The following is a transcript of Episode 27 of Championing Justice. You can listen to the full episode here, or watch it on YouTube.
Darl:
Thank you for listening to the Championing Justice Podcast. My name is Darl Champion. I am a personal injury lawyer based in Marietta, Georgia, just outside of Atlanta. I own and run my own personal injury firm called The Champion Firm.
On today's podcast, we're joined by Ryan McKeen. Ryan is a plaintiff's trial lawyer based in Connecticut. He has owned and managed his own personal injury firm for more than a decade and is the author of the bestselling book series, Tiger Tactics, which teaches lawyers and talks about how to own and run their own law practice.
In 2024, Ryan launched Best Era and is one of its founders, working with lawyers across the country on leadership, firm building, and creating practices that are both successful and sustainable. Today, a major part of Ryan's work is building tools and services that help lawyers prepare for and shape the future of the legal profession.
Ryan has written extensively about artificial intelligence in the legal field and how it will impact the practice of law. And that is going to be our topic today, is talking about AI in the legal industry, generally speaking, but also it's going to have a specific focus on how it affects personal injury practices. So thanks for joining us, Ryan.
Ryan:
Well, Darl, thank you so much. And I was hoping you would condense my bio to just big Darl Champion fan. And I greatly respect your views and your takes, particularly on the personal injury space and the practice of law. I really look at you as sort of the moral center of the profession. Thank you. I'm very excited to be on with you.
Darl:
Thank you. Thank you. All right. So Ryan, before we talk about some of the specifics of AI in the legal industry, tell us a little bit more about your background and how you got to where you're at today with your own consulting firm, Best Era.
Ryan:
Yeah. So I was one of these people who, if you look at my sixth grade yearbook, it was like, what do you want to be when you grow up? I wanted to be a lawyer. And really I just wanted to help people. That was my thing. So when I went to law school, I didn't want to work for the government. I didn't want to work for big companies or defense work or anything like that. It was like I wanted to represent people.
And I was very fortunate to get a job right out of law school as an associate at an old school general practice firm. I don't know if these even exist anymore in very many places, but it was the kind of place this was the best firm in the town. I mean, people went there when they had their car accident, their kid got arrested, they got divorced, grandma died and they needed the state done.
And so it was a really old school, community-based, full-service firm. And the lawyers were older and they really were like, "Yeah, I don't want to litigate so much." And so I was the young associate who got thrown into a bunch of litigation of all sorts, commercial litigation, zoning appeals, probate litigation, family law. I mean, just the gamut of litigation experiences.
Most days I'd be, most weeks I'd have seven to 10 court appearances, and this would be for years and years on end, just court to court to court, arguing motions, appearing for clients, taking depositions, just those things. But what I really began to gravitate towards was the personal entry work. I felt like that was, to me, the work that I could be creative on. And the notion of telling a story and also the flexibility and upside of contingency fee models, where it's like by me investing all this time, I'm not billing the client because most of my clients did not have that much money.
Their entire litigation budget for cases I handled would be $7,500. There's only so much one can do in that space. So in personal injury, that constraint wasn't there. And I began to grow that area of practice and I began to actually land wrongful death cases.
And I was like, do I want to refer these out or do I want to learn this stuff? And that set me down a path of really learning David Ball and Don Keenan and going down to Atlanta and going down to Keenan Trial Institute and really trying to develop that craft as to what it is that I wanted to do.
But six and a half years out, I said, you know what? This is an older school firm, their paper, they've been incredibly successful of what they've done, but this is when the cloud was coming online. And this is when the opportunities to build more modern practices were, I think, really viable at that time.
So I decided to open up my own firm and again, just doing more general practice litigation stuff, but ultimately I phased it out over time into really just focusing on personal injury and ultimately winning myself off of all other litigation and being a hundred percent all in on personal injury and then building a firm of very close to 40 people just doing personal injury work.
Darl:
Gotcha, gotcha. And you originally started as a solo and then grew your firm, or did you have a partner? Tell us about that evolution.
Ryan:
Yeah, so when I left on my own in 2012, I had a partner. I had my friend Megan Frieden, we went to law school classmates. I was like, "Hey, you want to start a firm?" Neither of us knew what the hell we were doing, and there wasn't the amount of resources.
At that time, Darl, back in the old days, you may remember, LinkedIn was just a place to you put your resume, and so there wasn't this sort of sharing of information. There was lawyerists, and I would read what Sam Glover was writing and things like that, Carolyn Elephant words out there.
But we went into business together and as often when two people don't know what they're doing, and it didn't last, we didn't make it a year. And look, their firm grew to be ... We're friends, their firm grew to be a very successful family law firm, but the sort of compatibility of doing family law and personal injury was not there.
And then I did go solo, and ultimately I ended up in business with my now former partner in 2018 when we started building Connecticut trial firm after a period of iterations of me as a solo or a solo with some help.
Darl:
A very successful trial law firm. Y'all got a hundred million dollar verdict, which not a lot of lawyers claim on their resume, so congrats for that.
What led you to want to go this direction of focusing on helping other lawyers grow and manage and run their practices?
Ryan:
Yeah, so I mean, really, I look at my career and I think of it really in tenure blocks. And the first 10 years, I really spent a lot of time learning how to be a lawyer and being a lawyer. The next 10 years, I really spent a lot of time trying to figure out the business and learn the business. I've been out of law school. I've been a lawyer now for 20 years.
And when I left my firm, it was like, well, I have the opportunity to rethink what I want the next 10 years to be. And what was really speaking to me at the time and speaks to me now is really teaching. And it was, well, how do I do that? And I'm an adjunct professor of legal entrepreneurship at University of Connecticut School of Law, but that does not pay the bills.
And I'm like, how do I really do this? And so creating this company that really allows me to do a lot of teaching and consulting and work, I listened to myself. It's just where my mind was going, where my interests are, what energized me and what excited me, and ultimately just being confident enough to go in that direction.
Darl:
And so what does Best Era do? Give us the elevator pitch on what your company does.
Ryan:
So here's what our pitch is at Best Era. Our team is comprised of five of us from Connecticut trial firm. We brought over Brittany Green, who was our marketing director. So Brittany ran marketing with me at the firm.
We brought over Allison, who's my wife, and Allison did our operations at the firm growing it, brought over Marissa, who was one of our top paralegals and helped and wanted to launch paralegal training, Para Era with Marissa. And I brought her-
Darl:
Great name, by the way, Para Era.
Ryan:
Para Era brought over my former executive assistant, Pam as well. And Pam doesn't do any client-facing work, but Pam does all of our social media marketing and webpages and graphics, and she edits our podcasts and things like that. And she's in Mexico. So it's all five of us there.
And what we identified as needs in the space and what it is that we wanted was we wanted to in some ways build a one-to-many model. We didn't want to just be trading hours for money. And so we're like, we want to build solutions that we wish we had as a firm. And some of it was we created a very basic operating system that we called The Way that developed internally as we started to work with firms.
And we sell that as a book for $9.97 on Amazon, and it's 40 some odd pages. And it's a basically a simplified version of EOS that really works, in my opinion, for law firms, especially those under 10 million in revenue, which are the vast majority of law firms.
And we started building things like we built our community and for a long time, Darl, I didn't have any money. And so we made it a dollar a day and it's like you get access to me, my team, people ask questions, you connect, we do webinars, we do ask me anythings. So we're adding that and we have 366 people in that community. We started building online classes. Brittany does a mastermind for marketing directors across the country.
So it's information about that where I think she's probably got about 15 or so marketing directors from around the country working with her. Allison has an operations mastermind and it's all virtual and it's really at accessible price points. And then what we do is we take on a limited number of private clients who we work with on a very deep level. We work with them on their biggest opportunities in their firms.
We look at their firm holistically. We try to meet them where they are, try to take them where they want to go and leave them with clarity action and offering them the supports to implement it.
So that's sort of what we call our bespoke work, and that's how it is that we operate because I didn't want to grow a consulting firm where it was like I'm getting people in and then I'm turning them over to people I don't know who know what they're doing. And it's like for us, it's like we've all been in these roles, I know this stuff, and I'd rather limit my client base in that way and offer products that other people can access essentially easily.
Darl:
So you've written a lot on AI and I've seen, that's obviously a hot topic. That's like the buzzword in the legal industry is AI, and I've seen a lot of your posts on it and they're super thoughtful and very in depth.
And that's one of the really big things I want to talk to you about on this podcast because I think anytime somebody's looking at the end of the year, there's always the predictions for next year and what's going to happen and what are the latest trends. Tell us a little bit, just generally speaking, how you view AI and its current capabilities and uses and the best uses for lawyers.
Ryan:
I think why I love this so much is my mom growing up before I was born was a legal secretary, which we would now call a paralegal. And my dad was in the Army and he trained in telecommunications. And so he worked with computers. He actually worked for what is now Brooks Brothers. And when individual stores would have to report back to main office way back in the '80s and '90s before the internet, there was intranet. And so that's what he worked. So I've always had a big interest in legal technology and how it impacts work and what the possibilities are. Because as you know, Darl, it's really hard to be a lawyer. It's really hard to run a law firm and do it really well. And you don't have that many leverage points that you can push. And especially if you're in personal injury, you can't just raise your rates and costs go up every year.
And so one of the only ways to create leverage, especially in personal injury, is through the use of technology. And so much of what has gone on, I would say honestly, before ChatGPT comes out, which is just three years ago, that the sort of biggest development in legal tech was email.
Somebody could look back on the first 30 years of legal tech and say, email was the big-
Darl:
Ryan, I got to tell you this story. So I started law school at Mercer down in Macon, Georgia in 2004. Westlaw and Lexus were a thing, but they were kind of new. Integrating that into the law school curriculum, we started out, they showed us how to do the good old-fashioned research with the books and how to shepherdize cases and all this.
And we're like, why are we learning this? And it was almost like the Westlaw and the Lexus was kind of an afterthought, but it's like that's the centerpiece. And then working in law school with lawyers who had been practicing since the 70s, 80s, when they were on typewriters, the technology differences are just, it's crazy. And I think a lot of it's generational. I think you see, I like to think of myself as a younger lawyer, but there's certainly younger lawyers than me.
I'm in my early 40s, but lawyers in their mid-twenties, they didn't come up in a world. So to your point about the AI, and I think you may see this in your own kind of consulting practice, older lawyers might be a little more apprehensive about adopting it, right?
Ryan:
There's interesting things that we can talk about this because I think it's a good place to talk a little bit because there's sort of interesting things that I think happen with AI, one of which is to me very corporate-driven, which is hallucinations.
So it's like buy our specific product because hallucinations and hallucinations are getting lawyers sanctioned. And look, to me, that's just sloppy legal work. That's not necessarily a deficiency in AI. I mean, you have to understand how large language models work a little bit, which is, yes, it's weighted predictive analysis, but a lot of tools like Westlaw or Google Gemini, for example, give you the ability to verify and read those links at this point.
And I have not come across... I come across things when I do research in Westlaw, because as a professor, they give me access to full boat Lexus and Westlaw, which is really kind of cool.
And some of the AI summaries, yeah, it's inaccurate, but sometimes headnotes were inaccurate too. You have to, as a lawyer, if you're going to rely on something, use your critical thinking skills, apply contextual-based knowledge and dig into what it actually means and read the damn case and read the medical records.
And I think there's that sort of driven fear, but I also think there's an equal heart fear in this space as to why people don't use it, that they, in some level, realize that it's better than them at a lot of things. And being a lawyer is identity-driven.
And in law school, a lot of our grades were based on our ability to write briefs, our ability to make law review, our ability to get the clerkship. And you start looking at some of these models that are out now, like Claude's Opus 4.5, it writes better than almost any lawyer I've ever met.
The writing is that good. And so when paralegals sometimes look at these products too, it's like, oh, it does the MedCron better than me. And I think that there's a realization and a fear and a resistance that can come from this that leads to a lack of adoption or willing ignorance, kind of like, "Hey, if I just ignore that, it's going to go away."
Darl:
You're head in the sand. Yeah. Yeah. Well, this is a great time, I think, to talk about some of the tools that are out there, because I think this is the common misconception people have.
When you hear people talk about AI, some people just think of ChatGPT, they don't even know that Claude's out there. Some people that might be on Filevine think of, oh, Filevine's AI feature. And we're on Filevine and we can talk about that too. But one of the programs I use a ton is Google LM Notebook or NotebookLM, whatever it's called.
Ryan:
Amazing.
Darl:
Uploading depositions to it and having ongoing dialogues with it. What did so- and-so say about this? And I'll give you a perfect example. We had a summary judgment motion to respond to recently, and I uploaded all the depositions in the case, and I uploaded the discovery responses, I uploaded the complaint, the answer, and it was great to, one, just get a summary.
You can do a prompt, "Hey, summarize this and tell me these categories of information." But as you're working through the brief and you get an idea and you're like, "Huh, that would be a really good fact to cite. Where was that? " Just go into the chat thing, "Hey, what did so- and-so say about this? " Boom, there's the citation. Now, of course, you want to verify that that's actually what they said, but it's easy. You click the citation, the box comes up.
That's just one example. And I think people also, talking about the hesitancy to adopt AI, I think some of the hesitancy might come from what I would call the snake oil salesmen that are out there that are putting their label and upcharging things that's really probably no different than ChatGPT.
And so let's talk about some of the tools that exist out there and how lawyers can use them. And happy to start wherever you want, if it's with the popular ones like ChatGPT or Claude...
Ryan:
Okay, let's actually start with NotebookLM. And I'm just going to get ... What's really fun for me at this point is as a consultant, I get to have dinners and meetings with people who are way smarter than me at this, who are at the forefront of what all this stuff is and what it means. And I just get to think about it, which is to me just a ton of fun.
And so a tool, it's essential for lawyers to basically understand what different tools are and what is out there. And so NotebookLM is probably the most useful tool for most lawyers that is available. And it is available if you have a Google Workspace account, you use Gmail or whatever, and if you don't have one, it's $14 a month, so you can get it. It's completely affordable. And so what Google NotebookLM is what makes it unique is that it has an incredibly large context window.
Google is, to me, the absolute leader in AI because of the amount of information that it can process and contextualize. And it's also, it's been able to be trained on YouTube. So there's more ... Every single day, the entire Netflix library is added up in data into YouTube forever and ever in terms of videos. So there's a lot of language that those models have been trained on.
And what NotebookLM is also is there's something called a RAG. It's like retrieval augmented generation. I may be tired, something like that, but RAG. And what that is essentially saying is RAG is simply saying it's going to limit what its universe is to what you give it. So if you're using just ChatGPT and you're like, Hey, summarize this deposition, it may be going out on the internet and reading some Morgan & Morgan blog posts for you.
You don't know where it's pulling from. And so it's not hallucinating, but it's just the wrong tool for the job. It's trying to hammer in a nail with a screwdriver is essentially it. So NotebookLM is like, "Hey, I'm going to confine my world to the depositions that you give me and I'm going to go really deep into that. "
And so look, and that's just one sort of commercially available example. If you're just chatting with ChatGPT or Claude or Gemini, you're using probably about 10% of what it can actually do for you. And you should just know a few basic things about that. One of which is that, I mean, Google just announced that they have solved this problem, which is it's going to be mind-blowing what happens in 2026 because of this. Basically, when you give a chat right now, everything is new to the LLM.
Everything is like, they've never seen this before. Now, Google has solved an ever-evolving memory issue, so it's not anymore like the LLMs are going to be looked ... And we say LLMs, it's just large language models, most common stuff being Google, Gemini, ChatGPT, and Claude.
And it's going to not just say, oh, this is the case, and I'm going to go outline and read it. It's going to have read the whole law library, and it's going to be able to tell you that case in context of everything else, which is really with the job. That's what real deep legal analysis is.
It's understanding where a case is not a universe unto itself. It needs to be contextually specific to be valuable for advocacy. And so you need to understand essentially different things. For right now, it's useful to think of chats as people in a way. And if you start giving people too much, they get overwhelmed in the work quality declines. So breaking things up into smaller tasks is incredibly useful.
So if you're just like, "Hey, write this brief for me, " that's not really useful. If you say to it, "Here, I want you to analyze these five different cases on standards for summary judgment and summarize them so I could put it in a three-paragraph part of my brief to give the judge the legal standard, it will do that okay." You know what I'm saying?
Darl:
Well, I'll tell you, we talked about NotebookLM, and I used that for the cases too, not to find it for me because it doesn't. It's reading the stuff that you've uploaded, but it's easy to go into Westlaw and just put in a String site to download all the cases that the opposing party cited, download it, upload each case up to Google LMNotebook, or NotebookLM, and then ask it to summarize each case and then interact with it.
But also it's helpful for when you find cases on your own in Westlaw or Lexus or whatever program you use to download them and upload them. And I had a number of cases on different issues that were relevant to this briefing, and I was just able to interact with it throughout. It was 50 plus cases, I think, that I had in there. It still took a long time to write the brief, but it took a lot less than it would have without it. And I think my ultimate work product was probably better with that tool.
Ryan:
Sure. And you're using it in the right way, which is really like, look, when we do legal research and we do projects, you have to be asking yourself a series of questions. "Well, what about this? What about that? How does this impact this? Okay, I got this new piece of information. Does this change this? "
And also, Darl, I'm a wear-my-heart-on-my-sleeve kind of guy, and I get into something and I become a true believer in it. I just don't have the gear in my body to be objective when I'm being an advocate, I think.
And some of it is you can ask NotebookLM, what are some weaknesses in this argument? Based upon these things, what questions do you think a judge may have after reading this? What questions should I prepare for for advocacy? And you could start seeing it.
Part of when you work with AI is orienting it to what it is. I want you to be a judge who is reading this brief, who has read these three cases. What questions would you ask based upon this brief or these briefs?
So you're orienting it to the audience, you're saying," I want you to ask it questions I'm a lawyer to prepare for. What are the gaps? What are the strengths and weaknesses? "And it can be very good at being neutral and objective.
And look, one of my big things was, and it still is for PI practice, is doing focus groups and focus groups are incredibly valuable to do. You should be doing them if you're not, but it's also incredibly valuable to be able to focus group your legal arguments, but that can be challenging.
Lawyers are busy people. "Hey, Darl, can you read these five cases and read these briefs and help me prepare?" I mean, that's a big ask and instead you can do that through NotebookLM very easily.
Darl:
That's an interesting use that I had not thought of. I'll have to check that out next time I've got something like that going on. So NotebookLM, that's a commercially available product for everybody and has uses across a variety of industries. So does Claude, so does ChatGPT.
What are some of the legal-specific tools that you think lawyers really need to be aware of that could bring a lot of efficiencies to their practice?
Ryan:
One of the big things here for lawyers to be thinking about, okay, there's a recent McKinsey study that came out, I think in September of 2025 that really said that, look, despite all the money that's being invested in AI, firms are not receiving gains from it.
And this isn't like law firms. These are Fortune 100 companies, these are much bigger companies because of one-off tools like It improves efficiency a little bit. And what you had mentioned, Darl, which is you had to research the cases, you had to download this, you had to upload this, you had to ask this, you had to do this.
And in part, it's like you don't have the data in a way that's organized, you have to think through it, you have to think through the structure of it, you have to give it input. And so basically what they say is that that kind of usage of AI, which is far beyond putting something in ChatGPT and saying, make this read better or help me contextualize this or summarize this, that leads to about seven to 8% gains in efficiency, which is nice, but it's not actually saving versus the cost of implementing and the cost of some of these tools that can be very expensive.
But it says that firms and companies that are putting it at the core of what they do receive gains of 70% to 80%. And so one of the problems with AI usage or just using something like ChatGPT is that it doesn't have all the information it needs. It doesn't have all the cases, it doesn't have all the notes. Things may be structured in a way that are not useful to it and may be in formats that are not useful to it.
And so if you're putting ... The most successful firms at this in the next year are going to be firms that put AI at the center, they create, first of all, a single system of truth. And look, that could be Clio, it could be Filevine, it could be something else, but they have all their information that gets poured in a way that the system understands it and everything is there.
So it's not just the medical records, it's the medical records, the phone calls, the intake notes, the text messages from the client, the pictures, the images, and then you're going to need AI to understand the context. What is a pleading? What is a complaint? What is this? And that's a more nuanced, refined version of AI. And once AI has those things and as it gets a lot better, and what you're looking at right now ... Sorry, I love this stuff.
What you're looking at, just to give you, there's a bunch of metrics out there regarding AI, but my favorite is called Humanity's Last Exam, and it's basically all these expert level ... It's 2,500 PhD-level questions.
The average human would not do very well on this at all. And so you have to think about it asking questions to a history PhD, an English PhD, a physics PhD, 2,500 questions. And the scores on Humanity’s Last Exam last year at this time were two, 2.7. Basically, AI was able to do right around what you would get from guesswork, just pure guessing.
So if you give me the test and I just filled in the B bubble, it's roughly how it would do. This year, they've got it up to 37.5% on some of the most recent metrics with Gemini-3. Next year, it will be over 90, which is considered passing humanity's last exam. So the underlying tech is about to get a lot better.
Darl:
So I have a concern about that. Let me ask you about that. So I'm not aware of any ... Are there any programs or features within FileVine's AI program or Clio's AI program that can write briefs for you?
Ryan:
No, not yet.
Darl:
Okay. So here's a concern I have. Let's say that I've got my briefs and my case uploaded to Filevine. Is Filevine training its AI on my briefwriting and using that work product to benefit other lawyers?
Ryan:
You have to look at what their current terms of services are.
Darl:
Which I've never read.
Ryan:
And my belief is based upon just what I have heard and read is Filevine has a zero data retention policy, which is one of the things that you want to look at.
However, Filevine is obviously very into the legal world, and so they can understand the relationship between a complaint and a deposition, and it can contextualize what these things are. What their actual data sources are for that are probably a combination of aggregated data and human programming into it, frankly.
So you want to look at what those things are. Just for a security perspective, to simplify this for your audience, what you really want to look for is something that is at least SOC2 compliant. And SOC2 is a very hard certification to obtain. It's an independent audit regarding best practices of security and data management. You also want to look at something that is HIPAA-compliant for that.
You want to look at things that have zero data retention, which basically means that the AI does not take your material outside of your environment and it deletes it essentially because you have to think of it as there's your data and then there's essentially things that are exporting your data, and then there's AI, and then there's your data coming back in.
And essentially the data does not ... It gets anonymized to the AI and then reprocessed. So those are the basic things. And I mean, for lawyers, and I'm not a security expert, caveat, caveat, these big companies, whether it's Filevine or Clio or SmartAdvocate or Supio or Eve, they are up to what the security levels needed for people to, I think, practice fully ethically. I don't think that there's a dispute about it. It's where people get into problems is just feeding things into ChatGPT.
That should not be. You should be probably using an enterprise version of ChatGPT if you're using HIPAA data and healthcare information and things like that.
But yes, those companies have clients that are much, much bigger than the small firms that maybe you and I are used to who have very trained security people who have very high standards of compliance, including government agencies and businesses.
Darl:
So I love Filevine. I've been on it for over five years now, be six years next year, and it's got a lot of great features.
We signed up for some of the AI features on Filevine, and I think Filevine is something you're very familiar with, right?
Ryan:
Yes.
Darl:
Yeah. What are some of the AI ... If you were talking to a law firm owner right now and they were getting Filevine and they were kind of on the fence of, okay, I don't know if I want to get AI, pay for the AI features.
What are the ones where you'd be like, man, these things are really going to make a huge difference in your practice if you use them?
Ryan:
I mean, look, the best uses of any sort of technology are automating and using AI and the things that you do the most. So look, in my firm, we got to trial on maybe one out of a hundred, one out of 125 cases that we'd taken. Just our good cases settle, policy limits, and sometimes they just wouldn't.
So I wouldn't be using AI to really hone in on your closing arguments because it's an infrequent thing. That's not the thing I'm going to build. But where AI can really lend itself to is things like medical records extraction, where it's extracting the codes, the ICP codes.
It's really the ability to summarize medical bills, the ability to read the bills, the ability to spot things like priors, the ability to identify missing records. That is very, very useful. Filevine also has a feature where it can use AI to say, you can't change the phase unless you have this information.
So for example, I forget what they call it, but you can't move it from the intake phase to the pre-lit phase unless you have the insurance policy information.
Darl:
Interesting.
Ryan:
So you're forcing your team to essentially comply, you're forcing an audit and a compliance on them in a way that's really useful because you need to have that data and you need to have it entered in a way that makes sense. And if it's getting moved through the process without you having the data, that is a problem.
Where it's about to get a lot better in the next year, for sure, is the ability to use your templates and your documents and your structured data in ways to be generative. That's to me with the underlying advances in it and with what they're building.
Darl:
Can you give me an example of that? Is that drafting a complaint?
Ryan:
Correct. Correct. I mean, because right now Filevine has had their document generation, which is basically mail merge. I mean, that's 30-plus-year-old technology, which is, okay, you have information in this field, you do it.
And that's one of those things where most of the time a lot of lawyers can probably copy and paste an old complaint and look at a screen and get that done in- It's just as fast. ... 20 minutes. Yeah. Right. And so the notion that what will happen in FileVine is as the medical records come in, as the police records come in, the complaint will be automatically drafted and draft.
So the lawyer can look at the complaint and say, okay, so here's what it is based upon what we have. Do we want to finalize this? Do we want to check this? Where are their inconsistencies in these records?
Does one record say that it was their left arm and does this other record say that it's a right arm? These kinds of sorts of inconsistencies.
So where that's going to go, because it has all the data and it can look at your intake notes, your text messages, your phone call notes, is it can see those inconsistencies, it can see those consistencies, it can point out things that are strong and where it's going to go more from, it's going to go ultimately in a way where AI acts as very assistive technology.
So you can think about it as really not just like, "Hey, can you summarize these meds?" "Hey, I've summarized these meds for you. "You could imagine a really good paralegal. I've summarized these meds for you and there's a problem. There's a prior that our client didn't tell us about and we've requested those records.
That's where this is all going to go. So I think for firms out there, whether or not they like the AI, first of all, look where it's all going is in two, three years, AI is just going to be how you work. It's not going to be an app, it's not going to be a button.
It's just like that will just be what Filevine or Clio does. And so I think creating a clean source of data for your firms, putting in all the data into a system like Filevine or Litify or SmartAdvocate or Clio or MyCase, whatever, one of the big ones, getting that organized and watching what happens with AI in the next year to two is going to be really, those firms are going to be really far ahead.
Darl:
So we talked a little bit about this before the podcast about training people, and the more technology there is, the more important it is, I think, to have somebody who's training people, but also it's not just a one-time thing. It's a constant continual thing of these are the new features, this is what you can do, and staying on top of that, how should firms do that?
I mean, obviously one option is to hire Best Era to come in and help with that. But if you've got a firm that's a small plaintiff's firm, let's say they're 15 employees and under, how does somebody stay on top of all this stuff?
Ryan:
Yeah, I mean, okay. So look, this is all moving so fast. And Darl, every week I get somebody contacts me about some company that I've never heard from that solves a problem that I barely think exists. And so what I think is one of it is just learn about the products. First of all, if you don't have an AI policy, you have an AI policy, and that AI policy is your paralegals are dumping personal information into their personal ChatGPT.
So you have to have a policy that says, you can only use this. And if you use Google Workspace, which I highly recommend is just fantastic for AI at $14 a month, and you're like, "Hey, you can use our paid Gemini subscription," you've come out ahead. And we're not getting into complex things like building agents and stuff like that on this call, but I would then really look and say, okay, my firm uses Filevine, what are the AI features in Filevine?
And they do product announcements and webinars and confine your universe to what is going on with those products. Or if your firm is using something like Supio or Eve, look at their product announcements, attend their webinars, attend the functionality of them because a lot of times firms don't, they buy it and they have some implementation and then it just doesn't get used or built.
These are ongoing investments in both development of product and training and adapting to new technologies. So if you're going to be a firm owner and you're going to pay for these costs of these products, none of them that we're talking about are cheap, I mean, Google Workspace aside, but Filevine or Supio or Eve, these are not inexpensive products.
Really invest in ongoing monthly training on them. Really, one of the things that we recommend is find out who your champions are. Your champions are in this.
Darl:
I love that word.
Ryan:
And really empower them to be the educators and the drivers. So you've got 10 paralegals, one or two are not going to hate AI.
One or two are going to be like, oh my gosh, this is the greatest thing. Take the ones who are the greatest thing, give them access to training, give them access to webinars, let them learn the product, let them make suggestions, and have monthly meetings on it as a team. Here's what's new, here's how to use it, and also monitor and encourage usage of it.
So if you've bought into Supio like, "Hey, to the extent you have KPIs or something, your KPI is to use Supio for 10 demands this quarter and see who's using it. " Just very basic things like that.
Darl:
Have you used the Filevine AI demands feature?
Ryan:
Yes. Yep. We were the very first original beta client of that.
Darl:
So I'll tell you one of the ... And I don't know how it's pulling information, but I've tried a number of different medical chronology programs. And again, I love Filevine, but even within that MedCron program and some other third-party ones I've used, I get a lot of errors. And it's not hallucinations, but it's, I think, errors and interpretation of the data because for example, certain forms might have a lot of dates on them.
They might have the patient's date of birth. They might have the patient's first date, whatever, and it gets put in the wrong places in the chronology. Is that something that you see as a short-term problem that AI is going to figure out in the near future, or do you see that being a consistent ongoing issue?
Ryan:
Short-term problem. And this is really what ... It's Google Titan project for those who are out there. This is what the biggest limiting problem with AI has been because it's seeing the information anew.
And so what I mean by that is it's looking at the MedCron, but it's never seen the date of birth before. But when you look at how this is going to evolve, and look, all these products use combinations of Claude and ChatGPT and Gemini and whatnot.
So it's not like Filevine has its original AI. They're using these systems. And as this develops, it's going to say the AI in Filevine is going to be like, no, no, I've seen their birthdate, it's 12/19/1974, and now I'm seeing it on this medical record again, and I know that's not the date of service. And that's where you're going to see in 2026 this get a lot better.
Okay, interesting. I would be shocked in two to three years if we're talking about hallucinations the way we are now.
Darl:
Interesting, which by the way, it's the first time somebody got caught hallucinating in a brief, submitting a brief with hallucinating case citations, I thought, oh, well, now it's going to be out there. Everybody's going to know and it's never going to happen again.
It is happening all the time. And I have no idea why lawyers haven't gotten the message of, "Just check your case cites. That's it."
Ryan:
And part of it is too, just understanding what it is. And so AI at this point is not a finished product and you should feel comfort in that because if it really gets to being a finished product, oh, what is our usefulness?
Darl:
Thank you for that. That makes me feel very good.
Ryan:
But look, one of the rules of life, pretty principle, delegation, if something can be 80% good, then it's worth doing. And honestly, look, believe it or not, I can make mistakes, Sara, and good paralegals can make mistakes too. And most of the time it doesn't matter.
And AI is at this point no different, but really the AI functionality becomes useful is if it gets you 80% there. And if it gets you 80% understanding the meds, and yeah, you've got to spend time clicking. It doesn't make sense, you're ahead.
Darl:
So what is the difference between generative AI and agentic? Did I say that right? Agentic AI?
Ryan:
Yeah.
Darl:
So that's something I've seen a lot about is agentic AI. What is that?
Ryan:
Okay. Agentic AI is, they say 2025 is the year of agentic AI, and yes, and 2026 is going to be the year really of agentic AI. So agentic AI, okay, so large language models are just that. It is trained on all the things ever written on the internet. And so agents are sort of, you have to think of them as very specific people almost, like specialists. So an LLM is a generalist, an agent is a specialist.
And so if I'm building out, say I'm file finding AI, I'm building out an agent whose job it is to have a background in understanding all sorts of medical literature and also reading medical records. That's going to be that agent's purpose.
I'm going to give it information designed to do that purpose. I'm going to have an agent that does medical billing, we'll say, and it's going to be very good at understanding billing codes and what those means. Those are going to be separate people on my team. They're separate team members.
It's like I have an offensive lineman and I have a place kicker and I have a running back and I have a wide receiver. And so the advantage of that is that you build specialists who are trained in very narrow things. And then what happens is you build an agent that manages those things.
So it's like, okay, we can call it the paralegal agent who manages the medical records, the medical bills, the medical chronology, the demand, da, da, da, da. And so what that agent does is works with the other agents and tells it, "Do this, now do this. " And then it says, "Hey, agent A just did this. Agent B, now you do your thing, and agent C, you put it together."
And so what it's doing is it's taking skills from different programs or different softwares, and it's putting them together in ways that are specific. And you can create agents really easy. I'm using make.com or Zapier because what Claude developed was cool, which is a language, model context protocol, MCP, and allows these things to talk to each other and do these things. And so you could have, say, for example, a very simple agent that one agent reads all your pleadings.
Say you get into court notices. I'll do court notices. One agent reads all your court notices. So it monitors your email and its whole job is to find and read court notices. It identifies those court notices, and then you have another agent who is trained in here's the format of these court notices and extracts the dates that you need from the court notices.
And then you have a third agent who then takes the data from agent two and puts it in your calendar and automatically puts it in your calendaring system. And so that's an example of a very simple agent build that does this. So agentic AI is, yes, that's a very big thing in legal because if you think of agents as people, you understand that the people who are really good at accounting aren't the very best creative storytellers, and you build a team of people with diverse skills.
Darl:
So before we wrap up, I've got a few more things to cover. What does this mean for paralegals and legal assistants in the future? What do you see their role?
Ryan:
Yeah, I mean, their role is really going to be, in my opinion, look, AI is going to get really good at inputting data, processing that data and extracting that data.That is going to get where AI is really good at. So you could imagine a situation where it can request medical records, it can process the medical records, it can tell you what it is, and it can suggest things to do.
So maybe this is, for example, you get a PI case in and the AI can go out and request the police report, can read the police report, can analyze it, and then it can say, "Look, do we want to put this case into suit? Do we want to drop this case? Do we want to refer this out? Do we want to do a demand on this? " Somebody is going to have to be making those calls and that person is going to be a human.
Look, I can imagine some situation where there is no more work, but that's not going to be the immediate future because AI is not going to have the context and judgment necessary to understand all the dynamics at play to actually make those calls.
And I think there is that. And I think that there is client service. We are in a very human profession with human needs and human concerns. And so I think the people who can, paralegals who can understand AI, understand how to leverage AI, embrace AI, and who can work on their skills regarding judgment recommendations and client service are going to be where a lot of this is going to be.
So to me, the work is going to be the work one paralegal should be able to handle more than they're currently able to handle, and that will in fact be more valuable and it will also just create different work.
There's all sorts of weird pro se litigation going on because they have access to ChatGPT and there'll be more work.
Darl:
Last question. What are your predictions for the short-term, the next year to two years of AI and what people need to be aware of?
Ryan:
They need to be aware of. And the thing that I think people underestimate is the speed at which this is all happening. And that is the reoccurring theme from various leaders that I talk to is that this is happening so fast. And the key is to, one, knowing that, accepting that, and two, really adopting your organization and thinking about what it means.
And I know, Darl, you shine through on this with your values. Okay, this is all happening and fine. Are there going to be direct to consumer claims settlement AI bots? A hundred percent that there are. How do I position my team in a way that is different? What is my value proposition? Who am I serving? What does the structure of my firm look like going forward? What am I really going to need? What am I going to want to hire for?
Because hiring the nurse paralegal may not be the best use of your dollars. AI may be able to read every medical journal and every record and be very good at that. But I think what law firms need to think about is how are they going to teach judgment and experience?
Because those are the things to me that are going to be the differentiators. It's like, look, it's no different. I can watch a YouTube video on how to fly a plane, but you can't get me in the cockpit of an A320 and expect me to land the damn thing.
And I think don't have that false sense of confidence, have that humility and think about what your value proposition is and think about how you are going to be adjusting your workflows, which I'm going to suggest to you are going to be ... I mean, AI is going to happen in the background.
I can tell you, Filevine will request medical records, analyze medical records, and make recommendations to you in ways that come at you. Your activity feed is not going to be there. It's going to be very, very different in the very.
Darl:
I could do without the activity feed. That's my number one.
Ryan:
Well, you will because what will happen, Darl, is you can imagine a universe where it's like an assistant and it calls you in the morning and you have your morning meeting and it's like, "Hey, we got the police report in on the Rodriguez case yesterday and it cites our client.
Our client said that this was not his fault. The cops were wrong. What would you like me to do? Do you want me to set up a call with him?" "Yeah, set up a call and we'll talk to Mr. Rodriguez about what this means. Would you like me to send it to him in advance of the call? ""Yes, I would. " "Okay. On the Jones case, look, there's a $50,000 policy. We got the original ER bill and it's $42,000. Do you want me to send over the ER bill to the adjuster and try to close this case?"
"Yeah, send over the ER bill." So this is where this stuff is going to be going in the very, very foreseeable future, whether it's next year or the year after. And so sort of just pause and think about what that means for your organization and what it means for your ability to serve clients. And I think you'll be okay.
Darl:
Fascinating and terrifying at the same time at times. I think anybody that hears these things, it's going to get a little ancy because change is antsy, makes people antsy and nervous, but it is fascinating. And I think the information you have and your insight is incredibly valuable. So I appreciate you joining us today to talk about this.
Before we close, tell us how can our listeners find you? If they want to reach out and talk to you about ways BestAera might be able to help them with their firm?
Ryan:
Yeah, LinkedIn. I'm on LinkedIn. Slide into my DMs and if you've got something for me, you want to work with me, you want to talk to me, you have a question, answer it.
If you're some vendor who's listening to this and you want to sell me leads, I'm going to just respond, "I already have a girlfriend," because I don't appreciate the volume of spam, but that's the easiest and best way to get ahold of me.
Darl:
Understood. Understood. Well, thank you, Ryan. This has been very entertaining, so I appreciate your time.
Ryan:
Thank you.
Darl:
You can follow us on Spotify, Apple Podcasts, and YouTube so that you can see and hear our latest updates when they go live.