Legal Tech Talk Midyear Episode

I'm Daniel Gold, BDO principal and eDiscovery managed services leader.

And I'm Eric Derk, BDO's managing director of legal operations. Let's get started with this episode's exciting discussion.

And welcome back to another episode of BDO Legal Tech Talk podcast. I'm Daniel Gold, and as always, here with my cohost, Eric Derk. And, Eric, we are definitely doing something just a little bit different here today. We're not sitting down with a single guest, and we're not replaying one episode.

We're actually gonna take a step back in halfway through the year and asking a much broader question, which is this. What have we really learned from our podcast guests from the first half of twenty twenty six? Because, look, when I look across the conversations we have had so far this year, and we've had some really amazing guests, really very diverse, I don't see a random collection of interviews. There's actually a through line, Eric, that has gone through all of this.

And we've talked to legal operations leaders like yourself. We've talked to privacy lawyers. We've talked to eDiscovery professionals. We have, for the first time in four years on this podcast, talked to investors, which is very important.

And we talked to law firm technologists, AI governance thinkers.

People really sitting at what I think is, and it sounds very cliche to say, but really at the intersection of legal data, capital, which is new, and change. And, Eric, I think that there is a very, very clear pattern that is being created here that we created, right, from the first half of this year. AI is obviously the headline. We can't go through a single sentence without talking about AI, but I don't think that's really the real story.

I think it's too simple to say that, you know, AI is the story or say that AI is coming. That's over. Right? I think that really what the real story is that what AI is forcing the legal industry to confront, This is what we've been taught by our guests, which is how lawyers learn and how legal work is priced and how is it governed and how does trust get built and how much responsibility do humans still have when the tools become more powerful.

And that's obviously something that, you know, I'm very, very passionate about. So, Eric, given that this is like our mid year review, right, I wanna really start with that. When you think about the conversations with this prestigious guests that we've had throughout the course of the first half of twenty twenty six, what is the first thing that's standing out to you?

The thing that stands out to me for the most part is that it's almost not about the technology.

I think that the market drives the idea that technology, AI technology in this case is gonna change the way they work. Now that might be true. But in order to actually get away from this shiny new tool idea, A lot of our guests were talking about how legal teams need to change how they work. And that that's really supported by that underpinning of process definition and design. And we know that in the legal space, process is a dirty word. Right? We don't really like process because lawyers and supporting folks tend to think that, well, every situation matters different and requires unique sets of thinking.

But I think what we're now starting to see is it's not about a shiny tool. It's about how these tools can support processes, yes, processes, in a more effective way and standardize certain types of work. Now it's not all work, but there's lots lots of substrates of work that can be actually better. I'm not saying automated.

I'm saying that the shiny new thing is not the thing. It's about how we think about changing the way in which we well, let's talk about predictive coding and tar and what we did with analytics. What we did with CLM, contract life cycle management. Now we're doing with Gen AI.

What really stood out to me was it's about process, governance, accountability, value, and then technology. Like but if you read the papers, it's technology first. Buy my thing, and then we'll we'll retrofit everything into something that we are accustomed to. Right?

And typically speaking, legal teams would just get handed technologies and they'd have to figure them out. And that's kind of just an IT function. Right? Well, IT is gonna solicit and figure out they're gonna go to these shows and they're gonna figure out what might best work for a legal team.

Well, that's always been the wrong way to do it. But now you're seeing a leadership vision come forward with a lot of these these teams that say, hey. We wanna we wanna recraft the thinking about billable arrangements. We wanna come up with different fee arrange.

But at the end of the day, it's all about values and outcomes. And so we need to challenge our industry to say, how do you measure value? Isn't that a qualitative metric?

How do you measure outcome?

To me, just think that we're not used to thinking like that. We're very quantitative. We think about billable hours and we think about costs and that's about it. So I think that a lot of the podcasters that we interviewed were really focusing on this kind of value idea and how we change some of the antiquated constructs.

I love that. And, you know, and what you just honed in on is great, is the outcomes. And what I wanna do as far as an outcome for this episode is let's talk about all of our guests in five different buckets. Right?

So in the next forty five minutes, we're gonna basically tackle. And what Eric and I did was we found five very good themes with the help of AI, of course. Five themes that continuously came up with all of the guests, and we've had a lot of guests so far in that first half of the year that we saw throughout this. And one of the things we saw was that, to your point, is that, you know, we look at this as being an efficiency and we look at this being a productivity tool and look how, you know, how we can unlock all of these great things.

But to your point, there is a hundred percent in outcome focus and outcomes are created by whom? Humans. So let's start with the very first theme that we're gonna cover here, which is and I love this. This is not our word.

This is something that came up by one of our guests. AI as a superpower with an apprenticeship problem. And I love this. So so one of the clearest themes from the first half of the year is that AI is in fact becoming a superpower for experienced professionals.

But that same superpower actually creates an apprenticeship problem for the next generation. We talked about this over and over again. So the superpower word came from Kim Wolf from State Street Bank, and she captured the challenge beautifully in her words, not ours. She said, but, you know, when I think about it, for me, AI is my superpower.

It makes me better. It makes me smarter and quicker.

Look. That, Eric, is an optimistic side. For someone who has already built judgment, a hundred percent AI is going to be your superpower. Alright? It happens for you on what you do on a daily basis. It happens for me as well.

We can pressure test our thinking. It it in some sense, while we pressure test our thinking together on a daily basis, Eric, in our day to day job, right, and what we're doing, we constantly utilize what we call our AI friends to be able to pressure test how we are thinking. And, you know, and then Kim immediately raised a concern that I think every single law firm and legal operations professional, they should all be talking about. She said, in her words, and if they're not gonna cut their teeth learning all this stuff at the level of detail that they can spot an error because AI does make lots of mistakes and I can catch them, who's going to catch them when I'm retired?

So, Eric, you know, we have talked about this over and again about the next generation of workers. So whether they be legal operations professionals, litigation support professionals, paralegals, and, of course, attorneys, next generation of workers that come in, if AI has always been their go to, unlike us who started off you know, I started off practice with books, because we've talked about many times on this podcast. I started off with books. I started off legal pads, sticky notes, highlighters, and a Sharpie.

Right? And and so and copy machines and faxes. Right? And dictaphones. Right? That's how I started things off.

So I learned the hard way, which was the only way to find case law, to find statutes, to find codes. Yep. But if every single person comes in and they're just using AI, Kim has such a great point. Right? That may be one of the most important questions of the year. What do you think, Eric?

Well, I think Kim's I loved talking to Kim. She was awesome. And I think Kim's ultimate insight was about we talk about governance. Right?

But we really talk about oversight. And it's a little bit different. But if you see what AI generates and the tools that we use every day, a lot of times it's AI sloppish. Right?

It's stuff that and in fact, you and I have started to actually come up with tells.

Right.

And we're using AI against itself. So Claude has a whole you can see there's a whole slew of folks in the industry now. There's there's you can go on TikTok or wherever you wanna go.

And there's there's small little businesses that have said, hey, make this more human. You know, make this more human. I mean, make in other words, make sure that it doesn't sound like a machine. Right?

Well, that's Kim's main point. Like, you have to have governance and oversight, but you have to know and you have to be able to read and research. Your theories on KRINO, which is essentially a quality control process, really pounded on that. We can get into that later, but I think it's an efficiency discussion at the heart.

The industry keeps asking like, is AI gonna save time? And some people say, of course it will. Other people like, well, yeah, but what's the professional output information if if it does save us time? What do we lose if AI is placed in the wrong places?

Like, do we lose the ability to critically think? Are we just gonna take this this output writ large and say, this is what AI told me. I created a prompt, a Google ish type prompt. Well, no.

That's not really that's not really the answer. You have to know what good looks like, and you have to know what the output is in the qualitative standpoint and when it's maybe off. So we work with AI tools all the time, and I continually go back and say, hey. Check double check your work.

And, Daniel, you know this. Like, how many times does AI come back and apologize to you?

A hundred percent. Right. Yes.

Many many times. It's only the humans that know the exceptions and they only know the context on from the institution. A machine does not know the institutional contact. I think that's a really important point. So as we're all kind of getting fueled by all this great stuff, tools, process, tech, people, strategy. What we continue to come back to is, hey, wait, you better double check your work. Just like just like a partner would check an associates work or vice versa.

You know as an expert or somebody who's a qualitative, you know, analyst, you'll know if something comes forward that's just gobbledygook. We call it AI slop. Here's the other thing. So I talked to my friends in banks and, you know, all over the place.

Right? And one of my friends who we've had on the podcast, Emily Diamond, has said, hey. Listen. I'm really worried about this AI thing.

Like, what's gonna happen to the first and second year associates or maybe the first and second year litigation support person who's coming up as an AI native. Right? What where is their muscle memory and what is going to be that push to train the mind. And the other thing I'll say, Daniel, is we had James Sherer on the podcast.

We've talked to James a couple times and he's got this really cool cognitive miser idea.

Yeah. Why don't you talk about what that means, a cognitive miser? It's a really important point.

So essentially, when we have so much information to look at, your brain needs to control and quarantine certain things. Right? You let's say you're building an argument. Right?

Well, you might have hundreds and hundreds of ideas that are given to you by AI, but you need to constantly refine and hone and tune. Right? You need to be a miser almost in the information that's coming to you. Now the flip side of that, and this is where I think when you go back to, you know, other podcasts, you know, the stuff that Kim said was really cool because she was really talking about expanding the knowledge landscape and bring other ways of working other knowledge sets.

So we're kind of at this kind of inflection point where we have so much information that's being fed into us and we have an opportunity. Now I don't know if anybody's gonna really kind of dig into this too hard. We have an opportunity to change the efficiency and outcome thought process. Right?

We can bring in more insights. We can say, yes. That's good. Yes. That's bad. We can tune and hone.

And what you and I have done with other people here is that we've learned that it's like it's data is just a big ball of clay. It's it's Gumby. Right? And you have to basically shape and tune and shape and tune and shape and tune.

And it almost, I would argue that some of the podcasters have said, it's creating more work for us.

So That's right. That's right.

I think that you and I both witnessed this firsthand. I like, oh my god. I created a prompt and it gave me a hundred page output.

Right. Wait. What do you do with that? Yeah.

Yeah. Now now now I've just created a ton of work when all I'm looking for is a one page memo.

Correct.

To build an argument. So, like, we have this cognitive miser idea that's maybe not aligning to, hey. You can do more. You can you can understand data in a much more broad broad context and you can build a better outcome. Now, don't know if everybody's gonna be into that, and I think that's an inflection point.

Well, so it's interesting because in our first season of BDO Legal Tech Talk, we had Phil Weldon from Hecker Fink, boutique law firm in New York, be on the show. And I was just in New York, as you know, Eric, and I was talking with Phil. He's gonna actually come back onto the program to talk more about this in the latter half of twenty twenty six, which I'm very excited about. And one of the things that Phil said was is that he actually has increased the amount of work because he uses AI.

And his clients know that he uses AI, and so as a result, he's getting more work. He's doing it more efficiently. He's doing it better. So the outcomes that he's getting it is great.

And I think that's a terrific use case, but I wanna go back to some of our guests that we've had on the show. You mentioned James before. And and you've also talked about responsibility. And James talked about responsibility.

And one of the things that he said so beautifully on the program besides giving us this cognitive miser phrase that we use all the time, if it crosses your desk as an attorney, as a practitioner, it's your responsibility. And you're right. First of all, that goes for sure to the KRINO framework that I created. We could talk about later.

But the KRINO framework is basically it's giving voice and words to what we already know in the rules of professional responsibility that we've talked about on this podcast, which is at the end of the day, in this entire legal industry, it is ultimately the attorney that signs her name that is always going to be responsible for the work product. Always.

Totally.

And and, you know, one other guest that we had on the program, Hannah G. Odo, she made actually a related point from a completely different angle. She said, in talking about technology and people, she said the one sustainable resource of the company is people. Right?

In other words, it's the human. Right? That can use the technology and that can adapt to change. And I think that that is something that you mentioned right up front in the very beginning is that technology should be last.

So you gave us litany of different things that need to be considered when talking about AI. And it's interesting how you put and I think you did it mindfully and with intention, Eric, the way you said it, is that we have to talk about outcomes first, and then we can talk about technology last. Right? Even though it's under the umbrella of technology.

And I think that that is in fact what when we talk about AI being a superpower, I think that it's enabling the humans to have better outcomes, but you have to know what that outcome is before you start buying the technology. What do you think, Eric?

Can you talk a little bit about let's let's just dig into KRINO because it was a concept. You developed it from a lot of these conversations. Right? Let's just dig a little bit into KRINO because, like, k r I n o by the way, and he's written about Dan's written about it.

But I think what it really gets down to is you will have an accelerated like, let's say back in the olden days, you know, we opened up a Redwell, we did our document review, we unitize things, we put them in piles, we said this is this is hot doc, this is not. And now we're doing that at a monumentally different scale. I wouldn't say it's always like that was always efficient, but it did form. It helped build an argument.

Right?

Correct.

Now now AI is just compacting that and expanding that ability and that could be good, but maybe just let's just kind of underpin this quality standard that needed to come forward. Everyone's talking about governance and everyone's talking about oversight. Now those are very important things.

But tell us tell us what you think people can leverage from this concept of KRINO.

Let's take a step back about what what that word is. It's an actual word. Right? So it comes from the old Greek word which means to judge or to discern, to separate the valid from the invalid.

So why choose that word? Because it's very apropos because that's what we need to do when we are putting the inputs in to any sort of GenAI platform, whether it's eDiscovery, contract life cycle management, or or DMS system, a document management system, whatever it might be, or just a general, you know, frontier program, we have to be able to leverage our knowledge, skills, and expertise, as they say, the SKE. Right. Right?

In order to be able to discern the valid from the invalid. And so, in order to give voice to this idea, I developed, this framework, which is basically off of the word KRINO.

Well, I helped you.

Now, wait a minute.

Alright. Fair enough.

Fair enough.

Fair did. Yeah. I will I will give you credit. I have a footnote in my presentation that yeah.

Thank you, Eric. That's the footnote is what it says. So so CRINO is a CRINO is an acronym. So it makes it easy to remember because it's a word that we don't say in our daily conversations.

So it turned into an acronym. So the k stands for no. What does that mean? You have to know, obviously, what the subject matter of whatever it is that you're working on is about.

You have to know context. So we talk about context a lot lately because of AI, because the more context we give, the more we know, and the more context we give, the better the outputs can be. So you have to know. Right?

Then you have to r, review.

And so you have to review the outputs against your knowledge, skills, and expertise to know whether or not what you are reviewing is as you had called it AI slop. Right?

But then the I is the most important part.

And the reason why eye is the most important part is because we do this all the time, Eric. You have to interrogate the outputs. You have to. And what does that mean to you?

Well, okay. For every single lawyer who's on a call and you've interrogated a witness, that's exactly what it means that you do to a Frontier AI tool or a tool within a software application. You have to interrogate it against your knowledge, skill, and expertise. And you have to ask and challenge it, how did you do this?

What's your logic behind this? And then I don't mean interrogate in a bad way. Just mean to interrogate. Very similar to if an associate gave you an output, gave you a work product, you would interrogate or ask questions of the associate in a positive way, not a negative way, but in a positive way about, like, how did you get here?

What was your thinking? Right? And then n stands for normalize. So normalize the output to not copy and paste.

How many times did we talk about this? Right? And the case law supports this as well. Don't just copy and paste.

So normalize the output against your own judgment. And then finally, the o is as important as interrogate. O means to own. Own the data.

Right? Own the output. Because at the end of the day and how many times did we talk about this in the first half of twenty twenty six, Eric, with all of our guests?

As an attorney, you are an officer of the court. When you look at it through that lens as beings we don't often say this. We know you're a juris doctor, you're an esquire, you're an attorney, you're a lawyer, whatever. We don't often say that we're an officer of the court, and it's so important to remember that. We are an officer of the court. Rule one talks about a speedy and just fair disposition of every single matter.

So that you have to own it. Judge Kim has said before that if you cannot tell a judge or any tribunal in plain English what the AI has done, then you do not know what AI has done.

Right? I mean, so anyway, so there is KRINO and I've yes. I've been evangelizing the KRINO framework as much as I possibly can because I think it's just so important for as many lawyers as possible to really understand and paralegals because I've I've evangelized this to paralegals as well.

You know, it kinda goes back to what we heard from Hannah and what we heard from Kim.

The asset, the durable asset, it's not the thing, it's not the tool. It's the person. Everyone kind of is stuck on this human in the loop, but it's beyond that.

It's really And you know I don't like that phrase.

I know. But your philosophy that should be a practice is it's you need to find if you're gonna have a workforce strategy to use AI, the first and most important thing is you gotta have people who can discern, who can adopt, who can interrogate, who can revise and shape the ball of clay and put it into make it into a gummy. You have to supervise it and you have to apply it. And then, like, you and I know all the time, like, when we work with clients, it's not what they want.

Sometimes it's something they want some something very different. It's like, well, we thought based on what you wanted, the outcome was this. And they're like, yeah. But now we have this expeditious way to go back and change things.

So now back to kinda Kim, AI is the superpower framing, which is really cool, but it only works if you're super precise. And I mean super precise. Like, you can't just it's not Google plus. Right?

Correct.

If you have, like, superhuman strength and you use it in bad ways, well, if you don't have judgment, that's bad. I mean, if you don't have supervision, that's risky. And in a legal context, it's horrible. So training can't be a strategy.

It's gotta be a substrate practice. You have to be able to stand by people and tell them, like, what's the most fundamental thing that we're talking about? Did you put the right data into the model, and did you ask it the right question, and did we agree on the right question and that's we're hearing a lot about prompts being discoverable and we talk about that, but the but the idea is when you and I have challenged each other a lot about what what did you actually interrogate? How did you do it?

And the oh oh, hey, Daniel. The context is slightly different than what I think this outcome needs to be, and you'll go back and revise a prompt. Right? So I think we all are needing to wake up too.

So when I am engaging in a tool here, I'm I'm not engaging with a person. I'm not engaging with Daniel about what do you think the outcome of this should be? I'm I'm oftentimes coming up with an outcome before I even talk to you. And so I'm interrogating a tool and sometimes I'm off a path that's just wrong.

Like, hey, man. That's not even close to being what you're doing what we need to do. So what I what I really come back to is we're almost using AI as a siloed ability to develop information, not knowledge, but information. And if we don't have peer reviews, critical insights just like an associate and a partner would have in a law firm or a law department, you really are missing the boat.

So let's So let's don't take this stuff for granted, in other words.

That's my biggest fear.

Well, let's pick up on that theme because that actually puts us in bucket number two over here, theme number two that I think you're gonna absolutely love talking about. It's what we heard from Tyson Ballard. It's what we heard from Stephen Dooley. It's what we also heard from James Sherer.

And and it's probably your favorite area. So I don't know how we're ever gonna get through this spot over here, Eric, because you're gonna just yeah. I know you're gonna scream from the mountaintops like I do with KRINO. But let's talk about governance before tools.

This is so important. And this came out pretty pretty much directly when you and I went to the Clock conference in Chicago, and we interviewed several very, very good guests while we were there. Kim was obviously one of them. One of the guests that we talked to at CLOC was Larissa Young.

And Larissa had said, before you buy the tech, make sure you have the processes in place. And, man, I thought I thought you were gonna, you know, just I couldn't pick your mouth back up, Eric, because you were just so excited that Larissa had said that.

And and, you know, and Tyson Ballard, he when we interviewed him, he had said, you still need to be questioning the why at the end of all of this. You know what is the problem that this process is trying to solve and what is the human need at the end of that process.

In other words, it's not what tool should we buy, it's what AI feature should we be turning on, but what problem are we solving for whom? And even Stephen Dooley talked about this. He's like, as opposed to following the hype cycle, let's really kind of get in and unpack it. And and I love this.

So what did everybody really say between Tyson, Larissa? James also talked about this in some matter as well. He loves the idea of talking about, you know, governance. We really wanna talk about this through the context of the AI is important.

We have to focus in on what we always focus in on as consultants. Start with one zero one. What is the problem that we're really trying to solve for? Because if the answer is, I just wanna be able to be competitive with the other law firm or the other legal department that has decided to operationalize AI. That's not the right answer. Right? That is not going to lead to full scale adoption.

And we know that there's a hidden cost. There's lots of hidden costs in AI, actually, in GenAI. And this is one of them. It is the governance.

So okay. I'll let you have the mic here now. Talk to us about about the governance piece of this because I don't you could disagree with me. I don't think you will.

But, Eric, is the governance of AI any different than the governance of any other software as a service application that is being procured by a law firm or legal department.

If you wanted to, like, simplify it, I would say, quote, it's process, silly. You know, that's really the, you know, the whole that's the ball of wax. But as I said before, process is a dirty word in legal. Right?

So But why up here? Unpack that a little bit, Eric. I mean, you say that, but it's important to unpack.

The historical excuse has been, well, I'm dealing with all these different part I have my own team and we can define our process, but then I have to rely upon third parties other, you know, maybe plaintiff's counsel, maybe other supporting counsel. I've gotta deal with a court docket. Gotta deal with a judge or a magistrate and they all have their own things that they're going on. So we can't really unify process here, right?

Well, that's nonsense. You can define the process and what's the biggest enemy of process variation. But we know that in legal variations kind of commonplace. We but we have to be able to identify process variation.

And that could come in simple things like, oh, well, we had a change in a calendar or or we had a change in the way in which we decided to bring different types of people, custodians, datasets into an argument. Okay. Well, that's part of the organic nature in which we develop arguments. But if you don't define process, technology is not gonna serve you well.

It's it usually amplifies the problems.

Okay. So so that was a really important point. So so double down on that. If you don't have good processes in place and you and then you procure AI solutions, it exposes or amplifies what was already not working well. Explain that because that's such an important point.

It exposes and amplifies what already is not working well or it offers alternatives against standards. So let's just talk about the first one. When you expose something from a poorly drafted process and technology can really show you that, hey. This is not really the right way to do it.

You're not including all the data sets. You're not building the correct argument. So that that's something that I think AI is, like, shining a light on right now. It's like, there's law firms out there that are buying AI tools and giving them to associates, and they're using it for discovery.

Like, bad idea unless you have very standard disciplines in place to say, here's the acceptable so we talk about governance. What is governance, really? It's underpinned by acceptable use. Do this in this way with this dataset and don't use it for, you know, this don't use it in these ways.

Well, that that takes away the legal creative process. Right?

But I think that as we mature, we will see that, okay, as we define these process and guardrails that it's not gonna be a big deal to say, okay, in this use case due diligence, we're gonna use AI to do these seven things. Great. In this use case, we're not gonna use AI to do these other three things because it's bad. Right?

Or it doesn't give you the right outcome. So let's talk about variation and why variation is so helpful to us. Right? So I'm I have six sigma in my background and, you know, variation comes forward with, you know, this Drucker and these ideas that, hey.

Listen. If you wanna you wanna sew a shirt, you can't really have a whole lot of variation in the line worker who's, you know, showing sewing a scene. Oh, that makes sense. But that doesn't take a lot of thinking.

Right? That doesn't like, sometimes if you talk to a lawyer, they'll say, yeah. But you know what? I just found out some new facts and ideas and I have some new insights and, oh, we uncovered some some different ideas that we wanna basically build an argument from.

That should be allowed.

But within a context, right, you don't wanna just say, hey. I just found all these new things. Tell me what to think. You have to go back to the KRINO idea that says, we're incorporating new datasets that does include variation from what we thought about from our original process.

Tell me what you think I should do with this. Then you bring your peers, you bring your partners, you bring your other people in to say, we incorporated these new data sets. What do you all think? Get a human in the loop, sorry.

Get a KRINO process in place Exactly. That can can really interrogate whether or not that that variation is acceptable or not and move on. But the most important thing here and that we don't do when you talk about legal processes, we don't have measurements. We we don't have really good hard and our friend Maury, you know, Maury Kabri, he writes all about, you know, KPIs and you're a big OKR.

Right? His episode will be released in the second half of twenty twenty six. Correct.

But but what what's so harmful about measures? Like, okay. The idea that we missed the goal, we missed the target, we missed the timeline, we over budget. Those are good ways to learn.

Those are those are things that are critical to building better hardened processes. Now the last thing I'll say, Daniel, is that when you institute standard processes, and I think Kim pounded on this, is that if if you don't have training and awareness and persistent evangelization, like, remember, like, some of the things that we heard last year were like, oh, well, we used to buy matter management systems or we used to buy doc management systems and we would turn them on and everybody would just use them. With AI, there's constant learning and continuous improvement and managing process variation and how those things are are utilized.

Those are that's a very different way of thinking about how we use technology because we're in a place now where well, may maybe years ago, well, amount of management system is not changing too much.

Maybe they have a new bell whistle or a new configuration, but it's not changing dramatically. Well, AI is introducing these compounded changes to us daily. And that is something that, like, the human mind I mean, maybe this is just too crazy, but the human mind is having a hard time adjusting to, oh, wow. There's a new way I can use a notebook, or there's a new way I can use this MCP, Model Context Protocol.

I can join data together. I was never able to do that before. So what does that do to my process? It changes it dramatically.

But it's a constant continuous improvement, but you have to be able to create measurements and control frameworks that say, cool. We're gonna change in this direction now because we've hardened it. We've had a KRINO process, and we think this is a better way of working. Very difficult for us to grasp in legal.

It it is very difficult. And then I wanna throw another layer on top of that, which actually brings us over to theme three, which is, you know, data is a risk layer that goes underneath everything that we do. We heard this from Hannah for sure. Right?

So AI gets a lot of attention. And I think that what you're talking about just now is you're talking about systems level thinking. You and I talk a lot about this as well, Eric. It's systems level thinking about processes, about measurable outcomes, and look, that's hard work to do.

That's not easy. So so the concepts that you're talking about are very important, but we also heard a theme that absolutely came out from a few guests, including one of my favorite people, Debbie Reynolds, who I've known for so long.

It it was this data and it being a risk layer. So let's talk about that a little bit more. So Hannah gave us a very global version of this point. She discussed Cross border privacy frameworks, and she said there's a lot of differences for sure, but we can find common themes.

Now, yeah, that that sounds easy to say, but it's a really important point that I think we should take some time to dig into because our guests in the first half of twenty twenty six talked about it. So organizations are already dealing with things like GDPR, CCPA, China's PIPL, data localization requirements, sector specific rules. So we've got cybersecurity issues in here. And now now there's AI governance expectations, and all of the details are very, very different on everything we just said.

But there is a common theme that keeps showing up, which is that there's notice, there's consent, minimalization is very, very important, purpose limitation, transfer restrictions, security accountability. I mean, gosh, Eric, I mean, the list goes on here. This seems like this is a big challenge.

You know, you and I have talked about this a lot. And when you compare the privacy constructs are have really forced us to become hardened in our in our standards. Right? Because those are rules that are written like, hey. If you get a subject request based on the reg, here's the things that you have to do based on those subjects ask.

And discovery, it really hasn't ever evolved that way because it's highly it's far more discretionary. Now there are, you know, rules of civil procedure, but those are highly interpretable. Right? But I think the most important thing that I come back to is if you distill privacy, cyber, legal apps, eDiscovery, data governance, you pick your construct or let's just pick on departments. If if you were to walk into any corporate American situation and you talk to their back office functions, I'm not talking about their commercial functions.

You would ask them, hey, what are the assets? What are the materials that you create? Tell tell me what those things actually are. Let's define them.

Well, they would all say, well, we have to know what our content is. We have to know where it's located. We have to know how to extract it and maybe conform it to a specific, you know, model. So maybe that's a data map.

So privacy, cyber, all these teams map information. All these teams analyze information.

We're seeing a convergence. I'm seeing a convergence of how all these practices are coming back to common tools and themes.

Now what does that do to corporate America? What it should say is, hey, listen, the privacy folks over there, they built a really cool data map. Or let's say the discovery people, they have an unstructured data map that's for prep and response for a case. Great.

You know, the these teams have similar ideas around data. And a lot of it, Daniel, isn't just about identifying the data, but it's also about how to access it. Because it's not always about in the file shares. Sometimes it's in a database.

Sometimes it's in a more difficult place to actually extract. And we know that we've seen burden arguments and reasonableness arguments come way farther along than the first publication of the FRCP. And that's good for everybody. Right?

We we don't want everybody to boil the ocean. Hate that. But but what AI is doing to us now is it's kind of flipping those things, flipping that capability. We can ingest more data.

We can do more stuff. Well, remember, when you're talking about outcomes, you have to have a control framework that says we're not gonna look at everybody's email in the company. We're not gonna look at every bit and bite in a in a database. We're going to have a human that defines those things first.

But that requires governance.

Saying that requires governance and accountability.

You know, I want I just wanna ask you a question about process. How open do you think people are to writing down processes? Another common tool would be like a RACI or an accountability matrix. How willing you think people are to get to that level of standardization in the work that they do?

Well, it depends on what people you're talking about. What role are you talking about? Right? Yeah.

So Yeah. If you're talking about a trial lawyer, I think that a trial lawyer is gonna be much more focused on the client that they're serving and the outcomes that that client wants in order to solve that particular matter. And I think that is ongoing. I think the paralegal's outcome is to make sure that they're doing the work for the attorneys.

And so To I think a paralegal is more willing to be able to sit down to document processes so at least that that person knows what those processes are in service of the attorneys. But then I think the litigation support director and I think the legal operations professionals are far more adept adapted to this idea of sitting down writing processes. But there's an underlying point to your question here, which is that, and it's something I talk about a lot, which is sometimes you have to slow down to get faster. Right.

And I feel like sometimes what we've heard from our guests as well is that the speed in which we are going is that we're not slowing down enough to be able to write down the processes so that we can get faster and become more effective and achieve the outcomes that we're trying to measure.

It's such an important point. And for those listening, let's just have a counter argument to this. Right? It's Friday night at five o'clock and I got a call from a lawyer that says we have to do a production on Saturday.

Oh, well, you're not following the process. Well, good luck with that. Right? So like we again, process variation is really important.

We have to be able to flex. No one likes that five o'clock call for a production. But that's the business. Like, that's the other thing that I think that we really need to pound upon.

You shouldn't be so disorganized that you can just call somebody up on the front end and say, give me a production for a couple hundred documents. That's not really that's not good for anybody. But process is supported by SLAs, service level agreements. Right?

So tell me about how you think process and let let's keep talking about AI governance because I think it's really important. How do you think process and these service levels help the Apex teams? Like, do they care about SLAs or do they say, hey. Listen.

I gotta get something done. I got a plaintiff who's, you know, in my in my face about something. How how will the market change so much if we define processes to say, no, but it takes us two days to do a production. Sorry.

We can't do it until Monday. Or will we get or will people just lose jobs and get fired because they're not they're not really kind of appeasing the Apex teams?

So it's an interesting question.

And and I don't think that we had a guest that really honed in on this as much, but I do think Debbie Reynolds got closer to Yes. Your question.

And and that is this, is that you need to have processes period. That's how a system works. And every system is defined by the elements upon which the system actually operates. And all the elements get defined by the people who are running the system.

And so when we think through systems, you've got as you and I both know and those that are system based thinkers, you've got stocks and flows. So okay. So let's think about stocks as being, you know, it's a clothing retail store is the easiest analogy. So you've got a lot of clothes in the back room, that's the stock.

Right? And it's gotta flow into the showroom where all the clothes are, where the mannequins are wearing the clothes that people are gonna buy them. And then all of those clothes can sit there for a long period of time. Alright?

And so then you've got another stock issue. Right? You gotta constraint no one's buying them. You don't have enough foot traffic coming to the store.

The flow is from backroom to the showroom to leaving out of the door. K? So there's a process by which that happens. Right?

Those are the elements of the system. Right? The you know, you've got the clothing, you've got the merchandise, you've got, you know, the people that are in there, etcetera, etcetera. What we have to do is define the elements of the system first.

And you have to have the processes in place and you have to understand what is the what is the outcome that I wanna be able to achieve. Right? And once you define all those things and that usually gets defined at the beginning of any business. Right?

Or in the beginning of any business transformation, I should say.

When that gets defined correctly, Eric, then you have a system in place. But a system should not be so rigid so as to not permit the flexibility or the agility to be able to walk outside of that process on five o'clock on a Friday afternoon Right. But be able to still follow a process that has been created even though it's Friday at five o'clock. So you have to have a system, but you can't be so rigid in that system that there's no flexibility for the system to be able to grow a little bit.

And and, you know, when you talk about data governance, it's also a system. Okay? And so Debbie Reynolds talked about this in a way I think everyone could understand, which was I use the clothing analogy. She uses food.

Right? And I love this. And she says, in order to get your dessert, you want the dessert. Right?

But you have to eat your vegetables first. And she called the dessert being AI, and the vegetables are record management, data classification, retention, deletion, access controls, privacy impact assessment, data maps, and defensible disposition. Right? Like and she rambled that list off pretty fast.

Yeah. Right? Yeah.

Yeah. So but she said the dessert is AI. And I and I think going back to your question is like, look, the vegetables are important. Defining the elements of your system are important.

Defining what processes are are important. Nobody wants to do that work really. Right? But the dessert is either AI or the dessert is the incredible outcome that you're going to achieve on behalf of your clients.

Does that make sense, Eric?

Yeah. You can't have you can't have your pudding if you don't eat your meat. Exactly.

Exactly. But but here's here's the thing. We're talking about these underpinning vegetable components as Debbie's called them. But what about what about the product?

Like, you can have the best processes in the world, and if you don't have a good product and this is where I think the AI platforms can actually increase value, but also risk. And in this way, I'm thinking products could be some could be a shirt, it could be the data, it could be, you know, your qualitative insights, your quantitative insights. But if you don't know what you're trying to actually deliver against the product, how do you figure out like how much energy you want to put into process definition? Because you know that in startups the last thing that they want to do is eat vegetables.

Right? They just want to they just want to go blingy blingy blingy. Right? They they wanna get all the sugar in the pudding.

So tell me a little bit about how you think that is going to change with AI. Like, is it are we gonna be able to have startups that understand some of these governance structures? Because I'll tell you what, if they don't, they're in a lot of trouble. And this is the first time in history I think that if you don't have standards of acceptable use, governance, outcomes, KRINO, it doesn't matter if you're a small little shop or you're a big time startup.

You are going to find yourself in deep doo doo. So tell me what your thoughts are about how vegetable eating needs to come to the desktop for small guys, for small companies, well as other companies.

Yeah. So so what happens, Eric, when you eat a sugar filled dessert that tastes really yummy and you eat that cake, eat that cookie, whatever it is, you eat that dessert that's sugar filled. Right? What happens?

You get this high burst of energy. Right? Why? Because that's what it is. Sugar.

Right?

Sugar I'm happy.

I like it.

Right. And then what happens in a couple of hours? You crash. Why do you crash?

Right? I'm tired. Right? You're tired. Right? And why does all that happen? Right? That's all neuroscience.

Right? That that's your system inside of you actually working. Right? The way it should be working.

It's great for like for a short burst accelerated burst of excitement excitement and stimuli, and that's exactly what the point of all the sugar in the processed foods are. K? But what happens when you eat vegetables or you eat the carbs, you eat, you know, the protein? Right?

When you eat the right macros of and not to get into, you know, health, you know, science here, but when you eat the right macros of protein, carbs, and fat, and you eat the right ratios of it, what happens in a couple hours versus eating the sugar? What happens is that you have much more sustainable energy. Right? And that sustainable energy is that you're satiated.

You're filled with the right fuel to allow you to continue to think correctly.

When you're off on your ratios, by the way, maybe you have, you know, too much carbs versus, you know, maybe too much fat versus protein, what happens is that you feel like you wanna go to sleep. Right? So Yeah. The so the point of all of this and talking about this analogy of food is this, you talked about the startup.

The startup doesn't wanna sit down and work on the processes. They're sugar filled. Right? Like, yes, we have a company.

Yes, we wanna get going. Like, let's like, we are fire powered. Right? But without an underlying system or the underlying processes, if you don't have that in place, then it's just gonna be that short burst.

It's not scalable. It's not sustainable, and it will crash because you don't have the right processes involved. And I do think that that's where AI can in fact help where somebody may not be so focused solely on the implications of what happens if we don't have a long term strategy in place here. We might think we do, but AI can really challenge that.

And that's where I feel like, you know, either, you know, the KRINO methodology comes into play or some other context focused prompts would come into place as well.

I'm thinking a lot about some of the things that we heard from Brendan Brendan Sullivan. He talked a little bit about data data quality. Right? And, you know, we know that, you know, legal departments, law firms typically have just massive diaspores of data.

It's hard to aggregate, you know, in the context of discovery, what discovery people have done very well is that they've been able to triage data. Right? And then they go through a lot of junk and the whole system, actually, the whole industry was built on data triage. But when you when you think about the vegetables, you know, and building the vegetables and maybe even building the commercial products, the sugary stuff, tell tell me a little bit about where you think AI can can solve for the dirty data problems because I'll give you my thought here.

Because when you think about this vegetables, this this nutritional thing well, the first thing that we think about in information governance is, well, you gotta have good clean data. The first thing we think about in managing any system from a migration standpoint, from a running the day to day shop is, well, you're reporting on bad data. I'm working with a client right now, and they don't have really good data. They have three sets of rates that are different from the same outside counsel.

They don't even know what to build the outside counsel. So tell me about how good clean data based on what we learned from from Brendan is like a really important point here, whether or not we're talking about AI or something else. But let's talk about, like, the quality of data and why AI can maybe help us push forward with that qualitative, you know, lens.

Yeah. Brendan Sullivan, made such a great point. So he comes from the, you know, the legacy data side, and it is, to your point, so much more important than ever. And one of the things that Brendan told us was if we don't know what's in that legacy data, how can we confidently get rid of them or how can we confidently access them? And so what he was really talking about was, you know, a problem that nobody really wants to talk about, which is, you know, the forgotten archive problem. Right? The fact that we're storing all of this data.

And and you asked the question about AI, and I think Brendan beautifully answered that question. And he said AI is an opportunity to tackle that because, really, when I think about what we've been asked to do for the last twenty years is all risk based. That is an eDiscovery way of thinking about things, not just an AI way of thinking about things. And and it's not just, you know, hey. Can AI summarize documents? But can AI actually help us understand unknown data populations so we could reduce cost, reduce risk, and make better legal decisions. And that's really where I think, you know, Brendan was so helpful and I think it's really where he landed on that one.

So maybe maybe that idea goes back to this, you know, controlling process variation. Maybe if you have better data that's going into a system, you can control the variation. You're not gonna actually have three different rate cards and not know what card to put in or what rate to go into an invoice. So maybe maybe his ideas are really more about, hey.

Focus on cleaning up your data shop, whether you're a startup or not, but start with fundamentals of data management. And then and then, like, it all these different teams that try to leverage data to provide insights are gonna be better off. Sounds kind of Pollyanna ish because we know that with AI and heck, you can you can read the latest new discovery reports. Data is exploding and there's all these new data types.

How does one get in front of this this idea that there's too much? And maybe back to James' ideas of the cognitive miser. There's there's just too much stuff out there and when I do my little privacy routine or I do my little brief response stuff, I just go at these types of data. And, yeah, I do my cleansing, but I'm not thinking.

I'm not allowing AI to be my cognitive partner. Maybe I'm just saying, listen, man. I'm good with my process. It's cool. I've always done it this way.

It ain't broke. I'm not gonna fix it. So how do you incorporate this idea of continuous improvement into these traditionally siloed teams to get them more unified to not just develop a better outcome, but to use the same good data over and over again?

Like, that's gotta be really tough, I think.

So I think what you just hit on, and I think Brendan Sullivan hit on this point as well, I think that's the gold mine, no pun intended, but that's the gold mine opportunity right now. It really is because what you're talking about is two things simultaneously, Eric. On the on the one hand, what you're talking about is you've got all these outputs that AI gives you. So you're creating much more data, especially on the enterprise layer.

You talked about the hundred page output report that an AI tool created. Right? That that created a lot of data on an enterprise level. Right?

Sure. That create a lot of data. This recording that we are doing right now for our BDO Legal Tech Talk podcast creates a lot of data. But what we also have is all of the historical data.

That was the stuff that Brendan was talking about. So Right. So if you combine what Brendan was talking about with what Debbie was talking about, with what James was talking about, all of a sudden you how you see where AI plays a different role than just asking for an output. Right?

So you've got data that's discovery. You've got data that's privacy related. You got data that is records related. You've got data that's security related.

And all of a sudden, you could use AI as the through line with all of those various different functions. And whereas we have talked about, quote, unquote, cross functional behaviors throughout an organization, now we have the chance with AI to actually have those functions truly one hundred percent talk to one another. And that, I think, actually lends itself well into this idea that I think that the economics of legal work are really being rewritten, and that leads us to our our fourth theme here, Eric. And that is what is happening in the economics of legal work.

And this is where we had a really interesting conversation that came up in several different angles.

And one of those angles that came up was from Lee Minkoff. And he brought this private equity and services investment perspective that I hadn't heard of. Scott Mozarsky brought the investment banking, legal tech, capital markets perspective.

Kevin from Husch Blackwell, he brought the law firm operations perspective. Tyson, who we talked about before, brought the legal operations and entrepreneurial perspective. Stephen Dooley over at Selvyn Cromwell brought the law firm technology implementation perspective. And I think what ties all of those together you tell me if you disagree on this point, but I think what AI brought together, and it's a good segue from what we're just talking about, is that AI is not just a productivity and efficiency story. There's a business model story here, Eric.

It's so complicated to kinda distill all these different brilliant people that we spoke to, but it's the models that we are going to revisit and we talk you know, Lee and Scott talked about this managed service outsourcing.

And what that is essentially allows law firms to be lawyers and export essentially outsource all of the other functions like HR, finance, tax to a third party provider. Now I think it's really interesting because when you talk about the new the new models that are out there, what's gonna be blown and we'll we'll talk about this a little later, but what's gonna be blown away is the leverage model.

And the leverage model is billable hours, and that's partner profitability. And when you when you start thinking about law firms being challenged to only offer up senior level people, and then you hear all this clamoring about why am I spending three thousand dollars an hour? Well, your frequency and intensity of hours from that individual resource have gone down. Right?

Because the law department is trying to do more. We've seen plenty of reports from Thompson and Walters that say, hey. Listen. People are using AI to do a lot more and now there's metrics coming up.

You know, they're they're seeing a fifty percent, let's say, sixty percent reduction in the amount of time. Not necessarily cost because remember, we're also seeing this other thing like partner profitability is way up. Right? So we are at an inflection point right now where one of those is wrong.

And I think what's wrong is law firms are still riding the coattails of companies that are just interested in having an outcome regardless of the rate or the cost. But now we are seeing clients ask us, what's my staff allocation ratio from lawyers per revenue per billion? How many how many legal ops people should support my my legal function? How many powers do I need to have in house?

How can I bring more in house? So I think that, you know, we have to let this roll a little bit, but the idea that, you know, the models are gonna stay the same is hooey. The models are gonna be vastly changed. There's always gonna be the client partner relationship that will never go away, but that relationship will change.

And we're seeing tons of pressure from law departments on law firms to say, I'm gonna use this tool and you're gonna give me a qualified analysis based on the work that I've assembled from my own information. So what we talked about this a lot, but when you when you hear this idea about, oh, we wanna do more inside. What does that mean? Well, we wanna hire people.

That's increasing your your cost, your operational overhead cost.

But ultimately what it's doing is it's keeping institutional knowledge in place. And traditionally speaking, many big companies will just rely on Mary, the lawyer at this big law firm who actually has worked with this company for twenty five years to provide the excellent advice and guidance that they need for a merger. Well, what if you started to bring that in house? And I think that people are starting to get this idea that, hey.

We don't need lots more people because we can use AI to aggregate and build the Sullivan comment, is build better data quality and maybe start thinking about how we change the models around fee arrangements. Now is that happening right now every day? No. But we're seeing plenty of law firms out there that are essentially just expert led firms, and Pearson Ferdinand is an example of that.

And what the idea there is That was the guest that we had on the show last year.

Yeah.

Yeah. Yeah. Right. But what we're seeing is, like, there's twelve firms out there that are just having partners work with clients, and they're outsourcing everything else through the MSO. But the idea is, like, yes.

The idea of experts coming forward to play specific roles, that's a model change. And and again, it might be, I'm gonna pay for an outcome and I'm gonna pay you a fixed rate for an outcome. I'm not gonna pay you a billable hour until I think the outcome's good enough.

It's interesting. So our guest, Scott Mozarsky, he he framed this as a AI as being a much broader disruption. He said that we all tend to over index partially based on news cycle towards technical disruption, whether that disruption is AI now or whether it was machine learning when technology assisted review became a thing over a decade ago. And, Erica, I think that's a very useful reminder because AI is very massive, but legal has seen disruption before.

This is obviously the biggest disruption. But I think the difference now is that AI is intersecting with capital. It's intersecting with ownership models, access to justice, legal technology services, and Yeah. Also law firm economics all simultaneously.

And I think that when you look at it that way, you know, Scott's other point that he said was is like, look, if the clients are protected, you know, everybody really wins. And I think we started seeing that with some of Tyson Ballard's comments as well and also with Kevin and Stephen Dooley's comments as well. And I think what we're starting to see here is that as AI continues to grow, I think what one of the things that you identified that none of our guests really discussed as much, but there's a competence trap here. And that competence trap here is like, if you've always been defined by the billable hour, that time and materials, and AI sort of erodes that in some fashion.

Right? Then, you know, what do you have left if you're no longer defined by, oh, this is this is an investigation. This is a litigation. I know it's gonna take me this many hours, and it's gonna cost this much.

Well, I need this many, you know, attorneys, the paralegals, etcetera. I need this many outside consultants as experts, and this is what exactly what it's gonna cost. I've done this a thousand times. Here's the cost.

Right? But now what we're saying is, wait. Hold on a second. We could streamline those hours down.

And so now it's like, well, wait a second. Why should I then use AI? Right? So the economics of legal work are being rewritten here because one of the things that we talked a lot about with Lee, and we talked about it with Scott as well, as well as other folks that will be on in the latter half of twenty twenty six, is this idea that if everything else is changing with law firms and the economics of how law firm is being done, then why don't we talk more about those MSOs?

Right? Why don't we talk more about what it means to really run a law firm business now? It just seems like what AI has done based on the feedback that we got from all these prestigious guests on the show in the first half of twenty twenty six is that AI is not just disrupting the work that we do, but it's disrupting the economics of what we do, which is gonna disrupt maybe even the ownership models. Did you get the same impression, Eric?

And that gets back to this commercialization of the practice of law. Right? Big deal. Right? Oh my god. You're never gonna commodit you're never gonna standardize my process. You're never gonna commoditize my thinking.

Well, okay. I get I get that. If you're like a high high stakes m and a lawyer, yeah, you're probably right because you've worked with this private equity for thirty years and you know the ins and outs of this transaction based on these rules. Cool.

But there's so much other stuff And I think the MSO model has pushed a lot of these new new firms into this idea that, hey, man. We can actually commoditize a lot. We don't need to have associates and a leverage model to support clients. And I'll tell you one thing that we heard from all the private equity folks that we talked about is, and you and I pound on this a lot, which is it's all about protecting the client and it's all about building the relationship.

I had to ask you this because how are clients who have this trusted relationship going to react to a company, a firm that's using AI and not having a Creano process in place? Like you worked so hard for your relationship for the last ten years and all of a sudden you throw AI slop out there and a court we've seen all thousands. And a court says, hey, this isn't this is a false citation. What what do you think that's gonna do to a client relationship?

Oh, I mean I mean, it's a softball question because it's gonna hurt it's gonna hurt a lot. And and by the way, that what you just said, and I don't we won't dig too much into the hallucinations, but it is not something that just a pro se litigant or a small law firm gets in trouble for.

We've seen the headlines and law one hundred firms Yeah.

Are getting in trouble for this as well. And AI slop is another way of saying it's non supervision. Right? It's non KRINO thinking.

So the way to get around that is to really embrace that more. But let's dial back into this idea that we have, and Stefan Dooley said this, but in a different context is that AI is creating a multidimensional output. Right? And I think that we have this multidimensional output here with the ownership of law firms in a different way.

And and the over twenty years that I've been doing this, over twenty years that you've been doing this, I've never heard this conversation more loud and more well articulated than now. And I think that it's because we've hit this pivotal point where we're starting to see this intersection.

This convergence is probably a better word. This convergence of AI is disrupting how we do legal work. AI is now disrupting the ownership of a law firm because it's disrupting the way that we build the economics of a law firm. That's different. And this is the first time that we've had folks on the podcast that really have been so articulate about those points. Lee, in particular, was very, very passionate about this topic.

Lee's coolest comment was he wants to spend time with really good people. But, you know, it kinda rounds us out to this fifth theme, which which is really about human judgment, trust, and the evolving role of the legal technologist. I think what's important for us to talk about is there are really good capabilities that exist from legacy practices. Let's just pick on Records management, you know, those teams define the catalog of records and information that need to be retained based on some duration, based on some legal jurisdiction and citation. Great.

Those people can come forward today, and those people can be your content cultivation. I've worked with lots of companies and the records management people got pulled into this eDiscovery thing. Tell me where all your accounting data is. Well, now that question's different. Those questions can still be asked and answered. But now can say, hey, tell me where all the accounting data is that maps to this specific department, that maps to these seven people that I can put into a rag and I can interrogate it. So tell me about human judgment, trust, legal tech roles, privacy people, all of these other people that are are basically coming forward now who had these really good legacy skills, but now they can piggyback on this amazing technology again with the right process and governance controls.

I think this is the most definitive opportunity that we have ever seen in all the years that we've been doing this to have this convergence of roles of people coming together and leveraging AI to redefine how we do work in a meaningful way. Yeah. And it is the information governance, it's the data governance, it's the data privacy, it is the legal operations folks, and it's the lawyers. You know, it used to be long time ago when we talked about this a couple times here, Ralph Losey came up with this concept, so I can't take credit for it back in I think it was like two thousand eleven or two thousand twelve.

He was defining the eDiscovery relationship at that time as a triangle. And Yeah. Yeah. Remember that.

Remember that. So so those that may not remember this from well over a decade ago, he said that has to work in three components in order for everything to work correctly. And so the triangle was, you have the lawyer who knows the law, but may not know the technology and may not know the business of law. Then you've got the litigation support professional who understands the technology, but may not understand the law, may not understand the business.

And then you've got the c level folks. And the c level folks are the ones that understand the business, but may not understand the technology, may not understand the practice of law. And when you have those three things working together, he's like, that's when you have, like, that's an eDiscovery team was what Ralph's point was.

And and I love that. And thank you to Ralph because I've with his credit, I've been using that in a lot of CLE presentations I've given over the last decade. And so when we talk about where we are with human judgment, trust, and the legal technologist role, I think that AI has given us a different conversation to have.

And it is for those that know how to utilize it responsibly, that's an important point, who know how to explain it, who know how to govern it, who know how to connect the legal and the business and the technology problems together.

This is such a beautiful opportunity, Eric. And I think that our guests that we have had on the program so far this year have really talked about this. Debbie made the accountability point as clearly as anyone. She said, you can't trust these tools to be accurate, and, you know, the laws are about people, and they're about companies, so the AI tools aren't gonna be put on trial.

And James also put it really clear. He said, you know, if it crosses your desk as an attorney, as a practitioner, it's your responsibility. So what do all these things together tell you? It tells you that this is an opportunity for us to really work together in a much more meaningful way than we have ever seen as a result of using AI in a very different way than just utilizing it for, hey.

I need an outcome for a discovery project.

It sounds like these bridges that are being built, there's this really cool book, legal singularity and how how AI can make law radically better. And that is essentially this idea well, Lessig's talks about it as well. The singularity is near. Right?

And in legal, we have little functions like law, data ops, risk privacy fund like, these things are all coming together. The challenge the challenge to people that are in this space is will you allow it to come together? Will you strip your antiquated function and your maybe even your job title even in maybe even in the routine ways that you work? Will you say that doesn't work anymore?

Like are you going to be a big enough brained person to say that process is busted and we need to change it and actually be an advocate for that change? So one of the things that I think we've really heard a lot from all of these folks is that the most difficult thing about these types of evolutions is change and managing change. Debbie and Kim both talked about training as the underpinning of all this stuff. And if if you don't have good training and awareness and you don't integrate this KRINO idea into the ways of working and into the minds and hearts of how people operate and use these tools, you're gonna be lost.

And I'll say that the other thing that we really don't give enough credit to is this operational sustainability. Like, you need to make sure that you're doing things in a repeatable, defensible fashion, but you also need to be able to allow for process variation because things are changing way more differently than they have. So I guess my question for you is, like, let's just talk a little bit about how does change come forward in this organizational structure that has been built on these monolithic routines, you know, this guy reports to Joe and Joe reports to Mary.

Do how do these organizations, particularly in these new law firms that are operating very flat, know, essentially, client partner relationships. How do those things impact the way a traditional legal department or a law firm operates in the context of change? What is going to happen if you don't manage this idea of change?

I think that everybody who understands how to appropriately utilize AI to the benefit of their clients and to the benefit of their business is starting to see this convergence and apply this convergence. And they're the ones I think that are going to be much more innovative, and I think those are the ones that are gonna be more successful and stand out. Remember something that you said earlier on this episode that we've been recording and also came up with one of our guests.

There's a difference between information and knowledge. So Information is the output that we get from AI. Knowledge is what you understand because of your knowledge, skills, and expertise of what you understand and what you can discern out of that information that it gives you. Why is that important to talk about here? Because when you understand the knowledge that the information is providing for you, then you are able to leverage that through the context of I see the convergence and I see how I could differentiate. If everybody has the exact same information and everyone has the exact same LLMs, then what happens here is that information is in abundance and knowledge is in scarcity.

Yeah.

And the differentiator is gonna be those that can capitalize upon that knowledge. And if you could do that right, then I think that one of the major themes that came out in first half of twenty twenty six and what what they taught us is then you can have outcomes. And look, I think that if anybody taught us a beautiful outcome of any guest we had so far in twenty twenty six was Larissa Young who works at the hospital. In the realm of talking about AI and outcomes, what's the outcome that they're at the hospital? She's trying to save kids' lives.

Yeah. Very human. Right? Purely human.

Like, let's just come back to that for a second.

Right? And it's such a great way to round out this whole episode. It's how we started this episode out. And as a matter of fact, we started this episode talking about Larissa. Right? But, you know, when I think about it in the context of when you work in a hospital, you really are in the business of saving lives.

That's the outcome. The system in place, the elements by which it defines the system, is everyone that's working there, the technology and the processing tools. You know, gosh, that's so cliche. We talk about people process tech.

There's no better time to talk about people process tech now, Eric. Right? And for all of our listeners, you know, when we talk about what's the outcome, I look back at, you know, Larissa and what she is doing and the outcomes that she's trying to leverage in creating transformation and leveraging AI in a hospital to help save people's lives, that's the most human thing you could possibly come up with. Right?

Yeah. And I think that if we can strip away everything else we talk about, well, we have a flat organization and Mary reports to Sally, etcetera, etcetera. Let's just talk about it a little bit differently. And I think AI gives us this opportunity for this open space to slow it down so that we can go fast enough to have a different conversation.

And if there's any lesson that I've learned at all from all of our guests so far in the first half of this year is that we're having a different conversation, and we have to be open to a different conversation because, again, information is in abundance, knowledge is in scarcity, and those that can capitalize upon the knowledge are the ones that are gonna be most successful, Eric. And I think that Yeah. As we go into the end of twenty twenty six with our other guests, I think we're gonna start seeing some of that more.

Yeah. For those of you who are listening that are maybe just getting into AI, embrace it. It's here to stay. The message is not to not to avoid it.

The message is to be have a critical eye of it and to mature around it. So but to me, that maturity has a set of constructs. You gotta understand how AI affects the professional development. And I'm talking about the human development too.

Right? It's it's the psyche. It's the person that's trying to move up a ladder or develop a better skill. Like, AI can have a tremendous impact both positive and negative, particularly when you hear what my friends at PNC, Emily, has to say about, well, we're eroding this ability to critically think because AI is taking taking all that that functionally.

The other couple ideas are She was also a guest on the podcast as well.

Yeah. Governance is key. Oversight is key. KRINO is key. Don't buy a tool. Don't even start using a tool until you have a very specific construct on how you're gonna use it in the context of an outcome.

And then as Brendan said, gotta know your data. And if you have bad data, like, that's a good thing to know, but don't start building models with bad data because it's gonna really shift the outcome and that's not gonna be very good. And then I think the future is really thinking about rethinking these models around this value, not just revenue, not revenue per partner, not partner profitability, not ours. It's what's the value that we're providing this Scott and Lee both said, it's about the relationships, silly.

It's about relationships. And if you don't have good relationships and you're in a transactional space, I'm not talking m and a, but you're just trying to make money off of a quick hop and stop with a client, you're not gonna have that client very long. It is about relationships and if you don't invest in people that can build those relationships, who can connect technologies to judgment and then better outcomes driven through value, you're gonna get beat. I think that's really, like, to me in a nutshell, what we learned.

You know, that is the cleanest place to land there, Eric.

And I think what I just heard you say, and obviously our listeners can't see this, but I was smirking as you're speaking because I think that what I just heard there is that the greater this technology is, the more human we have to Right on.

Right? So Right on. We wanna thank all of our guests that we've had so far mid year. We've recorded a whole lot more that are ready to get distributed throughout the rest of the course of this year.

We want to thank all of our guests that have helped us and helped shape us and all of our listeners. We want to thank all of you as our listeners. And if you haven't heard any of these episodes or you want to catch up on some of these because we brought up some really great guests with some really great quotes and you want to go back through the back catalog, by all means head over to bdo dot com. Go over to any of the streaming platforms that you listen to.

We are everywhere that you want to listen to a podcast. And thank you very much. Until next time, this has been another episode of BDO's Legal Tech Talk.