In this episode of The Law Firm Blueprint, Jay Ruane and Seth Price are joined by Michael McCready and Jerry Zhou of Supio to explore how AI is revolutionizing the legal industry. The conversation touches on the challenges of implementing AI in law firms, such as overcoming the “human bottleneck” and building confidence in new technology.
Whether you’re a personal injury lawyer looking to reduce overhead or a criminal defense attorney eager to handle complex cases more effectively, this episode offers actionable insights into how AI can elevate your practice. Tune in to learn how adopting AI tools like Supio now can put you light years ahead of the competition.
#LawFirm #Blueprint #AI #SethPrice #JayRuane #MichaelMcCready #JerryZhou #LawFirmBlueprint
Links Mentioned
Hello. Hello, and welcome to this edition of the Law Firm Blueprint. I’m one of your hosts, Jay Ruane, and with me, as always, is Seth Price. But today we are joined by our Law Firm Blueprint legal AI expert, Mike McCready, back in this series, and Jerry Zhou of Supio, and we want to talk all about AI, but I’m going to throw it over to Mike, who is our guru when it comes to this stuff. Mike, what’s your first question for Jerry?
Well, let’s, let’s start by you, Jerry, telling us a little bit about what Supio does, and I know it does an awful lot. So let’s, let’s start with the top selling point that you are involved with, with Supio?
Yeah, you know, Supio primarily focuses on the documents in plaintiff practice areas, and what we do is we have specialized AI that goes and helps you understand and structure all of the data within these documents. And by doing that, you’re going to be able to use much more accurate and powerful automation tools across the entire case. That’s everything from understanding whether you have missing bills or medical records to understanding hidden injuries that weren’t diagnosed, all the way to generating demand letters, to using AI to help negotiate with settlement adjusters, to taking this case in the litigation, and being able to compare and contrast things like depositions to using our AI in court, and if you follow the news recently, we’ve helped some firms get up to $495 million verdicts against, in product liability suits by helping them pre-ingest and analyze their data ahead of time.
So one of the issues that we, that we run into with AI is the data set that you are pulling from, right? You know, we’ve all heard about hallucinations, and they come from, you know, getting your answers outside of a controlled environment. So how does Supio deal with a controlled environment in generating the responses that you’re looking for?
Yeah, it’s really about what domain you choose to go and build this product for. So you know my background and my Co-Founder, we’re all about documents and understanding these. So when you break down some technology like generative AI, there’s essentially two parts of what you can do with it in legal. One is to use it for problem-solving. So you know the latest reasoning models and all these things—not only can it help you summarize and find information, it can help you do reasoning and math. The second part is it also itself inherently has so many patterns that can help you answer open-ended questions. Hallucinations primarily are usually coming from the second part of asking really open-ended questions, perhaps about legal precedent that you’re relying on these patterns to be able to answer.
So to really combat that, there are a couple of ways you can do this. The first is, like, we focus on helping you understand all the documents and data in your case, and by doing that, we’re breaking this down into easier-to-understand different types of facets, whether you know, we understand the document type as an insurance document, so your query is trying to look for that, and you’re going to be able to pull that in, to also establishing these knowledge bases, where we’re taking your common objections from discovery and enriching them and classifying the objection types so that when we measure this against creating these types of discovery responses, we can create a much higher accuracy than what’s coming out of the box.
So obviously hallucinations are just a part of using this tool, kind of like when you surf the internet, you’re going to find bad and good data. And when I look at our users over the course of 12 months using this product, it’s—you have to build a relationship with this AI, where you understand what you think it’s going to be good at and what it’s not. And by doing that, it’s going to help people speed up and be better practitioners. But it’s not a tool that you can just take advantage of and instantly scale 20x without really understanding the technology itself.
So Jerry, your product does a lot of things, and it’s fascinating watching the run you had has been awesome. What are the two or three areas you think are the biggest, quickest wins that as a business owner, as a law firm owner, you see people leveraging through Supio to get the, you know, to turn what’s a cost center into a profit center?
Yeah, if you think about one of the most painful areas of personal injury, It’s probably going to be around the documents, around medical bills and medical records. There’s so much work around the human processing of this, where you’re trying to build out ledgers or establishing chronologies or creating demand letters. So when we digest these, we’re measuring this to the 99th percentile of accuracy when we’re extracting, and so you’re going to be able to get human level accuracy from demand letters and chronologies that you can instantly use. And then the software layer that we put on top of that allows you customization and better understanding of your data.
So in pre-litigation, this is something that is an instant value for any firm that can just go and dramatically reduce the overhead of doing this. On top of that, what we find is that you know, when you have these latest breakthroughs in reasoning models, you can run these models on top of this, like really clean data now and now, suddenly you understand all these injuries, like you understand that maybe this person had a concussion, but they actually call the case manager. They’re like, I’m fine. I just want this settled with.
I just had a case today where we were just working on this with some of our collaborators, and you know, at that point you should tell this plaintiff to kind of slow down, right? And like, we need to go get this properly treated and diagnosed. So as we work through these cases with our customers, we think that we’re also going to be able to uncover and coach them in the higher settlements and making sure that their clients are getting the proper treatment they need. So that’s how we think about the value here.
So before I toss to Jay, what do you see as the biggest mistakes? You look inside a lot of law firms, as the three of us have access to as well, but you know, these law firms are not always rational players. They’re not always set up particularly well. What are some of the bigger mistakes you see as people try to implement AI into their firms?
Yeah, I think the biggest thing about AI, right, is that there’s a level of learning involved in engaging with a chatbot, so finding the right champions of AI within a firm, like the best ones. My favorite stories are like, right when we started, Seth, when we got connected with you, we were working with a few firms. One day, one of these lawyers just introduces himself to me. He’s like, “Hey, the firm told me I need to learn AI. That’s all I got.” And then within three months, he’s so good at using our chat and using our AI that he doesn’t even really write full sentences.
You can always tell, like, in the beginning, when you talk to the AI, you’re saying, like, “thank you” or “please.” Later on, it’s just like a complete, like, a really quick evolution there, right? So when we think about this, it’s like, you’re not really just implementing software. You’re almost teaching someone how to use the internet or, like, learning how to use a computer, right? There’s a fluency here, and that’s the investment that if you can get your firm on this, it’s going to just revolutionize how you practice. And the earlier you do that, I think the earlier you’re going to be able to take advantage and understand how to change your hiring processes and also how you manage your people.
I’m just smiling because it reminds me, Michael, before I go to you, of the early days of the internet where I’m sitting teaching Flavor Flav how to surf on Google. I feel like Jerry had taken that mantra and had to, you know, show me how to do this stuff. But sorry to cut you off, Michael.
Yeah, no, no, you hit on about three or four things that are very near and dear to my heart. You know, you can learn AI yourself. I have spent a lot of time. I love it, you know, it’s all about framing the query and having a control data set and teaching it. AI learns, right? You give it feedback. But the benefit, and listen, you know, ChatGPT can do medical chronologies, but they are not going to be as good as some of these other, you know, AI programs, you know, like Supio and others, because you have already trained that machine to get rid of a lot of the mistakes.
And, you know, I think a lot of people misunderstand how much work goes into creating, you know, a custom GPT. It’s not just a matter of, you know, “Hey, let’s upload this and run.” It’s giving the feedback. And that’s why, you know, companies like Supio and others, you know, they’ve done all of that groundwork for you to get to that 99% success rate.
100%.
Jay?
So, you know, it’s interesting to me, coming from a non-PI perspective, I can probably walk into Seth or Mike’s office and you guys will have a file that will be this thick of paper for a $10,000, you know, PI settlement. And I, on the other hand, have a file that’s this thick for someone who’s looking at life in jail, right? Because it’s not really paperwork-heavy in the criminal side. But I’m thinking to myself, wow, I need to reach out to Jerry.
I need to, because we’ve, you know, I’ve been lucky enough to work on some Innocence Project type cases where people have been in jail for 30-40 years, and we’re finding stuff wasn’t disclosed, and that type of thing. And that’s when we’re finding, you know, we’ve got 18 bankers’ boxes worth of materials that get scanned, but then we have to figure out what was there. And I think there are tools out there now that could help us really sort of process this volume of information that we’re getting.
And I’m thinking to myself, how many Innocence Project cases could we ramp up and take now if we have tools like this? So I think, you know, on the criminal side, we’re on the cusp of being able to tap into these things. But Jerry, I want to ask you a question about the abilities, you know, internally. And I know Mike does this internally. You know, you keep track of adjusters and that type of thing. Is there the ability to extract across the board?
Hey, you know, these adjusters respond this way to these types of cases. And, you know, with a, with an 8% PPD of the thoracic spine, they’re going to be paying X amount of dollars historically, if that stuff is tracked. Is that like the next level, where it could be anonymous data that’s supplied? Because, like, as Mike and I talked beforehand, I’m sure the insurance company knows their adjusters and what they spend and what they’re susceptible to. And if I was the CEO of Allstate or one of those other companies, if I had an adjuster who always paid top of the line for certain injuries, I wouldn’t send them any of those injury cases. I’d send them the cases that they pay the least amount on. So there’s got to be a thing. But I don’t know if PI lawyers are sharing that data among themselves, but it might be able to get that stuff anonymously out and provide that as a service, right? It seems like that’s capable.
Well, let me address that in like kind of two levels, right? When we think about structuring a case, you basically look at—let’s take a step back in history a little bit—and we look at something like a practice management system. We see this a lot in product liability cases, where, you know, in order for them to go and query these cases at scale and understand some of the data, like settlement values and things, they have to think ahead.
Like, you’re processing, you know, 10,000 cases of Camp Lejeune. You’ve got to go create all those fields beforehand. You got to get the people to put the data into the fields. And if you can get them to do all these things correctly, suddenly you have all this structured data, like you have an adjuster name, and then that adjuster name can now be matched against other adjusters that have the same, and then you can pull out the other relationship there, right?
The problem I see, even for our own business, is you understand what you want at the moment, like you’re working on right now. I don’t have the foresight of understanding, oh, I needed to go pre-index, like, 500 adjuster names in my previous cases to go do that. That’s one of the biggest challenges, and that’s what AI is able to solve really well. Because what AI does is it lets you query against unstructured data.
So now instead of, like, having to query this, you have all this data that’s just text, and you can just say, “What was the adjuster name of this case?” There are definitely processing limitations and how we work on things, but one of the biggest things that we help firms understand is relationships between their cases across the board.
One example is there’s a large, high-volume firm that we work with, and they’re both working on a Tesla mass tort, and they’re also working in single event. And then one day, they were like, “Hey, how many of my single event cases where the defendant was driving a Tesla could be eligible for this mass tort?” Right? So these types of questions are questions that we can help people answer, and as we genericize that interface and help people go connect this data in an easier fashion, connecting the adjusters and the relationships between these things are just places that we’re marching toward really quickly.
Wow, that’s amazing. Mike?
Yeah, no, I like to explain to people that, you know, when you were in law school, you had to master the Boolean search, right? Slip within four of fall and, you know, and if you had a good Boolean search, you could find cases. Then comes Google, and you can type in natural language, but what Google gives you are references. It gives you websites and URLs, and it does a pretty good job of answering your questions. But now, with generative AI, you know, it’s giving you the answer, right? It’s not giving you a website, and it is pulling from wherever you direct it to pull from.
And the whole idea of ChatGPT is having that conversation back and forth. And if you don’t get the answer that you want, you change the query. And what I’m trying to stress to people who want to get involved in AI is, you know, learn how to frame a query. Now with Supio, you’re making it easy, right? You’re making it so you don’t have to, you know, become an expert at queries. But that’s where this AI is going. And having a company such as yours is going to really speed up that learning curve for people that don’t want to invest the time to kind of learn the foundations of AI.
So you’re telling me, I don’t have to say please and thank you to ChatGPT, because I’m like, “Can you please tell me,” and “Thank you” in my queries? And I’m probably just wasting water by adding those extra, uh, words into the query, right?
100%. I would like to add to that a little bit, which is there’s more to generative AI than just the asking of questions and getting results back, right? Like the actual understanding of data helps you automate workflows. We just released a new product called Case Economics, where we go through and break down every single one of your medical bills and ledgers and insurance documents.
This is something that is extremely time-consuming, and no one likes really doing it, right? And we’re not really relying on the question-answering part of generative AI. It’s really the data ingestion and formatting that into structured data part. And once you can break that into the basic components, now you can do work with it in a spreadsheet. You can add it up. You can understand the differences and balances, which is traditional software engineering and like Excel, right, to kind of go do this type of work.
In order to go and really understand this data, which is way beyond what I think a foundation model can do, you really have to understand the different types of bills, whether it’s a Kaiser super bill, or it’s an insurance document, or it’s a medical insurance or med mal type of case, right? So basically, being able to understand all these things allows you to be able to focus in on creating human levels of accuracy, on extractions, and being able to build specialized models for these things.
Jerry, one of the things that, you know, we struggle with is there are bills and there are bills. And Mike, this is more Michael’s world than Jay’s world, but there are doctors we can send people to that will run up a bill, but the insurance companies know who these doctors are and immediately give it a substantial haircut, versus other doctors that go and sort of like the general, you know, you might actually go to yourself category.
How close are you guys to being able to figure out any data you give them, it’s over time? But to me, those are some of the things that are pretty exciting, that you can start making decisions that will end up helping people by avoiding things that they think are helping them, but really aren’t going to economically benefit them in the long run.
100%. I mean, even to that particular issue you’re talking about, as we work with our partner firms, right, like they always mention, you can get the initial costs from the medical bills, and then you have to get the insurance ledgers to get the actual payment amounts, right? And so when we think about building out a data set, you know, we call this a knowledge graph of the case. So all these different line items, every single cost, they’re all related to the event, or it’s related to the other types of costs, payments, and insurance companies that are actually doing this work, and they’re all interconnected in a way now that when you are trying to automate a workflow like generating a demand letter, or you’re trying to build a medical ledger, or you’re trying to figure out which lien to pay, we can show you the correct relationships to do this type of work.
I mean, people have to remember that AI is predictive, right? That’s the foundation of it. It looks for patterns. Everything goes down to zeros and ones, but you know, it’s looking for patterns, and it’s creating your answers based on predictive analysis. And it’s already pretty good, but the more you program it, the more you put into it. That’s where you can get to your 99%.
How good is this stuff? And I don’t have a lot of exposure to it in my practice area, and I’m sure other people—how good is this stuff at like reading a person’s handwriting? Because in my world, everything is handwritten out. Like, you know, I can see medical records. They’re keyed into systems and everything is clear, nice, but I’m dealing with, you know, a photocopy of a photocopy of a handwritten statement from 1993. And, you know, it takes us humans forever to sort of, okay, in the context of this sentence, what word could this be? Is AI at that point yet? Or is that something that I need to be getting ready for, to be able to use in the coming years?
It’s definitely not at that point yet. So, you know, when we think about building a product like Supio, we try to establish a confidence level of how well we recognize this document, and if there’s handwriting or anything like that, we escape that into a human quality control process, where we transcribe that into text. If you’ve ever seen a chiropractor bill or chiropractor-like case, like the check marks on the graph where they circled anatomy, that’s something that I think AI is still a little bit far away from being able to read.
But we come from the other side. Like when we work with a law firm, we think about the user experience of the law firm. We can’t offer a product and tell you, “Hey, send us all your data, but it only works for this subset of things.” So like, remove all the handwriting pages, remove the chiropractor bills. We want this to be a unified experience that makes it a lot easier. And as we approach a higher degree of accuracy and extractions and all these things, I think that, you know, it’ll become more and more efficient. But for now, I think people have to still intervene in these types of data.
I mean, it makes a lot of sense, but it’s awesome that you guys are heading in that direction because that gives, you know, that gives me something to be—I need to be pinging you back, you know, every couple of months, “Hey, where are we with this?” You know, just so I can—because the minute it becomes available, I want to jump all over it, you know, because that can really help. I mean, I have some—I have clients in jail we truly believe are innocent, and we’ve got so much handwritten stuff that it’s, we literally have law students trying to decipher things. And if we can get any sort of help, you know, that’s going to be phenomenal. I’m looking forward to the future for sure.
You know, so, Jerry, everything is changing and evolving so quickly at an exponential rate. Where do you see AI at large going in the next six months?
So the fact that this is being recorded is funny, because I’ll make some statements and I’ll just be totally wrong, and you guys can remind me of this in six months. But, you know, there’s a common saying right now in the industry of people who work with these models, that it’s becoming somewhat asymptotic, meaning that the more parameters you train with them, they are improving, but they’re not improving in the ways that they have been in the last year or two.
So what that means is, like, as we ingest more data and we’re able to understand and create a level of accuracy, that’s probably going to slowly, incrementally improve as we move forward. Where the actual improvements have come are in two other areas. One is really around the ways that you can calculate and ask questions, right? So, like the latest release from OpenAI with the 0103 models, they call that chain of thought.
Chain of thought is fascinating because instead of asking a direct question, you can break down the question and ask multiple substeps of that question. So it’s really planning before it answers. And what they were doing with that is they measured that against some of the top math questions in high school and college, and they were able to get really, really good answers to the point where it’s almost as good or better than humans are.
And the theory there is, the longer you think, the more accurate the answers are going to be. So that’s an example of very innovative ways where you’re taking two models, and one is guiding the other, and you’re getting the answers that way. So I think, problem-solving-wise, AI is improving dramatically, really quickly.
The other side is what they call multi-modality, which is really around—you can think about, like, we do text really well, right? Large language models are generated on text through the internet. So, you know, it can predict the next word in a sentence. But it turns out that you can use these models in different ways, like you can predict bits and ones and zeros and patterns.
So images or videos—these are places where there are huge breakthroughs. So, you know, when we talk about something like personal injury or even criminal law, like processing something like hundreds of petabytes of video to find out, like, what actually happened in the crime or what happened in an accident—that’s something AI can now do really, really well, because you can just automate that versus having humans go and watch this.
Of course, we’re still at this point where we’re trying to measure the quality of these things for really, really specific industries like, you know, criminal law or personal injury, but that’s what the promise is, I think, for breakthroughs in the next six months.
All of this takes incredible computing power. I have noticed, even with my enterprise GPT, you know, it’s taking longer and longer to generate this stuff. And do you see—where do you see—or is technology going to be able to keep up with the demand?
At some level, do you feel better than—it almost feels spooky when it’s that fast, like, I like the fact that it has to think about it for a second.
Yeah, there’s a couple of things. Actually, the latest breakthrough is, what’s funny is that they came out of China, which was like DeepSeek, was the most recent model that came out. The U.S. is in a trade war with China, obviously, so they restricted all the computing power there, so they had very little computing power, and they were able to produce results on par with these very big, large language models from OpenAI and Claude.
So, you know, I think the innovations to condense this type of processing are coming now. It’s kind of like people who are trying to hit the gym, right? You can either bulk up or you can cut, and it’s really hard to do both at the same time. But once we hit this efficiency, I think it’s going to be a lot more about being able to build smaller and smaller models that can help you create a lot faster answers.
Yeah. I mean, listen, you know, AI is fantastic for, you know, coding and huge, huge projects. But that’s not what most of us are using it for, right? We’re using it for, “Hey, I’m planning a trip to Portugal, and I want to do this, this, this, and this, give me an itinerary.” You know, that does—that should not take a lot of computing power. So I do think that, you know, there will be some restrictions, just because it’s taking up way too much power.
Well, I’ll throw this to Jerry. You know, look, the groups that have hit the legal market are not inexpensive. One of the things that I’m hearing is that OpenAI and Gemini, these guys are almost giving us loss leader costs, if not free, in order to get us hooked. Do you expect the public to be paying significantly for what we right now think of as free AI?
I’m a big believer in Moore’s Law and the doubling of capacity every single year. That so far has not broken. I don’t think that will be the case. I think it’s just more where the public is entranced, right? Like, building really big brain, like trillion-parameter models, is something that everyone gets caught up on. Like, “Look at this, it can answer these questions. It’s so smart.”
And now we’re getting into the era of how you utilize them. Like, it’s not—I don’t think training even more larger models is going to be something that’s going to substantially change people’s lives anymore. One of the challenges you can think about is literally just the interface. Like, Michael, you were talking about formatting questions for the AI to answer, right? If you think about a practice area or something that you’re going to go ask a question, there’s a limit to how much you can communicate through a single query.
So, you know, at some level, you can’t even go give it all the context that we know. So I think at this point, we’re just trying to figure out the more innovative ways to use these models and guide them to create better results for people. I always say that Supio is more of a measurement company than an AI company. So we take the industry’s best breakthroughs, we look at the work that our customers do in plaintiff law, and we try to measure whether it’s actually able to go and create quality of work that’s the same or better across the different types of facets. Tools that help us measure faster will help have more breakthroughs in this particular industry.
You know, so what’s great about AI is that, you know, at least in law in general, there’s a lot of things that are done the same way all the time with every case, and applying AI just increases the efficiency as one benefit. But then, once again, the analysis that you’re able to bring by referencing everything in a case file.
So, you know, I tell lawyers—and see if you agree with this—that AI is here now, and the people that are working and using it now are going to be light-years ahead of everybody else. You know, the people that say, “Oh, I’m just going to wait and see. It’s not there yet. I don’t trust it.” You know, the time is now. Would you agree?
I 100% agree. I think this goes back to an earlier point we were making, which is, with any new technology, you have to build a relationship with this technology. So the sooner you do this, you’re just going to become an expert, and when the next feature or breakthrough comes through, you’re going to be more comfortable embracing it. At some point, if you’re falling behind, it’s just so far behind that you can’t really catch up.
When I think about AI in the perspective of practicing law or some of the work that our customers do, it’s really about eliminating the human bottleneck—not only in automating the existing work that people do but also the fact that, like, if you have a theory about a case and you need to ask an associate or paralegal to go chase that down for multiple days to go understand, you get that answer in an instant. If you remove the human bottleneck from that level, imagine how much better the outcomes can be and how many more theories you can look at.
So, you know, it’s really hard, I think, for people still to understand the effects until, you know, this technology is a little bit more mature. But learning how to use it is undoubtedly going to be a game changer for anyone.
I think that when people first start dabbling in AI and you hear what can be done, it just blows people’s minds. “Wow, I had no idea.” And it just keeps getting better and better. And products like Supio—you’re constantly adding more and more features to make it as user-friendly and as helpful as possible for the lawyers that use it. Can you give us any idea of projects that you’re working on to include?
Yeah, so, I mean, we’ve been super deep into pre-lit workflows for a long time, right? And we talked a little bit about this case economics, where we understand medical bills. There are signals that we’ve built recently, which help you understand missing documentation and missing injuries to help you better get your clients treated.
But on the other side, on litigation, I think that’s where things are really exciting, at least as toolmakers like ourselves, because litigation is quite a fluid workflow. So a lot of what we think about is—it’s almost like software engineering, where, when you write code, you have all the existing documents within the codebase in context. So you can use something like AI to help you write the next, the next actual class that you’re writing.
So in legal and litigation, there’s a set of documents from both the defense and yourself, which we put into context. And then we can load up any type of Doc X that you can have, and you can basically go and start generating that from the existing case data that you have. And we can also load up precedents, and we can also understand how to break down these different types of documents in ways that create a lot more reliable responses than what’s available in the market today.
I mean, it’s fantastic to realize, you know, one of the pain points of litigation is answering discovery. And, you know, when you have access to the entire case file, it largely answers the discovery, at least the first draft, from lots of different sources, and it’s pretty complete. That’s one of my favorite features.
Yeah.
That’s—it’s really, you know, I’ve been able to mess around with it a little through a friend, and it really is impressive what you’ve done, Jerry. You know, I look forward to one being available in my niche, but for personal injury lawyers out there, it really, for me, from my perspective, seems like a no-brainer to jump on a platform like Supio so that you can take advantage of this deep computing power and ability to get out from your data the type of things that can make a difference maker in your files. And you can do it at such speed.
I mean, I can remember when I did PI, you know, 20 years ago, you know, doing a medical chronology would take a paralegal—you know, you task the person on Monday, and you get it on Friday. And that doesn’t happen anymore. Now, with programs like yours, it’s, “Here you go.” And that’s a wonderful thing because it allows you to see the whole context of the injury journey. And I think that that is something that allows lawyers to do better lawyering because they have the knowledge so quickly that, when they’re in the flow, they can gain what they need. And that’s something that, you know, heretofore hasn’t been available. So bravo to you and what you’ve been doing.
Thank you.
You know, and that’s great, Jay, but that’s after the fact. I think where the real benefit is analyzing the data in real-time, right, like Jerry had said, being able to find a possible TBI right from certain things and recognizing that, hey, a medical record that came in and mentions a CT scan, and we didn’t know anything about the CT scan, and alerting. So really proactively applying AI to cases—I think that’s the real breakthrough. It’s all well and good when the client’s done treating, and you have all the medical records and you can do a chronology. But, you know, once the client is released from treatment and they’re done treating, the case is over. It’s never going to get any better, right?
So if you miss something during the treatment, and if Supio or AI can flag that, you know, that’s going to be a value driver. That’s going to potentially increase the value of the case. And having access to this kind of stuff in real-time—the analysis as the client is treating—I think is the real benefit.
Reminds me of why I should not be doing PI cases.
Yeah, I like—this goes to kind of removing the human bottleneck, right? If you think about just how much data is coming into a law firm, they’re like a giant document processing system. And there’s just—it’s not cost-effective for any human to sit there and look at every single page of a medical record. And when you look at kind of the lower end of claims, especially, it’s really hard to be able to go do that work.
So this is a perfect application of AI, where the AI can sit there and reason about this case for 30 seconds and try to go understand things that would take people hours and hours to do. And we think you can increase that standard of care across every single one of the cases that are coming through Supio.
You know, as an example, what we always do anytime there’s an emergency room record—you know, we always request it immediately. We always review it because we’ve got to find out what the client said at the emergency room, right? Sometimes it’s not what they told us happened, or they neglected to tell us that they had had 12 beers beforehand. So, you know, it’s really nice as soon as an ER record comes in, you know, to run it and find out what the history was, which was given by the plaintiff.
And we’ve caught cases that we’ve had to cut, right? We had a—you know, we’re not going anywhere with this emergency room record. And, you know, now we don’t have to worry about, you know, manually a person looking through all of this. You know, as soon as it comes in, it’s analyzed and gives us an idea of where we’re going on the case.
Awesome. Well, thank you so much. That was great, Jerry. I appreciate it. Any final words as we wrap up here?
I think we’re in an exciting time. Like, I look at—I’m a technologist at heart. I remember when, you know, when I just joined Microsoft a long time ago. You know, I look back in history, and I looked at the invention of spreadsheet tools like VisiCalc in the 1970s and 80s. And then you look at databases and cloud databases and how that’s kind of invented practice management systems. And now for knowledge workers, AI is something that will truly revolutionize this space.
And I’m happy to be here and work with our customers and prospects to kind of build out the next generation of these tools.
Amazing, amazing. All right, folks, that’s going to do it for us this week on The Law Firm Blueprint. Of course, you can always take us on the go by looking for The Law Firm Blueprint wherever you get your podcasts. Be sure to leave us a five-star review and a comment. Of course, you can catch us live every week, 3 PM Eastern, 12 PM Pacific, live on LinkedIn or live in our Facebook group, The Law Firm Blueprint.
But Jerry, thank you so much for being with us today. This is Jerry Zhou from Supio, along with me, Jay Ruane, Seth Price over there, and all things AI for lawyers. Mike McCready—that’s going to do it for us. Bye for now. Bye.
Subcribe to our newletter to receive news on update
