Maximum Lawyer

Watch the YouTube version of this episode HERE

Are you a law firm owner looking to change how you run your business? In this episode of the Maximum Lawyer Podcast, Mathew Kerbis, a lawyer and founder of Subscription Attorney discusses how AI is transforming legal work and why the traditional billable hour model is becoming obsolete. 

Mathew talks about the framework for using AI effectively within the legal space. It is important to remember that AI tools, like ChatGPT, are not calculators. They have biases and are reinforced by the humans who designed them. They are also not perfect and should be used as an aid. For the legal space, AI should be used to give you all the information before giving you an answer.

Mathew delves into the topic of the billable hour model and why firms should move to subscription based models. The billable hour includes doing a bunch of tasks for a client within a set time frame for a price. If a client only pays you for one hour, you are only working for that hour. But switching to a subscription based model with AI in mind means you can scale your business better. You can develop better relationships with clients because there is predictable revenue.

Listen in to learn more!


  • 4:38 The Latent Legal Market Opportunity
  • 9:11 Framework for Using AI Effectively
  • 13:25 Retrieval Augmented Generation & Tool Selection
  • 16:07 AI in Legal Practice
  • 19:10 The End of the Billable Hour & Subscription Benefits\


Tune in to today’s episode and checkout the full show notes here. 


Connect with Mathew:

Creators and Guests

Host
Tyson Mutrux
Tyson is the founder of Mutrux Firm Injury Lawyers and the co-founder of Maximum Lawyer.

What is Maximum Lawyer?

Maximum Lawyer is the podcast for law firm owners who want to scale with intention and build a business that works for their life.

Hosted by Tyson Mutrux, each weekly episode features candid conversations with law firm owners, business experts, and industry leaders sharing real strategies and lessons learned in the trenches.

If you're ready to grow your firm with less stress and more support, this is your next must listen. Subscribe today.

Mathew Kerbis 00:00:00 All right, all right, all right. Thank you. Thank you very much. Thanks to maximum Lawyer. Thanks to Max Lacon and Tyson and Becca and everyone in the the team. When I was in the middle of the pandemic, hating billing by the hour, tracking my life in six minute increments. And I was thinking, gosh, there's got to be a better way. And I started to think about the subscription model. But how do I start a firm maximum lawyer, the podcast, the Facebook community was there for me, so thank you so much. So who here has heard of the latent legal market? Let's see. A show of hands which is like impossible to see with the lights. But anybody anyone. All right. So just a little bit about me before I get into the presentation, if you go to subscription attorney, that's my law firm website. You'll see all my prices, all my subscriber benefits. And if you scroll to the bottom, I have hyperlinked my engagement agreement.

Mathew Kerbis 00:00:48 I put this out there for my clients, but I also put it out there for other lawyers who may want to do this the same way or a similar way that I do it. And I do have some copycats out there, and I don't mind copycats because of this idea of the latent legal market and the opportunities for law firms to do that now. My podcast is law, so I have a lot of free content on there about the ins and outs, on the nuances and how to make decisions on how to price your services. And I interview other lawyers about how they're doing it in their practice. That is not today's conversation. So if you do want to see those actual specifics, I highly encourage you to check that out. So AI is changing the game. If you are billing by the hour, you're already probably a little bit worried about it. Legal trends reports from Clio from last year. I think we get the new trend reports in another week have showed that 75% of total firm billable work, including lawyers and admins combined and staff and paralegals.

Mathew Kerbis 00:01:40 75% of that is being automated by AI. Automated. That's not reduced billable hours. Those are gone billable hours. So this is an important conversation that we're going to have today. So first I want to tell you a little story about a client, a real client of mine. But I changed your name and I changed your business. It's a little subtle, so I'm not sure if you could see it out. So let's call her Sally. And let's say she sells seashells by the seashore. Okay? And she's really good at it. She's good at business. In fact, she's so good. She took over the business from her aunt. It's a family business, and she doesn't know anything about the legal stuff. Her aunt's attorney did all that legal paperwork years ago. But she gets this employee, you know, she's good. Business is going well. She's selling tons of seashells. And this part time employee that she hires. Well, it turns out her dad's a lawyer. And my client, you know, we'll call her Sally.

Mathew Kerbis 00:02:38 She, she has a conversation with this part time employee. And it turns out this lawyer, dad says you don't sign this employment agreement. It has restrictive covenants in their non-compete non solicitations, and you're not being paid enough money to enforce this. It's actually contrary to current Illinois law. This happened in 2023. Okay. So like right sort of as AI was, you know, still kind of early days. Right. So she reached out to the attorney that her aunt used, and it turns out he was dead. And so she didn't have anyone to go to to ask questions about. Well, like, what am I supposed to do right now? She's she's on the younger side. She's in her mid 30s, but she's running this business. You know, it's it's doing better than when her aunt had it and she doesn't know where to go. So at the time she turns to Google. Right. GPT search wasn't around yet. Perplexity wasn't around yet. And she's searching around and she talks to lawyers and they give her free consultations.

Mathew Kerbis 00:03:38 Right. Free. Except it's not really legal advice, because in order to get legal advice, she needs to pay them hundreds of dollars an hour. Well, she's running a business. She needs pricing certainty. Well, she did get lucky, and she found me. Now, I do advertise myself as an affordable attorney. My subscription started $20 a month. My clients pay $50 a page, and back then my pricing scheme was a little bit different. But right now there's a 500, 1000, and 2000 a month level based on how many legal projects my clients need per month. But she found me, and so my pricing is transparent. She was able to budget and understand what it was going to cost to hire an attorney. It's not that she doesn't have the money, it's that she needs to know what it's going to cost. She is an example of this massive market of opportunity. So let's talk about this market opportunity. According to Grandview Research, in 2024, almost $400 billion was spent on legal services in the United States.

Mathew Kerbis 00:04:38 400 billion. And if you're if you have a closed mindset, you think, get me that, get me my slice of that 400 billion and we're competing for that 400 billion. Except that according to the World Justice Project, 77% of legal needs go unmet by lawyers in the United States. That's actually 2019 data. That's the most reliable data I could find. And that gap is actually probably wider now. Sometimes this is called the Access to justice gap. But not everyone qualifies for legal aid. So I think of it as this latent legal market opportunity. Right now we're lawyers. So maybe math is not our strong suit. So to do a little bit of extrapolation there, if 23% of the market is $400 billion, that means there's a $1.3 trillion market opportunity. While I do do some handholding and some coaching for law firms, I'm a practitioner first like to to teach them how to if they need some hand-holding to teach them how to do this. I give away all the information for free on my podcast, and it's speaking events because I'm not worried about competition.

Mathew Kerbis 00:05:52 In fact, a lawyer in my backyard in Illinois just started copying my business model, even with similar pricing. He's my backup lawyer now for my firm. If something were to go wrong because he gets it and there are more than enough clients in this latent market in Illinois that I only really need 40 good clients, right? To have a sustainable business now, they're not going to stay subscribed forever. You don't stop marketing your firm, but it changes the nature of the game. So even if you're not with me on this latent legal market opportunity, although I think it's really important to think of it, let's just look at the reality of AI ten our tasks automated with AI. And I'm going to teach you some. I don't have a lot of time. I'm going to teach you some tricks of how to get good output from these AI tools. But if you use the AI tools, right, ten hours becomes ten minutes. And what really what is really happening is what used to take ten hours worth of work.

Mathew Kerbis 00:06:52 And I have an example of real world example that happened this year that I'll give you at the end of the presentation. So pay attention, of of how this happened is you get the answer in two minutes, but then you analyze the output, you go back and forth with the AI tools and then and you do that for eight minutes. And so you actually have better work product in a faster time. So if you're building $500 an hour, again, simple math, ten hours of work is $5,000. So 10.0 to 10 minutes billing in six minute increments point two $5,000 becomes $100. What the heck are you supposed to do? Do you raise your rates? Well, let's look at that. You would have to do a 50 x times your rate or charge $25,000 an hour. If you bill by the hour, ask yourself if you really needed to. In an emergency, could you afford your own billable rate right now at whatever it is? How about if it's $25,000 an hour? You know, by the way, most lawyers I know are part of the latent legal market, right? Like, could we afford our own billable rates if we really had to? And it was an emergency.

Mathew Kerbis 00:08:08 Think about that. What does that say for access to justice? What does that say for being able to serve this latent legal market that already needs pricing certainty? Yeah, I got you the thing in ten minutes, and but you're still paying $5,000. Well, there's just something that doesn't sit right. How about our legal ethics bodies in Illinois, where I practice? It's our state supreme court. Would they deem that a reasonable fee? You know, no legal ethics advice given. Everything is for educational purposes only. But it's probably not. Probably not ethical, right? Probably not a reasonable fee. So to get this kind of efficiency gain this 10 hours to 10 minutes, I'm going to give you a framework that no matter what AI solution you're in, and I have my, like, holy trinity of AI tools that I like to use. I'll tell you how I use one of them soon, but you got to think of this AI as a person, right? Like, what would you do with a person? Let's just pretend that whatever AI tool you like to use, it's the smartest person on the planet.

Mathew Kerbis 00:09:11 They've read everything on the internet, right? And and they're really smart, but they're an entry level worker. This is their first day on the job. And this is their first job, right? What would you do? You'd you'd give them a ton of context. Right. And if you were at the event last night, you know, we don't call it prompt engineering anymore. It's context engineering, right. If this is again, think of this as being a real human right or even like a first year associate. You know, if you tell them, summarize this deposition and it's their first day on the job, and it's the first day as a lawyer or a paralegal or whatever, like what do you what kind of work product are you going to get from that? You'd give them context. You'd explain what you're looking for, what you need. Maybe you'd give them an example of another deposition summary and what that looks like. Or here's an example of motion for summary judgment. Or here's a form master services agreement.

Mathew Kerbis 00:10:15 Right. You'd give them as much context as you can as you know. Again, smartest person on the planet, right. They're going to figure it out if you give them a lot of context. And then if they give you work product back and it's not perfect, it's their first day on the job. Entry level workers. Super smart right? Do you say nope. It's trash. You're fired. No, no. You'd work with them. You'd iterate with them. You'd say, okay, you got some things here. You did some things there. Now, realistically, that's going to take a lot of time with a real human with AI. You have the output in seconds or minutes, right? Some now more complex tasks. These deep reasoning tools, like they take a little bit longer to get results, but you still analyze the output and you work back and forth and you go back and forth with it. Right. So this is the time to finish. Product has completely shrunk. You'd also not trust this assistant as a source of truth, right? Just because you've read everything on the internet doesn't mean that, like, you could perfectly regurgitate what you've read.

Mathew Kerbis 00:11:22 And also, if you're doing legal work, maybe what they read wasn't statutes, case law, you know, treatises, contracts. Right. So when you ask ChatGPT to do legal research, I don't think anyone here in this room does that. But if you did, like, that's what you're doing. Of course you're going to get made up stuff. Also, you know, they're not well-adjusted humans, right? So the framework I want to give you to think about this or remember this is that AI is not a calculator. And when I say AI, I mean large language model based generative AI, right? Not not machine learning, logic based AI, but the AI tools, the ChatGPT or Claude's or Gemini's. That's what I mean when I say AI. It's not a calculator. Two plus two equals four is an immutable truth. An objective truth. You program a computer or a calculator. With that truth, it will always get it right. Language is all made up. We made it up right.

Mathew Kerbis 00:12:25 If I were to spell out the word l I v e and asked you to say it, some would say live. Some would say live. Context matters. Right? So since language is all made up, and these large language model based generative AI tools are answering based on Probability. Right. What is the next? What's the word that's most likely to come next based on this massive training set of data, but also reinforcement data. Thumbs up. Thumbs down. Reinforcement training. Right. And so they're all also going to have their own biases, which is one of the reasons why I like to use multiple AI tools when I'm doing work. Not just one, because they're all I don't know what their training set is. I don't know what they're what the reinforcement training is. So I'm going to use multiple tools at the same time. And they're all relatively affordable. So it's probabilistic. Right. So you can't rely on probabilistic tools as a source of truth on their own either, which is one of the reasons why I want to introduce you to one more concept quickly.

Mathew Kerbis 00:13:25 I'm including this chart. I can't explain all the charts at the time I have, but essentially what we want with our AI tools is we want them to retrieve information before giving us an answer. I do not use AI tools if they're only answering me based off of training data. Just like the smartest assistant on the planet, it's not reliable enough for the work I do as a lawyer, so it needs to use what's called retrieval augmented generation. It's just jargon for the AI industry. But basically what it boils down to is it may or may not be trained on the sources of data. It's better if it is right. Just like if you have an open you imagine you're being tested on something. You're in law school, you have your your your law school exam and you didn't study, but it's an open book exam. Well, when ChatGPT when you give documents to ChatGPT or notebook or these other AI tools, you're in a top right category or your yeah, your your top right category, like you'll do.

Mathew Kerbis 00:14:19 Okay. And if you're the smartest person on the planet, you'll do pretty well. But if you actually got to study the material, if you have an industry specific tool or large language model, you studied it and it's an open book exam, then you're going to be more accurate. In fact, when I find in my use of these tools, they almost never hallucinate. Hallucinations happen, people make mistakes. The AI makes far fewer mistakes than humans, right? We still have to check the output. And some of these tools are built to make it. So you could check the source right away. I only use tools that give me citations to everything, and not necessarily case citations, but citations to the sources that it's drawing from using this type of architecture. Here's another example that I want to give you. And when people who think they could use ChatGPT for everything, imagine you hire a contractor to do some work at your home and they only bring over manual tools, screwdrivers, saw, and you're paying them by the hour and you're like, well, I, I'm a savvy consumer of construction services.

Mathew Kerbis 00:15:19 Like, I know power tools exist. Where are your power tools? I'm not paying you by the hour unless you use power tools. Construction guy goes, okay, okay. Contractor leaves, comes back with just a power drill. And it's like you're going to you're going to build me a new bathroom with just a power drill. It's like, don't worry about it. Puts out a two by four. You know, long piece of wood takes the drill and just drills sequential holes in the wood. Right? And then they're close enough. And then when he gets to the end, he uses the butt of the power drill to just smack it down in half. Right. It's like a rough edge or whatever. But hey, he saw that in half. Except it took way longer than it would to even use like a manual saw. Right. When we use one AI tool to think it does all the legal work, like legal research like ChatGPT wasn't built for that. We're using the wrong tool for the job.

Mathew Kerbis 00:16:07 We need purpose built tools for what we do. And just like the contractor can't use one power tool for that job, we need to use multiple tools to get the job done as lawyers for the work that we do. And when we do this, we're more efficiently able to deliver legal services, allowing us to price differently to serve that latent legal market. So here's a real world example of how I use a tool called notebook LLM Pro. All right. If you're in Google Workspace you already have access to this. If you're paying $14 a month or more, or you could just get access for $20 a month to the Gemini tools. Client is paying me $1,000 a month. He's at my business plus subscription level. Franchise owner calls me, schedules a 15 minute call because I don't pick up the phone. My clients schedule calls with me even though I don't bill for for time. I protect my time. I put boundaries around my time. My clients schedule calls with me. So we schedule the 15 minute call, which he's entitled to.

Mathew Kerbis 00:17:05 He gets three projects a month, and he calls me and he says, hey, Corbis. Because my my clients call me corpus, and he's like, so I already booked a booth at an event that's outside my geographic region where I'm licensed to sell for my franchise. Am I allowed to market outside of that region? I said, that's a great question. Let's take a look. So I pulled up his 600 plus page franchise disclosure document for 2025. I read it back in like 2018 when he first became a client. They updated every year and I through. I'm familiar with it, but I don't know what keywords to do. Keyword search for. To find this. I don't know where this is in this document. It's really been years since I've had to tell him exactly what he can or can't do based off it. Throw it into notebook LM in less than a minute. It's, you know, analyzed, vectorized. You know, my backgrounds in philosophy and law. I'm very good at using these tools.

Mathew Kerbis 00:17:58 I couldn't explain them too technically. Right. But in any event, after less than a minute, I'm able to query against this document, this 600 page document using semantic search, searching for the meaning of the words, not just the keywords. So I say I am an attorney, my client is the franchisee, and the included source is a franchise disclosure document. And then I explained what my client wants to do market outside his region. Is that permitted? And so in seconds it comes back and says it's probably allowed. I mean, it was like more of like a two paragraph explanation, but concluded it's probably allowed, but I click on the citation inside of notebook and it opens up the document and scrolls immediately to the relevant paragraph. Because I'm a lawyer, I want to read the document. I want to go to the source. So I read the relevant paragraph identify right away. Oh yeah, it does allow that. So I could advise my client accordingly. He's thrilled not just that he's allowed to do it because he already spent the money, but that I was able to give him an answer within the first ten minutes of a 15 minute call.

Mathew Kerbis 00:19:10 How much time would it have taken to analyze a 600 page document and write a formal opinion, legal opinion for this client, right? How the heck am I supposed to build for that? My clients. I'm making $1,000 from this client that month. Right. And he's a client. He's been a client for years, even before I used the subscription model. This is why the era of the billable hour is over. And if I still didn't have you yet. Right. In the last couple of minutes here I have. Let me let me appeal to your legal brain and and desire to comply with the rules. Right. So according to the model rules. Comment 1.5. Comment five. Most jurisdictions have a version of this. It ends with a lawyer should not exploit a fee arrangement based primarily on hourly charges by using wasteful procedures. Well guess what? If you're billing by the hour and not using AI, you're probably in violation of your rules of professional conduct. Right. But when you use these tools and you ditch the billable hour and you leverage the subscription model, aligning incentives between you and your client, and you're just think about the lives of your staff and your employees and your associates when they no longer need to track their lives in six minute increments.

Mathew Kerbis 00:20:27 Right. Your your retention is probably going to go up even regardless of AI. I've been teaching this since before AI hit the scene. The subscription model lets you scale your firm without more hiring. And now with AI, like it's only like 100 x more, right? So there's predictable revenue for your practice. You can actually develop real relationships with your clients. They are not afraid to call you. I consider my $20 a month marketing my $20 month subscription a marketing earn rather than a marketing spend. It just gets some access. That's it. It gets some access. And so who are they going to call when anything legal happens? Or who are they going to call before something bad happens? Me. And so I enter the legal problem avoidance space instead of just the legal problem solving space. You just have to charge differently for it. So pretend for you that your bar association, your state supreme court, bans the billable hour, but they give you one year you have 365 days. What do you do? Plan for it starting today.

Mathew Kerbis 00:21:36 And what I'll leave you with is I've been talking about this for years. I've been doing it for years, but it simply hasn't been enough. And so I'm announcing something. This is a max law exclusive Tyson, for everybody who's here. Something that I've just hinted at on LinkedIn is that I've actually taken it to the next step. And I'm starting a company called practice that's going to help lawyers do this. But here's the thing. Whether or not you use this new solution that I'm working on, or you have your own solution, homegrown solution, or non-legal tech solution to ditch the billable hour and use the subscription model. I hope you're convinced today that you want to tackle this latent legal, latent legal market opportunity of all these clients that have money to pay but need pricing transparency that you say goodbye to the billable hour and hello to subscriptions. Thank you.