hard to get ready for trial. But if you already have an AI system that knows what you like, when it comes to trial prep, it can be a repository for that first draft or that first set of checklists so that you're not starting from scratch every time you start a new trial or a new trial team.
- [17:20]
A
Okay, you've tried four cases with Joyce Williams in the past year. How does the team familiarity enhance your use of AI tools during trial prep and division of responsibilities? And I didn't note. Who is Joyce Williams? Is she on your team?
- [17:34]
B
Yeah, I'm glad you asked. Ms. Williams is another partner at the Reed Smith Denver office that we founded back in February of this year. And I was very fortunate Ms. Williams came over with me from my prior firm. But as far as you know, the. The familiarity is something that is invaluable. You just having trials are by definition high stress, even ones that are going well and unpredictable. So just having other folks that you've been with through those environments is super helpful. The one thing that I really like doing, especially with Ms. Williams, is oftentimes she does the openings, and she's done multiple openings in our cases together. She opens, I close. But why I like it is twofold. One, I start to see the case developed through our initial presentations, the things we develop, you know, weeks and months before trial, and then she presents it to me. So I get a sense of where the case is heading and whether it thematically lands where I want it to land. And what's really fun about that is it's a combination of her perceptions in mine. And then we obviously make it through some visual bang with the actual presentation we're putting together for the judge or the jury. So that's super helpful from my perspective. And also it really helps front loading some of the work so that I can kind of work on the back end, get ready for key crosses, and then close. So certainly having that relationship and trust and confidence is obviously a huge part. And then similarly on the AI prep piece and how we've done some of those things, I think it's very helpful for witnesses is the bottom line, because you're going to get better and better feedback on what the quality of the witness is, especially through AI, who, like I keep saying, is. Is objective and much more objective than we are as the people that have lived and breathed these cases with our clients. We want our clients to succeed. But the nice thing about AI is it's going to give you an objective evaluation of the credibility and the quality of that witness as well.
- [19:27]
A
Okay. So that's how you use it to improve witness preparation. So I guess especially for less experienced witnesses. And you integrate that into your mock cross exams and role reversal training, is that right?
- [19:38]
B
I do. And what's interesting is I've started to think about ways to come up with a couple pieces here. So with your, your own witnesses, there's a couple thoughts I'd really like to see sort of develop on witness prep. I think getting a more qualitative level of evaluation of your own witnesses because you have control of them. And so I see a bunch of different ways that'll play out both in the past and in the future. Number one is obviously deposition transcripts, things like that we can submit into our AI system and get feedback on the quality that wisdom, witness inconsistencies in their statements, you know, far faster than we've done in the past. And then going forward, I think what'd be really interesting on a lay witness prep is including videos and things like that in the AI systems so that we start getting evaluations of micro tendencies. Right. So for example, your micro expressions, how does this. Does this witness constantly look down when they're searching for the truth? Stuff like that, Are they appearing untruthful or they are just preparing Thoughtful. Right. And that's always a fine line. So I think that type of thing, as we see video catch up to the text part of AI is going to be really interesting. But I think probably this, the thing I'm most interested in so far is really the cross examination prep of experts. So I can tell you a little bit more about that.
- [20:52]
A
Yeah, go ahead, please.
- [20:54]
B
So on the expert piece, like I said, one of my favorite things to do is tell the AI system that I want them to adopt the veneer or the personality of that expert on the other side. And the hardest part about cross is you come up with it in your mind. You then put it into chapters onto how whether you're using onenote or however you're going to present it for purposes of the outline. But then what's most important is, is the cross actually playing out the way you envision it in your mind or not? And we oftentimes kind of end at that next step of we build the cross, but we don't actually try it to make sure it's going to work. And so the first time we're trying the cross is when we're actually live with the witness who by definition doesn't want to help you out. So what I like to do is put my witness in this situation, obviously the AI system as the deponent or the adverse witness, and then tell Them I'm going to cross them. I tell them their answers need to be relatively limited, they can be a little bit difficult. But I need them to answer yes, no to an extent. And then since you can use talk to type with most AI systems, you can start crossing them in there and you can see one, are my questions clear? And then two, how are they responding? You know, maybe they took a path I wasn't expecting. And so I think it gets you reps in a whole different way. And the best part about it is it's, you know, reps that you're getting that it's not costing the client any additional money because we're not putting out another person, things like that. So I think that kind of testing is really interesting and it gets you more pre trial reps in a way that's I think way more cost effective for the client and probably gets you a better outcome too.
- [22:28]
A
Yeah. As you were talking earlier about how AI is going to be more objective and I was thinking it's also going to be less emotional and that it's not going to respond to the person physically. But then you start, you said you are pointing AI at video and I wonder, is that, is that a technology that's emerging? Was like you said, are they looking down? Does that indicate they're lying or that they're being thoughtful? So is that happening?
- [22:53]
B
I think that's the next wave. So.
- [22:54]
A
Yeah, we think so too.
- [22:56]
B
From my perspective, you know, we've, we've had success with using transcripts, a lot of written materials. I think AI right now is built best for that. What I'm seeing. And I think this future soon, like everything with AI, I think the video piece is coming up real soon. And then I think it's really interesting because I haven't had a chance to stress test this yet. But where my hypothesis of where we head is put in video in any form format you can get of the other side or your own witness and start getting critical feedback on all the other things including presentation techniques, you know, far outside just what they're saying.
- [23:29]
A
Okay, I feel like my next question's sort of repetitive of what you, you emphasize the importance of pre trial briefs in bench trials. So how does AI assist in drafting or refining these briefs to better align with the judge's expectations? You mentioned that too. You're using AI tools and platforms to create persuasive visuals. You've talked about that. Is there anything else you want to say about creating visuals with AI?
- [23:51]
B
Yeah, I will say and this is actually something That I just came to light in another case is it's AI's ability to ingest tons of information, especially in large data sets, that we get in Excel's financial outputs that are in all of our corporate cases, immensely helpful. So a prime example is I was dealing with a case here recently where we're dealing with a distressed asset on the other side, and we're trying to get a quick and dirty summary of financial information about that asset. And so having the ability, for instance, to summarize all the financials we get, which normally I'm paying an outside vendor or I'm having to spend hours of a paralegal to take complex Excel data sets and synthesize them. That's all stuff that AI you can dump in now. And I just get the prompts, you know, tell me all The AR that's 60, 90 days old. You know, tell me all the AR that's over a year old bucket in various Excel buckets and columns, what will be the, you know, projected financials for this issue or that issue, that type of thing? So I think it's going to be very, very helpful because, you know, the toughest part about lawyers, especially corporate lawyers that try cases is our specialty is in building cases and creating narratives and themes. We don't have all these other magical specialties. And so sometimes we get hung up on these sort of administrative components to get to the answers fast. But I think AI is going to be phenomenal for stuff like that. And then it's okay, I know what I want. I just need AI to give me a simple narrative of this financial history, of this business, this asset. I'm trying to show positive revenue growth because I'm trying to show no harm. I think kind of being able to talk the type to get that output is going to be super helpful and it's going to cut a lot of the back and forth that normally drives cost. It doesn't really get us anywhere. I think AI is going to replace a lot of those issues.
- [25:42]
A
You're obviously enthusiastic about the potential. Not just the potential potential, but how you're already using AI. And I'm just curious, how did you get started using it? Did you just see, this is just like, this is a great technology that we should be able to use as lawyers, or was something else pushing it? How'd you get into it?
- [25:59]
B
You know, it's funny, I'm like an elder millennial, so I grew up with a computer and like Oregon Trail, but I didn't, you know, I certainly wasn't you know, born with a computer in my hand the way like my children are now and have, you know, and their ability to navigate technology at 5 versus me at 15 was this. Sure.
- [26:16]
A
Yeah.
- [26:17]
B
So I will not say by any means that I had a tech savvy in these components. To me, it was two variables, which was certainly at Reed Smith, we already had an internal platform when I came over in February as part of the Denver launch. And so having that platform and getting trained up on it just opened up a bunch of doors to things that were in my mind, but I didn't have sort of the ability. And then the second sort of desire I think was not that I have any special technological skills whatsoever, but my constant desire is figuring out how can we try more cases at a more cost effective level. And to me, anything that I can do to that way. So in the past I've tried to cut down outside vendor costs to make trial more efficient, effective. To me this was just the next step because if I can reduce trial cost, I can increase the amount of trials we can try is just the bottom line because now the risk ratio is better in that respect. So honestly, it was just a desire to figure out a way to get to more trials in a way that works for the client and us. And this was the path forward.
- [27:18]
A
Yeah, it seems like you're, you're doing, you're doing all the right things. Everything I've read about AI is you're, you're training it on reliable data, focusing on the prompts, which as you pointed out, it's not just like doing a Google search from reviewing it as an experienced lawyer. You're asking it to look at logic, which is not the same as necessarily relying solely on case law, which is, as you said, we were using logic way before refusing case law. And then you're, you're, you're coupling it too with experience. And you mentioned, I'm sorry, I'm like Ms. Williams, you know, you said she's, she's strong at openings and so you really, you're really putting together the technology with, with the humans then, then of course you're reviewing everything. You're not taking anything straight out AI, just submitting it to the court, which is famously or infamously led to some like false precedents and things that were, that were hallucinated. But it sounds like you're doing all the right things. And I'm just wondering, as you said, you're going to see it developing in video. I think you're absolutely right on that. How do you See it evolving further in litigation, the kind that you're dealing with.
- [28:26]
B
Let's do near term, far term. Right. Because it's crazy how fast things have gone even, even since the first trial, used AI at this point, eight months ago, versus the other two trials of used AI. It's been remarkably fast, even this year. But every time we have a case, we're always trying to get a sense of our opposing counsel. Are they going to be difficult to work with? Are they going to be hard to work with all those pieces? So no one's really talking about the opposing counsel piece. And good counsel on both sides lead to a good outcome for all clients. So it's not to suggest that you can't have a good relationship. In fact, I think you oftentimes get the best outcome for both clients if you are in that world. But what's interesting is again, typically it takes us six, 10 months, a year to get a vibe for opposing counsel before we really understand kind of how are we going to coexist. Who are they, all those pieces. So I think starting to develop effectively avatars of your opposing counsel is a near term thing. Right. Here's my opposing counsel. They have this background, they have a clerkship from this, they spent this time in government work, all those things. I think we can start to get pretty near term sense of our opposing counsel pretty quickly. And why is that helpful? One, obviously, knowing your opponent just makes it better for you and your client. Number two, budgeting. Right. You know, a lot of times the budget in a case for litigation, which is difficult to budget in the begin with, gets really easier to dial in once I understand who the other side hires. Are they a firm that tries every case? Do they not try every case? Do they do a bunch of depositions, sometimes needlessly not? So I think getting that veneer for your opposing counsel far quicker is going to help with budgeting, predictability, all those things, just to give a sense of the client, like, what are we doing on this thing? Is a near term sort of evolution that we're going to see on the AI piece long term. I think once we dial in the predictive outcomes of trial in cases, I think it's going to be fascinating. So, you know, it's. It's hard to quantify the value of litigation because there's so much risk and improbability. But to me, how we demonstrate value to the client. We start the case. Let's say we're assuming plaintiffs side in a complex case. Okay, we start the case. We got to Survive a motion to dismiss, perhaps predictive AI in the future says when you file this case, you got a 15% chance of taking this case from the beginning to end. You got a bunch of roadblocks because you got just issues. But we prevail on the motion to dismiss. No motion to dismiss awarded. We move to the next phase. Some evidence gets better. In the case of now, our predictive outcome is more like 40% and then we get to trial and maybe we're at 75%. And so the point being is, I think we're going to be able to demonstrate how have we provided actual value to the case because we've gotten it better or flip on the defense side. Looks like it's a case that has a bunch of issues where you may have significant exposure, clear damages, you've got a 20% chance of prevailing defendant at the beginning of the case based on predictive AI. But through thoughtful motion practice, knocking out the other side's expert, you know, using your gunpowder at the right time, you've knocked that exposure down to 35%. I feel like if I'm on defense side and I can say, hey client, you had an 80% risk when we started this with a massive eight, nine figure liability. But through our work, we've cut your risk down to 40%. Let us take it to the last phase and let us get you down to zero percent at trial. I think it's going to make clients far more informed about their risk reward ratio and so that when we do go to trial, the risk is clearly assessed on both sides in a much more thoughtful way than currently. Tell me what my chances are. Which at the end of the day just isn't enough. I think the clients deserve more and I think we can figure out ways to give better quality. So I think that's down the road before we have a level of confidence in it. But it sure would be nice to give the client some objective facts and data to work off of when they're making decisions that could result in their company being around or not being around at the end of the day.
- [32:25]
A
Yeah, I mean I am seeing services saying that they can do this, they can help with predictions and stuff. So yeah, I think, I think that's inevitable, that that's going to happen. My final question is going to be around your clients. So they first, are they. Do you know if they're using AI in their operations and how do they feel about you using AI? Is it something they like or want you to definitely want you to use or. So tell me about your clients that way.
- [32:47]
B
You know, I think what's interesting is a couple of things on that front. Number one, normally it's the law firms, in my view, walking out, hey, we've got this new tool, we've got this new flashy item. But this is one where I think the clients adopted faster than the law firms. We're seeing it in a lot of RFPs and requests saying, like, how are you going to use it? And not just, you know, superficially, like give us real outcomes. So I think the clients especially, which is not surprising, they're early adopters far more than law firms are. I think they are absolutely ahead of us. So I think this is one where it's going to be client driven. And you can't just say, oh, we're not doing that. So that's my number one take. And then I think you are going to see especially like, you look at a lot of the things that really are mechanical driven, you know, maintaining databases for all your major contracts, things like that. I think it's going to be super helpful for clients and for the lawyers. It's amazing. I'm always surprised, even with large companies, how difficult it is to get documents. You would think there'd be clear repositories, and they're not, you know, so my hope is that when I go to them, I need this particular information. I did it far faster from the discovery process because they have much better document management system because they can actually find the stuff I'm looking for, rather than us just going through, you know, nine different bad paths to find the right one. And so I think you're going to see a convergence, but I think the clients are going to drive this more than anything else.
- [34:12]
A
Yeah, yeah. Certainly a lot of the big companies are big into it, both in their, in their operations and in their services that they're providing to customers so. Well, Adam Massaro, thank you very much for talking with me about this today. Wish we could have done it in person because I'd like to get to Colorado, but thank you for this.
- [34:28]
B
Well, we welcome you anytime and I really appreciate the opportunity. So thank you so much for delving into this really interesting topic.
- [34:58]
A
Sam.