Robert Klamser on Using AI to Make Chapter 11 Bankruptcy Easier to Understand

Russ Fordyce 🏆 (00:01.186) Welcome back to the Winner's Circle. I'm here today with Robert from Stretto. He is tackling one of the biggest problems in the law industry, Chapter 11, mess and chaos. Robert, welcome to the Winner's Circle and congratulations on the AI Excellence Award. Robert Klamser (00:15.196) Hey Russ, thanks for having me, I really appreciate it. Russ Fordyce 🏆 (00:17.57) Why don't you give us just like an elevator pitch of Stretto, get us kind of a baseline and your role at Stretto. Robert Klamser (00:24.316) Sure. So I would say the elevator pitch of the company is Stretto provides support services technology for a variety of professionals that are involved in in bankruptcy and that runs the gamut from small individual chapter seven bankruptcies up to the largest chapter 11 bankruptcy you'd see in the news every day. What we like to say is 90 % of the bankruptcy is filed in the US are either done with our software using our forms or with one of our teams. So The unsexy behind the scene plumbing that is the bankruptcy system in the US were involved in so many different facets of that. And personally, my background is in the Chapter 11, the large bankruptcy space. I'm not a lawyer. I like to say I've learned how to talk and kind of play and act like one. I had a business background, got involved in the industry by chance. and 25-ish years later, I'm still working with the same leadership team that I started with, and now I'm in charge of all of Streato's innovation efforts are really our AI-focused new initiatives, technologies, services, et cetera. And I like to say I'm a jack of all trades, master of none, and so I get to have, you know, roll up my sleeves and get busy in a lot of different areas, anything that AI could touch within the company. Russ Fordyce 🏆 (01:47.98) Well, this is such a ripe market for AI. my opinion, the whole law industry has got to be going through a boom with this AI stuff, just the amount of documentation and the complexity of it all. But bankruptcy is one of the most complex and stressful processes in business. What does that communication kind of typically look like if somebody's not using Conductor, your AI tool? Robert Klamser (02:11.673) That's a really good question. think you nailed it on both complex and then also just volumes of paper and volumes of documentation. You think about the largest bankruptcies, the most complex businesses that have to restructure. There's thousands and thousands of pages of information that's out there. And what we found is this is a niche industry. No one goes to school and thinks, you know, I want to learn bankruptcy. I don't even think a lot of lawyers go to law school and say, I want to practice bankruptcy law. And part of where we sit is we're kind of in the center of the information between the debtors themselves, the lawyers, the court, and then most importantly, all the stakeholders who are impacted by the bankruptcies. And that can be everyone from former employees to small mom and pop vendors to large landlords who are creditors of these bankruptcies. And so what Conductor does is we built this platform that communicates bankruptcy in a more clear user-friendly way so that when you know, I'll use the example of Saks Global, the parent company of Saks Fifth Avenue and even Marcus, etc. When they file for bankruptcy, the minute these documents are filed, you can go on our platform and you can say, you what do I do? I've never been through a bankruptcy before, but I am a sweater vendor and I supply sweaters to Saks and I'm owed, you know, six figures and it's a big deal for me. What do I do now? And we built this platform, Stratoconductor, to understand the documents and answer that question in a grounded, verifiable way, but also in a way that actually answers the question for that vendor. Or another thing that it can do is you're a lawyer, you just see, hey, someone just filed a thousand page document related to a court hearing that I'm on my way to. There's There's no way I can read the entire document, much less download the document. Conductor understands it you can say, me the highlights. I'm a vendor on my way to court, describe the key points of this document that was just filed and you'll get those answers immediately. it was very purpose-built for this community because it is complicated. No one really understands it, but we do. So we designed it for us as users. Russ Fordyce 🏆 (04:34.862) Well, and I remember I was in publishing with the business journals for many years. And I remember the reporters literally going to the courthouse and writing down the names of all the bankruptcies and then bringing them back and typing them in to let everybody know. It surely isn't that bad. What's broken about how information flows in these chapter 11 cases? Robert Klamser (04:45.077) Hmm? Robert Klamser (04:52.214) Mm-hmm. Robert Klamser (04:58.901) I think that's a really good question. would say what's broken is it's complicated and it impacts a lot of people who this is not part of their daily life. If you are a like I said a vendor, a former employee of a large company. I think about the airlines or the auto parts manufacturers, auto you know the car companies themselves that have filed for bankruptcy. that is not part of understanding that process is not part of your daily life, not something you would ever you would ever come across. We like to say when we're dealing with these cases and we're dealing with our clients, we hope we do a great job and we never have to work with you again because they don't want to ever have to work with us again. And so what was really broken is that oftentimes these large bankruptcies affect a lot of people who will never have to deal with it again. And it's a really critical. point there. It can be overwhelming, can be intimidating, and so part of what Stratto does just across all the different services outside of just Conductor, this tool we built, but a lot of the service that we provide is making it understandable for everyone that's involved. Russ Fordyce 🏆 (06:16.268) Hey, if you notice your screen is a little grainy, don't worry about that. The zoom that we're on is actually low res. You're recording in the backend, so don't let that stop you. Robert Klamser (06:19.932) Mm-hmm. Robert Klamser (06:27.687) Got it. Okay. Russ Fordyce 🏆 (06:31.244) Now for people outside of legal, you know, this legal world, why is it so hard for the stakeholders to access that basic information in a case? It seems like that ought to be, you know, delivered to my inbox somehow or, or, or something. There's, know, it seems like I should be getting notifications and all of that type of stuff. Why is it so hard to get the just, just this basic information if I'm a creditor? Robert Klamser (06:54.044) So there's two answers to this and I'll give you a little bit of history and I don't even need to apologize for it but everyone who I work with knows I'm a storyteller. I can kind of go off on tangents but I promise it's relevant. So bankruptcy code, is a, bankruptcy law is a federal. It was revised back in 1978 and there were a lot of changes and one of the key changes to the bankruptcy code allowed or what it said was certain cases are too big for the courts to deal with themselves. Because like you said, it used to be physical documents at a courthouse. If you wanted to understand what was happening, you might have to go and get photocopies of the document from the court. what the revisions in 1978 basically said, you know what? For some of these cases, they're too big. And so the courts, there can be some of the work that the courts would do can be done by an agent of the And so we will we will authorize certain companies under certain restrictions to perform some of the duties of the court. That's how our core business in bankruptcy originated is there's companies like ours who can essentially become the information hub in these bankruptcy cases. And so then at least things documents could be digitized. We could create websites and really that didn't start happening until the early 2000s. And so that was great. At least then, if you're impacted by a bankruptcy, you could go to a website and you could get some documents. And that's kind of been the status quo for the last 25 years, is everyone in the space had websites. And then everyone in the space started saying, you know what? You can subscribe to notifications. We'll email you the documents. That's all great. And it was really, it was good for bankruptcy lawyers, created a lot of efficiency, but that doesn't impact the, like I said, the stakeholders who are involved in a bankruptcy for maybe the first and only time ever. It doesn't impact and it doesn't necessarily improve things for the educational community or reporters or people who are trying to do research because, yeah, that's great that I can get a notification, but what if I need to understand how this notification impacts 10 other similarly situated cases? Or what if I'm trying to do research? What if I'm trying to understand this? It didn't help you understand it, it just gave it to you. And so what we did with Conductor is we said, Robert Klamser (09:18.339) We want to understand every single bankruptcy document in the context of the federal bankruptcy code, the local rules of the individual jurisdiction that you're operating in, because course cases are administered differently in the Southern District of New York and the Southern District of Texas and the Southern District of California. They all have nuance. They all have their unique local rules. And so we said, have to understand that. asking, how do I do this? in this case to impact my rights, you might get a slightly different answer if you're in a different jurisdiction. So we said, not only do we want to understand what the document means, we have to understand what it means in the context of the federal bankruptcy code, the local rules, and then really, what does it mean in the context of that one case I'm dealing with? You know, someone objected to this. What does that mean? How does that impact the 10 documents I've already read? That's what we're doing differently. That's what Conductor does. It really understands these documents. in the context of the hierarchy of bankruptcy, the rule structure of bankruptcy. And so that, what I think, is really democratizing the process. Because now, if you are new or unfamiliar or if you're trying to understand something bigger than just one document, you can do that because we used AI to really understand the document the way a bankruptcy lawyer would, someone who's been trained for 30 years would, another bankruptcy professional would. We've taught it everything that we knew from that lens and said, okay, now answer the question. Now answer what this means. What does it mean to you? What's next? That's what we built. Russ Fordyce 🏆 (10:54.946) So walk us through the of the experience. You know, had this very complex, very manual-ish process for way too long, right? It's been 25 years now. We've just been doing websites. What exactly is Stretto Conductor and how does it change that experience? Robert Klamser (11:06.989) Mm-hmm. Robert Klamser (11:12.695) Sure. So, and again, I'll story tell a little bit, but AI was new, newer in the legal space two, three years ago. And we said, you know, part of what we do is every time there's a large case, there's a big call center of people and we try to teach them the nuances of that case. We try to teach them the rules of that case so that people can call and say, what do I do? What is this deadline? What does this mean to me? What do I do next? And so, That was kind of the old way this problem was being attacked. And we said, look, this is exactly what AI should be able to do. We should be able to teach it all the rules of bankruptcy. Anything we would teach a new associate, anything we would teach a new consultant who was going to be working in a bankruptcy case, we should be able to teach AI something. But we also, like I said, realized that there's nuance between the 94 different federal bankruptcy jurisdictions. There's nuances and there's differences in all of those jurisdictions. So you can't just say, let's take ChatGPT and run every document through ChatGPT and just then ask it questions because the context was wrong. The understanding was wrong. And so we sat in this room and said, well, we know this. We know all the rules because we have to teach it to people when they come to the company. We have to explain it to unfamiliar people when we're on the phone. So let's teach it to an advanced AI powered platform and then let the platform answer the question. So that's really what we did is we said the issue here is something, we all understand it, but it is complicated and it is voluminous. And this is exactly what AI was built for. Understanding a thousand page document in the context of the right other documents and write rules. And then on top of that, we're dealing with lawyers. It is really, you can't be wrong here. You can't, we're dealing with a legal process. The difference between something shall happen and something will happen are very different. And we had to make sure that it understood and that it understood the law and that any responses that it gives as we're interacting with these people. Robert Klamser (13:31.752) are grounded in the right documentation from the right case and the right rules. And you know that. And so we just made sure to make, or we had to make sure that everything we built was both intelligent and then grounded and then performed properly. Russ Fordyce 🏆 (13:46.766) Well, the context matters and I hear that from a lot of AI innovators is, you know, these kind of specialized platforms, I liken it to, you know, tennis. You know, if you want to play tennis, you can go to Walmart and get a tennis racket. You will not win Wimbledon or probably any tournament with a Walmart tennis racket. You need to go to a specialty store, have it strong. You want it tight to your specification. You want the grip to your specification. So, know, these purpose-built AIs are really, I think, kind of that future where you can have that proprietary intelligence and really kind of, you know, be that domain-specific AI expert. Robert Klamser (14:22.93) Mm-hmm. Robert Klamser (14:29.18) Right. And that's, yeah. I was just gonna say right, but go ahead. Russ Fordyce 🏆 (14:35.554) Yeah, well, mean, so how did you guys decide that, you know, that chat GPT wasn't going to do it and that, you know, but, but at the same time, you're using that kind of chat interface, that, you know, single thread kind of interface. Robert Klamser (14:50.024) So, let me pause for second because now the battery on my iPad is low and I need to plug it in, which means I'm have to change the orientation. Russ Fordyce 🏆 (15:00.686) We'll see what it does. Robert Klamser (15:02.872) what it does. Okay. Is it gonna fix it automatically? There we go. Russ Fordyce 🏆 (15:11.433) No, I don't think I will. Robert Klamser (15:16.807) Is it good? it kind of back to that'll work? Okay. Let me just prop it up a little bit more. All right. Okay. want to? Yep. Russ Fordyce 🏆 (15:18.018) Yeah, yeah, looks good. Yeah. Give it, give it a beat. We'll back up a little bit. Russ Fordyce 🏆 (15:32.558) I think I'll go back to my tennis analogy. Robert Klamser (15:35.738) Yeah, let's do it. Russ Fordyce 🏆 (15:37.9) So in my world, I think about these kind of purpose-built AIs almost like I think about tennis rackets. And I know that's a little weird, but you know, if you want to play tennis or go fishing or something, you can go to Walmart and get a fishing rod and you can get a tennis racket and they'll have a variety there. But if you're going to win Wimbledon or the Bass Pro Tournament, you're going to be on a specialized rod that's dialed into you, to your spec, to your grip, all of that stuff. And it seems like you're doing the same thing for the legal industry here is building this purpose-built AI. Robert Klamser (15:55.44) Mm-hmm. Robert Klamser (16:06.199) Yes, that is such a great analogy and it's something that we've seen in the last few years as we've watched the evolution of AI in the legal space specifically is that a lot of the tools are very general. And to be fair, a general tool, like a general research tool where you're just saying, You know, I'm looking for a document related to an airline parts manufacturer bankruptcy. And I want to find something that's a that's related to contracts with a parts manufacturer. AI can help find that certainly. But when you when you start to say, but what is what does it really mean? What are the specific implications? That's where the answers are too general because you really need a tool that I want to back up. I feel like I'm just going down this path of cliches that I'm not gonna like. Russ Fordyce 🏆 (17:01.848) Cut. Robert Klamser (17:02.564) I like the tennis analogy because there are a lot of very general tools in legal space, especially over the last few years. That was kind of how everybody dipped their toe in the water, was with generalized tools. And they serve decent purposes in a lot of ways. But what we found is bankruptcy requires precision and it requires precision, like I said before, between the different jurisdictions. between the nuances and the differences on a case to case, judge to judge basis. And really, the example we like to give is if you want to understand an objection that's filed in one case, or if you ask what does this objection mean in one case, and you just, all you say is, know, just want to understand this one document, this one objection. That's your kind of Walmart tennis racket. ChatGBT can say, here's what this document means. But. that's not gonna do you any good if it was filed and then it was objected to three times and amended three more times. And really the answer to your question needs to encompass the whole history of what's going on with that. Because if you wanna say, what is this objection trying to do? Saying what this one document tried to do and interpreting one document isn't gonna do you any good in the bankruptcy space when there's eight other related documents that are impacting and modifying it. And so we... We had to build something that was very specific to bankruptcy because we said, if someone asks about this objection, you have to look for these other types of documents and you have to understand how they relate to that objection. And then you have to frame your answer with all of that in context. so, and the only way we could do that is we understood the market. The technology was built by entrepreneurs and bankruptcy lawyers. and founders and technologists. It was built by people who really knew this. I multiple people with decades in this industry. And it matters because we also built it for what we would want. We're not tech entrepreneurs who built something that we thought a lawyer would want. We're bankruptcy lawyers and bankruptcy professionals. And we built something that we would want. And that matters. Russ Fordyce 🏆 (19:20.162) Well, and in your nomination, you call out some issues like naive retrieval and reasoning errors that these general AI tools kind of run into all the time. Explain those a little bit and what does that look like when you hit it? Robert Klamser (19:34.581) I think, now I'm trying to think back to the article itself and the context was, think... What we saw when we saw these more general tools in the industry is... The. If you have a general tool and it's used by a general user who doesn't know specifically what they're asking for, you're going to get a general answer. And a lot of times you don't know what's wrong. If you ask, you know, what is the deadline that this document says, it'll, it might tell you just the simple deadline that might be alluded to or laid out in one document, but you don't know that there's other context outside of what you're looking at. And so these general tools, won't give you that answer either. wasn't that we found these tools in the market had errors. It wasn't like you were saying, you know, answer this question. And the error was I can't answer that question. You would get an answer. It just, you didn't know that there was more context that you needed to consider. And so we had to make sure that when we were answering questions about these specific cases, all the context of that case is considered, even if that wasn't what you asked about, because that was the only way to get the Robert Klamser (20:56.893) the real precision that this market demands. Russ Fordyce 🏆 (21:01.262) Well, and in the legal setting, having a father who was an attorney, being critical is root to most of them, and trust is the biggest thing to all of them. How are you building a conductor that's trustworthy to these kind of skeptics? Robert Klamser (21:21.992) That's a great question. And that's where it helps that it was built by bankruptcy lawyers because everything is citable and grounded, which now I think I will say is becoming more common with a lot of AI tools. And I appreciate that. And that was partly because we said you can't just give an answer. But for us, especially, if you ask a question in case A, the answer has to come from case A. It's critical. And so we... And you have to know that it is. So everything is cited. Every single thing it says is cited. And for us, it was also worth noting because this was a question that took a lot of fine tuning in how we set up the language model. We don't give legal advice. We're not a law firm. And so was very important that the public-facing version of Conductor isn't giving legal advice. It's providing the information that you need to make a legal decision. And we had to understand the difference. And so Every answer that Conductor provides, like I said, is grounded and cited either back to the bankruptcy code, the local rules, or the docket of the specific case. And then we really had to refine the language and the way that it presented information so that it's not legal advice. It's providing the information that you need to either get legal advice or make your own decisions. Russ Fordyce 🏆 (22:38.72) Yeah, now you had mentioned that you guys have been at this a little while and the platform's a few years old and AIs come into this mix. Do you find that your users are still kind of double checking the eye or has that trust point reached a point where we can kind of rely on the platform and they're comfortable with Robert Klamser (23:00.314) I think, anecdotally, what we hear from the market is they do trust the platform, which is fantastic. And I think a lot of it is because we're not trying to do everything. And we're not trying to generate, and this is key for attorneys, this is not a generative program, it's not a drafting program. You're not using it to draft your legal documents. We were very specific about what we wanted the use case to be here. It's to understand documents that are filed. It's to understand the implications of those documents. And so when you're not saying, you you always run the risk of dealing with hallucinations in AI. Everybody knows that it's what all lawyers fear. It's what all users fear. But we're not drafting documents for lawyers. We're answering questions and we're going back to a citation. And I think because we said, that's all we're going to do and we're going to do it really well, we have developed trust because you can see it. You can see, okay, it's... You know, it is telling me that I have to respond by this date, these, and these are the three steps, and there's the document. There it is right in the document that they're talking about. That makes sense, that's real. And it's not drafting your response, it's just telling you all the facts that you need to draft your response. And that has made a big difference in kind of the trustworthiness of the platform, but it's because it's not trying to do everything. It's trying to do one thing really well. Russ Fordyce 🏆 (24:21.134) Yeah, now, I don't think, you know, I never really thought about it until talking with you, but these lawyers are becoming the single point of contact for what could be thousands or millions of creditors if it's a big company. And, you know, they're spinning up basically an entire organization to deal with this bankruptcy. It's almost like a little mini company temporarily for two or three years or whatever the length is. Robert Klamser (24:34.318) Yes. Russ Fordyce 🏆 (24:46.114) and you guys are really driving kind of this platform to reduce those costs, what do you think operationally had to happen to make that possible? Robert Klamser (24:55.999) I mean, first of all, that's exactly what it is. It is exactly like a new company. A company that's involved in a bankruptcy is essentially operating as two companies. It's the bankrupt entity and then the ongoing entity. And so, I mean, this is one of the tangents that I talk about. But if you're the CFO of a large company that files for bankruptcy, two weeks after you file for bankruptcy, you're still trying to run the company. company is operational, you are trying to restructure that company. At the same time, there's this other legal entity that's involved in a bankruptcy case that your team is also trying to manage and the rules have changed. All of a sudden, you're not necessarily dealing with GAAP, generally accepted accounting principles, you're dealing with the bankruptcy code and they're similar, but they're also different. so what we've done and where we've seen this kind of operational efficiency is going back to kind of the question. Yes, a lot of these large cases, the result of a large case filing is that hundreds, thousands, millions of people are impacted by that. And it used to be, and it sometimes is, that they'll call the lawyer. if you look out in the industry, the billable rates for for bankruptcy professionals especially, it's such a niche and they're very valuable. And the billable rates reflect that. And it's not the best use necessarily of time for... a small vendor to be asking routine operational questions to a bankruptcy lawyer who's trying to solve this problem and make sure that a multi-billion dollar company emerges from bankruptcy healthy and operational. But we can do that. And it used to be that we would do that with our own expertise and our own call center, our own communication staff who would answer email. We were innovative with online chat and we'd be talking to people there because it's much, much cheaper than a bankruptcy lawyer. But now we have this tool that reduces the cost even more. And beyond the cost, it's the availability and accessibility. One of the things, it doesn't matter how many people you have in a call center. If you file a 500 page document, it's gonna take a long time for anyone to read the whole document, to understand it. Robert Klamser (27:08.053) and to communicate what that means to a 500 person call center so that they can answer questions. Conductor does it immediately. Within minutes, everyone, thousands of people at the same time can be asking questions to get an answer on that 500 page document that was just filed. So not only are they not calling the lawyers, but they're not waiting for days or weeks for someone to explain it to them. The other thing that we didn't even talk about is it's multilingual. One of our largest cases of the last year, is Azul Airlines, it's the airline, it's the largest Brazilian airline. 32 million customers were potentially impacted by their bankruptcy. And their primary language is Portuguese. And Conductor speaks Portuguese. And it speaks, so you can go to the website, all the documents are filed in English. But if you're a Portuguese customer who's supposed to go on a flight in two weeks, you don't have to wait in line on a call center, you're not calling the bankruptcy lawyers, thank goodness, but you can go online, you can go online, you can ask Conductor. Azul just filed for bankruptcy. have a flight in two weeks. Can I still take my flight? And as long as that answer is in the legal documents that have been filed, which it generally is in a customer programs or critical vendors, there's certain motions that would impact the ability to have this flight. But you can get that answer immediately in your native tongue at a fraction of what that would cost to call a lawyer or to have a call center. Russ Fordyce 🏆 (28:32.438) Yeah, or if you're downloading a 500 page document to look for a single line saying, yes, flights will continue. It seems like a total waste of time. What does that impact look like for a good size of bankruptcy? mean, are we talking, you mentioned there were 500 call center agents. Is that a typical call center for one of these guys? Robert Klamser (28:39.87) Yeah, exactly. Robert Klamser (28:54.548) That... That would be 500 is large. It's great number. There's very few cases that require 500, but I will give you an example. We were the claims agent on the on multiple large crypto exchanges that that file for bankruptcy. Each of them were each of them had between 1 million and 10 million customers. A bankruptcy like that would generally require between 10 and 20 live operators at all times. And lot of those cases, initially when those cases filed, they wanted live operators available 24 hours a day. And now we can do that with a fraction. You still have live people. That's the other thing that's important is all of these resources, we didn't trade it. We didn't say you can't get live support, you can't talk to anybody anymore. That's not our business. We're still here to make sure that everyone understands the process. We're still here to make sure that that information is flowing. But most people just want to know what's the deadline to do this? Where do I send this to? How can I find the document that I need? Is this the right document that I need? And now we can do that with three, four, five people with some of the busiest cases. the nice thing is it's so much easier to also understand what's confusing to people. One of the things, this is off, this is. somewhat off topic and it's maybe not what the award was intended for, but I'll tell you one of the things that using AI allowed us to improve that doesn't involve AI is now all of a sudden we have this archive of all the questions that are being asked in cases. And it's really easy for us to log in and say, what's confusing today? What's confusing this week? And we've actually gone in and we've modified language on our website to make it easier for people to get these answers because we can see what the questions are. Robert Klamser (30:50.437) We could still do that in a call center, but you had to wait a lot longer. And if the notes that people are making on the calls are formatted differently, sometimes it's easier to of surface those insights. Now, not only do we have a tool that's answering the questions with AI, we're also behind the scenes using AI to say, tell me what's confusing. What are people talking about in this case? Like, are the hot spots in this case? What can we do so that not only does Conductor answer the questions, maybe we can just put the answers out there more easily so people don't have to ask the questions in the first place. Again, that's what happens when it's purpose-built and we're also experts in this domain is we're learning from it as we go and we're iterating on other parts of the business using AI to help us understand all the data that's coming back. So I geek out about this, I realize it, but the platform is teaching us a lot as well. Russ Fordyce 🏆 (31:45.518) Well, you've been in this industry a long time and what I know from personal experience is that lawyers have been traditionally extremely slow at adopting new technology. I I remember I was working with my father on a contract last year and just find and replace for him. He was an older guy, but he literally printed it out and marked it up with a pen and he would say, well, this term is undefined. We'll have to go find it. And I'm like, well, it's on page seven, page 12, page 13, and it's in the appendix. And he's like, how did you do that? So a lot of them are very kind of in their way slow to adopt technology. Why is this AI adoption just accelerating all of it? Robert Klamser (32:27.055) It does. You're right that. That lawyers are slow to adopt, but I also think part of that is law and the courts are slow to adopt and I do. If there's anything I've learned doing this as long as I have is. As much as I've wanted to innovate and straddles a company that's wanted to innovate, I also have to understand why there is why it's slower and I think. court operates with caution. The court in general wants to act for the benefit of all. so the court is always very cautious so that change and innovation doesn't disadvantage any parties. And so that's why you see slower adoption. I think there's a lot of caution. And I'll say, I think the antidotes of new use of AI and hallucinations and hallucinating case citations, that sticks with people because this is an industry that has to be very conservative and they don't want to make those mistakes. And so I realized now I'm going to again, story tell a little bit and go off, not off topic, but even 12 months ago and 18 months ago, going to conferences, sitting in panels, hearing lawyers, both teaching lawyers about technology and AI and then also hearing other panelists and other questions. you realize that those narratives stick for a long time. If you're going to talk about AI, every single discussion you had talked about the case citations where cases were hallucinated and the attorney was reprimanded in court. Everybody talks about that. But you don't talk about the thousands of other hours that are saved and the work product that's improved while lawyers are using AI for research, you know, It's like the evolution of control F or find and replace because now it is it's not just find this clause again. Find me everywhere where this clause exists in the contracts and help me find it and I'll draft the revisions. Now it's more like, you know, find all this and find the language. Find the language that's ambiguous so that hopefully we don't have to revise this again in the future because we've identified language that was weak. That's happening every day. Every day. Robert Klamser (34:53.652) And we're out trying to really promote that. But every time you talk about that, people will also say, but there was that, there was that hallucination. There are these hallucinations. And so I think people are slow to adopt because there's, this is a precise industry and the inherently the way AI produces content. If you really look at it, it's not precise. You know, it's, it's, it, it, it's still happening. Yeah. Russ Fordyce 🏆 (35:17.934) Right? It's predictive. Robert Klamser (35:21.225) Yeah, so that was a long answer to a simple question, but I think it's that it's not out of naivete. I think it's out of wanting to be precise, wanting to be conservative. No one is going to lose their job for doing something correctly the old way, but it's risky when you do something incorrectly the new way, even if you kind of have to experiment and iterate with that for there to be progress. So hopefully we're creating safe boundaries for lawyers to... to experience some of that progress. Russ Fordyce 🏆 (35:53.378) Well, like you said, we're only a couple of years into this with the generative AI capabilities. Where do you see it in five years? Robert Klamser (35:57.236) Yeah. Robert Klamser (36:02.878) That we think about that a lot. think that when we 12 months ago, less I felt like what you were hearing, there was there was the data that was coming out and you started to see these different studies that said 30, 40, 50 % of law firms are even allowing lawyers to use AI. And I think there were still people even a year ago. there were probably a lot of people out of firms who thought, yeah, this is not something we're going to do. I think that that we've seen a tremendous shift there in the temperature and the tolerance around AI. And so I think that now you have to, and it's a matter of finding out what the right tools are for your firm. I think that within the next five years, what I'm hoping is that we're going to see more guidance and structure at a federal level. Right now you have, like said, have 94 bankruptcy jurisdictions. You have a lot of courts, and this is bankruptcy, but just in general in the court system, you have a lot of states and individual jurisdictions who are trying to come up with a legal framework to give guidance for lawyers to use AI, to use AI in their work product. And what I'm hoping and what I think will happen is over these years, the next five years, you're going to see a... more unified standard for what's acceptable, what's acceptable, how you have to disclose what you're doing, what is expected. mean, the example we like to use is the transition from having the bankruptcy code in law firms office. And when you needed to understand federal bankruptcy code, you had to go and look through the printed bankruptcy code. When I started in this industry 25 years ago, one of the things I got was a printed copy of the bankruptcy code. And you were expected, and it was OK that you went and used it. It was OK that if you wanted to answer a question, you pulled that out, you highlighted it, you had your post-it notes, and that was OK. It was OK to spend an hour researching the bankruptcy code to come up with an answer. That is not acceptable now. That's basically overhead that you shouldn't be charging a client for. Robert Klamser (38:25.418) And I think we're going to, we're at that point where very quickly the market is going to say, if I, if I hire an attorney, it is not acceptable for you to do something the old way when AI can do that for you much faster. That, is just what we expect. And so I think we're going to see this evolution. There's so many tools out there. There's so many great tools out there, but I do think we need a framework. We need a, we need expectations and then. I also think we're going to see a lot of consolidation in these tools. There are some brilliant minds out there. There are a lot of companies that are features for bigger businesses. And I think you're going to see a lot of consolidation in the legal space, in the technology space, where these great features are consolidated around different companies. I think you're already seeing some players emerge. And I think that's what Strato is trying to be, is we are the experts in our domain, and we are finding, building, assembling, acquiring these tools so that we can be an easier way for these attorneys to get their work done. Because no one's going to go and use 20 tools. That's not reality. So I think those are a lot of the things I see in the industry. But I do think there's going to be a better legal framework. And I think there'll be a consolidation of tools. And I think it's just the expectation five years from now is you have to use it professionally. Russ Fordyce 🏆 (39:37.901) No. Robert Klamser (39:54.51) you will have to use these tools. Russ Fordyce 🏆 (39:58.722) Yep. Well, Robert, fascinating conversation. You're definitely doing Jomans work for making this complex and tedious process a lot easier for all of us. So congratulations and thank you. Robert Klamser (40:10.6) thank you. It's like I said, I geek out about this. It's what we all know. It's what we do. so it's just, it's an honor to be, to get awards like this and to have someone acknowledge that we've added some value out there. We've added value to our space. It's very validating and we appreciate it very