From AI Hype to Legal Reality: What Litigation Support Teams Are Actually Using

Yeah. Thanks for having me, Daniel and Eric, and you nailed the pronunciation. So I think we're off to a rallying start.

Yeah. Happy to do that. So litigation support, I think, is one of those things that's super necessary for large firms that deal with cases of any that have any data volume or any size.

And some of the stuff, while I'm sure we'll get into, you know, topics like generative AI down the road, you know, and I think we can click off the box. We've said generative AI, so we're covered for the podcast. A lot of it is honestly more baseline of keeping things organized, you know, making sure that you're handling things as efficiently as possible, and frankly, keeping the teams together. And for better or worse, at our firm, we typically don't have super large case teams.

So, typically, it will be a partner or two and an associate or two, a paralegal, and then the litigation support department helps. We have a ginormous litigation support department of two people. So my colleague Alan and I run the department and, you know, some might say Alan does runs the department, and I help out. But what we try to do is help the attorneys get through the collection review and production process as efficiently as possible, which sort of spans all the entire EDRM.

The only area we don't really help with on that front would be on the collection side.

We do use some tools internally depending on client sophistication, but typically we use third party vendors to help us with the collection. But the rest of the stuff we handle in house, and our tools have sort of evolved over the years. You know? So we started with some tools and have migrated to other tools, but the main thing is to handle the cases and meet the case teams where they need things at the time that they need it.

And I'm not saying it never happens, but, you know, for the most part, our teams are very organized and timely with us. So that we don't have some Friday afternoon fire drills that we have to deal with, but typically, we're in a good position to know where things are and to have things planned out and handle things accordingly. And I think to the second part of your question was what are two or three pain points maybe? Is that what you were driving at?

I think one that stands out on the front end, which is a good indicator of how the case is gonna go, is sort of how negotiating the ESI protocol works. So if the other side is really pushing certain things, I think it's actually an indicator to us that it's gonna be a potentially tricky engagement with them. Meaning, if they're pushing back on what what I think are fairly straightforward or industry standard things, I think that at least indicates to us that they might be a little bit difficult to deal with. And I don't wanna paint with too broad of a brush, but I will say if we try to get intervention from the courts, I can't say that we often get a ton of relief in those scenarios.

So, typically, the judges, whether it be a magistrate judge or the judge handling the case, tells us to work this out amongst ourselves. And so that's what we're trying to do. And oftentimes, if there's a ton of pushback or if we have to lay out a lot more information than we typically would in one of these ESI protocols, we know that the other side might be difficult. And I say that only because I found that the standards that we get held to sometimes feel like they're higher than the other side is held to.

And and why I say that is typically we're the larger you know, we're the defendant in most cases and we're the larger entity. So I think what they're trying to set up is, you know, asks of us that they might not even have to deal with. And, again, I'm speaking in sort of general terms. I'm happy to plug into some specifics, but I think that's one pain point on the front end that you'll know.

And and where that sort of rears its ugly head is that we'll have a highly negotiated ESI protocol, what metadata fields we're gonna produce, how we're gonna do it, the timely manner, and smash cut to them producing me one large unsearchable PDF. And my response is, why do we spend all that time going over that? And I don't think at any point would somebody put one large unsearchable PDF as the production format. So maybe that's a way to cover it on the front end to know that it might be a pain point and then actually to see its rear its ugly head digitally on the back end.

It's a good question. And and frankly, I think in the generative AI world, the playing field has been leveled a lot more. In fact, the more that I'm going to conferences and talking with people that sort of reside on both sides of the v, I actually think generative AI might be helping on the plaintiff side more than on the defense side. I think these tools, while there is some front end work to get off the ground and running with them, I think that's lower.

Right? Because even some of the generative AI tools, you can ask them how to help you to get off the ground with them. Right? Whereas the document review platforms you mentioned before or the ones that are available today, there aren't many that you can ask and say, hey.

How do I administer this? Like, how how do I load this data? How do I push this data through? Whereas a number of the generative AI tools sort of hold your hand going through.

Now they may be holding your hand in the wrong direction. You have to make sure that what you're asking them and what they're giving back to you is actually correct and cited in the documents, but these tools actually help you push forward. And so I think that's leveled the playing field, and I'll be interested to find out. I think it's probably a you know, we used to say in years how these things would change.

I think we're more in months now. Meaning, when you say this is a twenty twenty seven initiative, I don't think people can get away with that in, you know, twenty twenty six here when we're recording. Right? I guess unless you're right up against the end of the year because these things are evolving so fast and they provide such utility that you need to push forward and utilize them where you can.

And I will say, I guess the last point I would make is that the appetite for attorneys and I can speak about my firm, but also other firms that I've talked to that might be similarly situated. The appetite is actually very high to use these technologies. Meaning, they're able to see the results and they're able to work with them because the technologies are more in their hands. Whereas I think a number of the technologies before that, they weren't able to see the tangible results.

So it led to a lot more front end selling to get buy in and to be pushed where you needed to use the technology because you had no other option, and then you did it. Whereas now, I might even say some of the technologies I don't wanna say it's the technologies that's wagging the dog, you know, the tail that's wagging the dog, but I do think some of these technologies are sort of being pushed forward and used more. So I actually think it's an opportunity. It's just being able to vet these, you know, make sure that they work and understand how to use them.

And, thankfully, that for better or worse falls on my plate, so I try to do that.

There is definitely an uptick in yes. So the short answer is yes. But to your answer to your question, Danielle, yes, it then results in being able to file more cases. And I think there are positive sides to that too.

You know, I think it can open up opportunities where people might not have been able to do that before. I think the hard part becomes we do need to have a baseline level of competency or something that we have to respond to? Because I don't think it'll happen in twenty twenty six, maybe not in twenty twenty seven, but I do wonder if there's gonna be a minimum threshold where some of these cases that are very clearly, you know, generative AI slop cases that have been filed against, you know, let's say, twenty different companies, the same thing just kind of rinse and repeat, change the name, file it, and just badger them.

I do wonder if there will be a generative AI solution, and maybe this is a business idea for somebody out there to just find these, run them through, know that this is a pattern that's happened before, and have a sort of canned response. You're having to respond to what might not even be a valid claim. Right? So it's a tricky situation where you can't ignore it, You have to respond to it, but it might not be the most productive thing in the world.

Right?

So the short answer is they do care. Right? They're into the technical nuances, and maybe it's just the type of firm that I work at or maybe this is me just complimenting the partners associates at our firm. Right?

So we can cut this clip out and I can send it to all of them that they are in fact interested. But what I'll say about that is I think part of why they're interested or why they wanna get into the weeds and know about this is that you don't have the initial feedback right away that you might have had in the technology assisted review world. After you review a handful of documents, pretty quickly you start getting documents you care about. Right?

Because you've coded documents that you think are relevant or not relevant, pretty quickly the machine learns that. I don't think we wanna dive into the nuances of the statistics associated with it or what you do on the back end to validate it. There is this thing where the attorneys right away are able to see, yeah, I'm getting more of what I coded. In fact, sometimes the problem becomes if we go back to the ESI example where we're producing more documents than maybe even we want to, sometimes we're actually coding documents relevant such that we know it's gonna bubble up a lot more documents that we don't really particularly care about, but we need to get through them.

I think that kind of validates the process. Right? You're seeing that happen. When you get into the black box component of it, I think we almost need to split it into two separate discussions.

I think there's the discussion of a confined universe, a small amount of documents that you might be using a generative AI tool on where you know what the inputs are and you can validate that versus dealing with a large corpus of documents because you don't necessarily know what you don't know. Right? And I think one of the things that's been helpful with the tools to allow attorneys to have some comfort is that not only are there sort of review capabilities built into some of these generative AI tools that allow you to get through them faster, hopefully? But then there's also sort of a chat feature available in some of the tools that allow you to backstop what you're seeing or lets you kind of go on a little bit of wild goose chase.

Because, frankly, the part of generative AI that I think has led to it being adopted so much, especially at law firms or corporations, you know, in the areas that we see, are that you can get feedback. Right? It's just the black box component of it. You need to work through what you're feeding to the actual system.

Right? So, again, small amount, pretty easy to validate. Larger amount, not as easy to validate. I don't know if we'll have time to talk about this today, but I feel like some of the generative AI tools that deal with small corpuses of documents are being compared to tools that deal with larger corpuses of documents.

And and it's not a fair comparison, right, because they're serving different needs or they're trying to provide solutions to both, and they're frankly being held to the same standard, which is not not necessarily fair to those tools. Right?

I think there's gonna be sloppy lawyering in any domain. Right? So if we're if we're leaving aside the sloppy lawyering, right, where you're citing the cases that are not actually real cases, that's its own issue. And and I think it creates headlines, and it's it you know, and it's even happened at large law firms.

And I'm not doing that to drag any firm per se. It's just it's out there, and that's a unfortunate situation because I think in the just to use the technology assisted review world, I think people made some generalizations based off of maybe small, you know, differences or, you know, sort of small thoughts. And I think the same might happen in the generative AI world. And and by no means am I saying, I don't think we're putting the GenAI back in the bottle.

Right? Like, I think we're going forward. We're gonna have to figure out how to use this. And, frankly, the example that you gave of you asked the same question, you kinda get three different answers, is a unique and interesting one.

And and I brought this one up internally. I saw a vendor present where they said, you know, let's just go over my bio. Right? And and they asked a question and it gave back the bio.

And then what they did is they put up one the proper version of the bio and then the generative AI created version of the bio. And Gen AI is just trying to guess the next word. Right? It's just trying to answer the question.

And so I almost wish there was, like, a confidence score because if we're talking about a twenty two percent likelihood that this is correct, that's a lot different than an eighty two percent likelihood that it's correct. And it was interesting to see sort of a heat map over the two bios where it was very clear some biographical information was always gonna be correct. You know, where the person attended school, you know, degrees they got, things like that. But then there were just some, you know, exaggerations, you know, things that were just thrown in there that it was just a guess.

And and I understand that you can't do the math equation on every word in something, but it would be great if you could essentially put that, you know, number together and say, yeah. This is how confident we are. It's correct. Because frankly and I just keep coming back to technology-assisted review.

But frankly, the help of getting rank scores on documents are that it's not how important the document is. It's how likely that document is relevant. Right? The same thing works in the generative AI world.

There there are scoring methodologies, and you work through those, and you do sampling of them, and you test to make sure that it's correct. But I will say it's very nice to see a graph of that and then to look at the outliers. Right? Because there could be instances where a document has a score that is totally different than the way it's coded, and there could be a valid reason.

And we're not able to dig into every single one of those depending on the size of the case. Usually, we review them, but it's not as if you can necessarily run out every single instance, but at least it lets you look at it again, double check that it's correct, or that you agree with the decision, and then you can push forward. And so I think that's part of it is trying to find the right areas to sort of catch those or to surface those so that it also improves the process and make sure that the results are correct.

Yeah. It's a very timely question. It's something we're dealing with literally on, a daily basis. Right?

Because the latest and greatest, like you said, is always coming out. And the ability to use those tools, I think we talked about earlier in our chat here, the barrier to using those is lower. Right? You don't have to stand up a bunch of servers, and you don't have to, you know, hire a staff to do this.

It's all SaaS based, and so you can kinda run with it. So there have to be protocols in place. And and, thankfully, our firm has, you know, a number of checks in place. I would say the first level check of whether you can even use a technology goes back to the outside counsel guidelines.

Right? Like, we've spent a ton of time curating our outside counsel guidelines, and we've actually used our DMS because we need to put in ethical walls and we need to put in restrictions depending on what certain clients require of us or don't allow us to do.

And so that's a good first step of knowing whether we can use generative AI for clients or not. Right? That's a threshold question. And then in terms of whether a tool will actually work for us, we have to kinda work backwards on what the result would be to analyze whether it's something that we can work with. And and some clients actually interpose getting approval from them for the use of certain tools because they might restrict how vendors work. Right? They might say, here's our preferred vendor list, or here's the things that you can do.

So I think there are a couple traps on the front end that are helpful. Does that mean that there are attorneys that are not out there evaluating tools on a daily basis and we maybe don't hear about it right away? I'm sure that's happening. You know? But I think before anything's actually used, it would come into the firm. We would do an analysis of that, and we would have sort of different checks before they're actually implemented. And I don't know if it's the perfect analogy, but I can say fairly, you know, generally speaking, but to talk about a sector of business.

When we went from on prem to the cloud, originally, there were certain sectors that were super nervous about that. So we do a lot of work for banking clients. Right? The banking industry was pretty skeptical on the front end of cloud services.

And pretty quickly, they went from skeptical, and I'm I'm talking generally. I'm not talking about a specific line of our firm. Right? But pretty quickly, they said, actually, this provides a lot more security than a sort of one off solution that might be hosted in someone's, you know, hardware environment that we don't even know.

Right? This is a much more secure, you know, corporate cloud structure. Again, I think cloud was a terrible idea because none of this is literally in the cloud unless they change how this works. But, you know, quickly that industry said, no.

Actually, we would prefer that you go this way. And and that worked out well for our firm because we were sort of moving from some on prem solutions to some cloud solutions, and there was some worry of if our clients weren't comfortable with that, would we have to change systems that we used? Right? Because I mean, I don't know that we wanna be on the bleeding edge, although I think now generative AI is forcing people to the bleeding edge, but we wanna be on the cutting edge of technology.

Right? And so if we were gonna stay on prem for some of these technologies, it felt like we were behind the times. And so we were pushing forward, and it was nice to have a, you know, sector that we service for our firm that was also kinda coming along with that. Right?

So it's it was kind of an I don't wanna say in partnership. Right? But, you know, it was one of those things where as the partners are having those discussions with the, you know, corporate clients, it just made sense. And so then that actually was an inertia to move towards that.

Yeah. It's a fair question. And to say that I don't still get some cold calls and emails would not be fair. Right?

Obviously, I'm still getting those. But I actually think the ones that resonate more or if they maybe they happen at a conference where you can tell that there's been some legwork on the front end to maybe either identify a problem that we might be facing or establish a relationship, meaning, hey. You know, I'm coming in for this, and I'd love to get, you know, fifteen minutes or a cup of coffee or whatever and discuss that and try to bring an actual solution as opposed to saying, I'm trying to sell you a seat license. Right?

And I get it that everybody has sort of bosses that they have to, you know, bring results to, but I think people are actually getting much more innovative about how they're approaching that and coming with solutions or frankly just hanging around. Right? Meaning available and because these technologies are moving so fast, the adoption is so much quicker that if you were to say, hey, you know, check back with me in q three, it might actually change.

And frankly, before when they were saying, oh, you know what? We're gonna roll out some new features to our tool. And, you know, we have our conference and it's always in the fall and, you know, we're gonna show you there and that'll be great. I mean, these features come out without even maybe salespeople knowing that the features came out.

Right? So I think it actually speeds everything up. And I think it was Eric that said it earlier that, like, we're not quite ready to, like, process all of this. There is a component where if some of these things actually do what they say they do, it's like it is kinda mind blowing.

Right? Now I don't wanna say I'm skeptical that they actually pay that off, but I don't know that they always do. So some of it's actually seeing it and then talking to others that have either tried it or looked at it because that's the other area is that there are two or three technologies that we've either brought in or, you know, got real close to the line of bringing in that they were mostly based on a referral from, you know, somebody in my position but in another firm that I trust highly. Right?

And so I think you're correct that the information is similar or, in fact, the salespeople might have more information than the law firm themselves. But what I think they do know is that their captive audience that use their tools and really like them, they need to be great with them. Now, you know, and I know we're not talking about specific vendors. The larger the vendor, you know, and, again, I'm talking more on the, like, software provider side, the less that they can move and change quickly.

But the smaller the vendor or the newer the vendor, they take that feedback quickly. And, frankly, I think they get in the door a lot more often because they're saying, hey. I've heard from Chris that he needs this, or I've heard from whoever that they need that, and we can execute on that and deliver it quickly. And, again, if it doesn't meet the needs of the customers, they shouldn't build it.

I don't wanna be the person that asks for a million things that are not useful for the market in general. But I do think generally speaking, when I talk to colleagues that are similarly situated, we kinda know the things that we need or the pain points that we have. And if someone else has something that addresses that, that opens the door quickly. Right?

And it gets the discussion going. And frankly, I'm less worried about people like who manages this or who runs that. Meaning, I think a lot of partners are also approached on this stuff. Whereas before, I think the idea of approaching attorneys was not possible because they couldn't sort of fathom how to deal with the overall enterprise component of this.

And and, again, that's where this stuff's driving, but part of it is, you know, trying to find an answer that helps in a little area and then does it make sense? I think my only maybe this is a place to finish my thought on it, but the only one that does worry me a little bit is that a lot of these solutions that are coming out now are super specific, and I think they provide a solution. But maybe it's so minute that if you had to license, let's say, ten of these to get something that you've got in one or two solutions, overall solutions, it makes it hard. Right?

Because it whether it means moving data, sort of vetting these, you know, the logistics of dealing with those tools, that hasn't caught up. Right? So in the generative AI world, the, like, vetting process hasn't necessarily caught up to the speed with which they're coming out. So that's the part where it gets a little bit tricky of seeing how it works in the overall platform or the overall usage at the firm.

So leaving lawyers aside, I think there are some administrative components that we're using AI to clean up and speed through that have been super helpful.

And my colleague Alan gave a great example to our group, you know, recently where when you do get the bad document production or the, you know, nonsearchable PDF, you can throw this at the generative AI. And if we have certain skills built in or certain ways that we cut through that quickly, you can actually pull apart that stuff and make it more useful a lot faster than you did before. And since we said we weren't gonna talk about specific tools, I won't bring up any names other than to say there are some that I think do a great job at one, you know, or two things, and then there are others that do a great job at larger things, like I said, about the sort of platform versus specific tools.

But I think it's really helped on the sort of back end administrative side. And then, frankly, something we're looking at internally, you know, we built, you know, simple ticketing system to handle different things internally. How we can improve upon that or how we can actually just sort of take what we have and then visualize it more, actually find out where there are areas that in the past, this is what we did. And then we kind of know that this is always the approach that we handle for this client or for this type of matter, it's super helpful on the back end for us because we're, like I said earlier, a ginormous team of two.

So we're able to communicate with each other. But, frankly, sometimes it's like, you know, like, today, it just happens to be that my colleague Alan's off, and then I'm I'm off tomorrow. So, you know, it's one of those things where it helps us from not having to bother that person. And, again, if there's something that's super mission critical and we need to talk about it, obviously, we're gonna connect and either talk on the phone or email, whatever it is.

But I think it's let us actually handle things on our own a little bit easier. Right? Because you need to have the stuff come in, break it down, and then try to find how it's reusable. Because that's ultimately where I think the savings will actually come in.

Because that's the gap that clients seem to talk about corporate clients talk about a lot is that there's a ton of talk of generative AI use and well, not a ton of specifics. Right? People don't really wanna show their cards.

But, ultimately, I think where this goes is if you're able to drive efficiencies such that it's more cost effective for clients, then the corporate client's gonna be happy about that. Now the counterargument to that would be that, you know, whether it's vendors or law firms were, you know, for profit businesses that need to have business available. So then it's either are you handling more matters? Are you, you know, handling them more efficiently so that it frees up your resources to handle different things?

And those are things that will shake out. Again, I don't have answers to that stuff, but it has offered, you know, what seemed like, you know, marginal wins. But I think getting those even small marginal wins internally, whether it be on a sort of administrative function to make it run a little bit more efficiently, that's great. You know?

And and those are totally encouraged because it's just less time you have to spend on something that you're gonna have to do anyways or that you might have to do ten times the rest of this month. Right?

What do the next six weeks look like for me? There are a couple tools that we brought in house or that we're evaluating that I think the timeline of actually what are we gonna use this for and what's the payoff to that would be like, oh, let's see if we can figure that out by the end of this year. Right? Whereas the evaluation period is so much shorter now and the results need to be delivered so much faster that in the next six weeks, I think there is one or two tools that will either become in house tools that we regularly use, and I think that has its own knock on effects that I don't think we have discussed a ton, or will be something that we tested and we've now moved along and decided that that's not gonna work for us.

And and the part that has not caught up of you might use this tool in the next six weeks is that, sure, there are certain pieces of litigation that are going faster. But, typically, these are matters that are handled over years, not months. Right? So as you bring these tools in, you're sort of saying we're agreeing to using these tools for x amount of time.

Right? So I think for the, you know, software provider side or the technologies that do in fact get in the door, when they get these short term wins, I think they do have these long term gains. Right? Because the likelihood that you can actually get off of those systems, you know, having moved our firm in many years back from a different platform to the platform that we're on now that was on prem to a platform that's in the cloud, The hassle of that, I don't think I have too many gray hairs about it, but I feel like it did take some years off my life.

Those same would apply, and I frankly don't know how you would pull that out of those tools and then move it to another tool. Right? Because they're so specific on what they're handling that I don't think you could actually take that and move on to another tool. So that's what makes those decisions pretty important and, you know, carry a long tail.

Right? So there might be a short evaluation that has a long life in terms of how it would live within our environment or what we would use. And so that's why we try to be, you know, obviously as judicious as we can with our, you know, firm's resources for what we bring in and how we use them because they're gonna be around and we're gonna have to support them.

And, you know, it'll be here for again, while the technology has changed by the months, the cases don't change by the months. They change by the years. So that's where we're trying to measure both and use them correctly.