AI Practice Notes by Socius Legal

The Lawyer-AI Sandwich: Defining the Human-in-the-Loop Architecture

• Michael Fahner | Legal AI & Tech Strategy for Socius Legal • Season 1 • Episode 2

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 43:02

How do you translate years of legal intuition into a digital product? In this episode of AI Practice Notes, host Michael Fahner talks with Marla Miller, a tax attorney and the Founder/CEO of 9to5 Legal Docs, about the transition from practitioner to builder.

Marla shares her insights on why generic AI models often miss the mark in high-stakes legal work. They dive deep into the "Lawyer-AI Sandwich"—an architectural framework where the lawyer provides the initial context and the final strategic review, leaving the "messy middle" of drafting and data synthesis to the AI.

Whether you're a skeptic or an early adopter, this conversation is all about how to stay "in the loop" while still capturing the massive efficiency gains of modern technology.

Episode Highlights & Timestamps

  • [00:00] – Introduction: Meet Marla Miller, tax attorney, LLM, and CEO of 9to5 Legal Docs.
  • [02:15] – The Logic of Tax: Why a background in tax law is the perfect primer for building logical AI systems.
  • [06:45] – Beyond Generic: Why generic probability isn't enough for legal work and how attorney-vetted expertise changes the output.
  • [11:30] – The Lawyer-AI Sandwich: Defining the Human-in-the-Loop architecture—human context on the front end, AI in the middle, and human judgment on the back end.
  • [16:00] – The "Review" Myth: Reframing the pushback that "reviewing AI takes too long" and identifying where the true efficiency gains live.
  • [23:45] – The Future of Legalese: Will AI lead to more concise contracts or a flood of AI-generated text?
  • [29:30] – Mentorship in the AI Era: How junior associates can use AI to accelerate pattern recognition rather than just automating grunt work.
  • [35:30] – The Business of Law: A candid look at the rise of "AI-Native" venture-backed law firms vs. practitioner-led innovation.

Key Quotes

"AI will do what you want it to do... but it doesn't understand why that's actually bad for you in six months or a year down the road." — Marla Miller

"The context window for a human lawyer is just bigger right now. We need to be the curators of the content." — Michael Fahner

About Our Guest

Marla Miller is a tax attorney and the Founder/CEO of 9to5 Legal Docs, a platform designed to provide attorney-vetted AI tools for startups and small firms.


Connect with Socius Legal

Follow for more on the future of legal work.

SPEAKER_00

Welcome to AI Practice Notes, podcast about the future of legal work and the lawyers who are shaping it. I'm your host, Michael Fonner, practicing attorney and the founder of Socius Legal. When we talk about legal AI, there's often a gap between engineers building the tools and the lawyers who are actually using them. This is something we're keenly focused on here at Socius. And we've always believed that the most impactful legal technology is going to come from practitioners. Our guest today shares that builder lawyer mindset. Marla Miller is a tax attorney with an LLM from Northwestern and a full-spectrum legal background. International tax in DC, big four advisory work, and time as outside general counsel for a venture-backed startup company. Today she's also the chair of the tax section for the Louisiana State Bar. But Marla is here because she's taken that deep domain expertise and turned it into a product. As the founder and CEO of Nine to Five Legal Docs, she's building attorney-vetted AI designed to handle the heavy lifting of the deal cycle without removing the lawyer's judgment from the equation. She isn't just theorizing about the way AI might change the profession. She's actually architecting tools that are doing it. Today we're going to dig in on practice versus product, why better workflows aren't enough to fix a broken model, codifying intuition, the architecture of translating legal judgment into responsible AI, and the lawyer in the loop, why the next decade of legal tech belongs to the builders who actually understand the risk. Marla, welcome to AI Practice Notes.

SPEAKER_01

Thank you. It's good to be here.

SPEAKER_00

Great to have you. Um, so I I would love to start by getting more background on you as an attorney and what led you to the legal profession to begin with.

SPEAKER_01

Oh, I would love to have a great answer for I floundered a little bit in undergrad. I started as a pre-med major and then was like, what am I doing? I don't like blood that much. You know, this is right for me. I switched to psychology and I realized I was too analytical. I was analyzing every human I came in contact with. And I took an aptitude test, and it was law or business. It was a very in-depth one. And I chose a law degree because I felt like I could do more with it.

SPEAKER_00

I don't know that that's like you're doing a lot, so that's great.

SPEAKER_01

Um what that was in particular. I got to law well, I got to law school, and I didn't know what kind of law I wanted to do. And um, our assistant dean was a tax attorney, and I took her tax class, and something just clicked. It wasn't the numbers um because I'm not a CPA or anything like that. It was just the logic problem part of it, essentially, and figuring out the rules or the parameters and how best to navigate them. That's the best way I know how to explain how my brain works and why it clicks that way. And I feel like tax is at the core of a lot of business decisions, right? Like you at least have to consider it when you're doing transactions or structuring or putting agreements in place, how it will impact you from a tax perspective.

SPEAKER_00

Yeah, I think taking income tax in law school was one of the more eye-opening uh educational experiences throughout law school for me. It totally changed my perspective on what tax even is, it's how it regulates things.

SPEAKER_01

Yeah.

SPEAKER_00

I often say to non-lawyers in conversation that the smartest lawyers are oftentimes the tax attorneys. So, with all due respect to everyone else in the profession, I I have a lot of respect for your specialty in particular.

SPEAKER_01

Well, I appreciate that. I think I feel similarly about securities attorneys.

SPEAKER_00

Well, that's great. But who I feel you know the the most uh particularly impressed by are those who are building legal AI tools. So that's um our topic of conversation today. And for me, when I first used generative AI a couple years ago, it was mind-blowing experience, however you want to describe it. I immediately felt gripped by it in the sense that it was going to change what we do as attorneys. And I started tinkering and dove in. I'm curious if you had a similar kind of, I don't want to say epiphany moment, but if you had a similar experience, if there was a point where you were like, I need to dig further into this.

SPEAKER_01

Yeah, I started several years ago trying to find a solution for the bottleneck of my practice and disseminate information. And this was pre-like AI becoming so um accessible. And I just put up a website with forms and information. That was the beginning of it. And then as AI became more accessible, I'm like, no, there's better ways to do this, right? And um, so that's really how I got started into it. But yes, it was very mind-blowing once you get into it, and even the little things that it could automate and do and um was or take, you know, used to you had to prepare all this information to or all these forms or all these articles, and here you could feed the AI some of that that you've already prepared, and it just produces a very accurate and great summary. So little things like that, you just start to learn, you know, how efficient it can become in the everyday lawyer life, I guess.

SPEAKER_00

Yeah, uh that resonates with me. I I use AI pretty much every day in my practice, and over time I've kind of come up with my mode of operating and how I use it. And as someone who loves research, likes digging in on things, it's just an amazing tool for its ability to distill things down. But beyond kind of like the individual efficiencies, uh, we're very interested in kind of the systemic efficiencies at law firms and in legal practice generally. So I'd love to hear a little bit more about nine to five legal docs and exactly what kind of problems you guys are focusing on, and how there's so many legal tech tools out there, a lot of talk about all this stuff. How do you guys differentiate from the the generic AI and some of the other legal AI tools out there?

SPEAKER_01

Yeah, so at nine nine to five, um, I think I started out wanting to have a drafting tool. And then I was like, no, we need something bigger. And as part of this, I really wanted to help startups. And um, they're I feel like they're in more startups now with vibe coding, right? And they just lack certain basic information and understanding of legal workflows. And so we developed the AI in the first context just to disseminate information, like try to cross that barrier, you know, and help educate founders based on lawyers' training, because generic tools are probabilities, right? I guess you could say. And it's different from when a lawyer trains it based on their expertise, experience, patterns, all the things, right? And so that was the first, I guess, um, version of our AI. And then nine to five, you can get documentation, you can ask AI questions, but once you make a purchase on the back end, you have like an account and it's more of a systems tool. You can do eSign, it it's the AI is contextual, so it helps educate you through the documentation process, but also it remembers you where you were with your documentation, gives you tips in that sense. And so it's really designed at its core to educate, but then become on the back end more contextual and assistant-like in the process. Um because there's so many areas, so many small details that can really trip startups up, right? And change the legal risk profile of what they are doing.

SPEAKER_00

Yeah, no, absolutely. So if if I'm a lawyer out there wherever listening, um how would I get involved with 95 Legal Docs? Do I I sign up? Is there a download or how does that process work?

SPEAKER_01

Yeah, so our next step is to have lawyers on the platform. So right now it's been in beta testing using me as a lawyer, because our tool on the back end, we want lawyers to be able to edit the documentation for the startups within the platform. So it stays all within the platform, not because we're being selfish. I mean, that would be great if everything was done within nine to five, but because the system also populates a corporate record for them and really helps teach that how to do that and how to create a data room. So if a lawyer can go in and really work within the platform on some on a base that they already have, um, it really eliminates the risk of um leakage, mistakes, all things that happen when there's all these disjointed tools and all these disjointed systems. So uh any lawyer can reach out to me about being on the platform. And and then the next layer will be white labeling to lawyers to utilize in their own practice. So I really believe that I don't know that all law firms will become AI enabled, but I think so. We will have to adapt or die, right? And so that's the goal is really to make AI tools accessible to solo and small firms.

SPEAKER_00

Yeah, that that's awesome. And it I mean, certainly resonates with me. I you know I use Socius in my practice, and I'm similarly constantly iterating on it, testing it, and kind of being the primary uh you know, hamster or whatever guinea pig. But I I I appreciate that approach. I think there's, like I said, a lot of legal technology companies out there who rush to market and are maybe are um not as well tested or you know, worked over with practicing attorneys. It's not necessarily about volume, but it's about that kind of constant feedback loop to improve the product. I think it's I think the whole profession is in beta when it comes to AI, and we can figure out how we're gonna handle this going forward. I mean, there's gonna be regulatory changes, it's gonna be obviously the technology is changing rapidly. So um I I have more of a maybe a long-term vision of that. I also have been very focused on this idea of like a an empty middle where you have the off-the-shelf tools, co-pilot. Every lawyer has Microsoft Office, right? It's available, it's really good, it's getting better. And these tech companies keep leapfrogging each other. Maybe Gemini is the top model right now, whatever it might be. That's always on one side, that's always available. And then you have kind of the custom tools and the tools that have been iterated on the way you're talking about. And it's that middleware where it's kind of like a skin over ChatGDP or whatever it might be. That's I think where legal tech maybe uh falls off the rails. And I'm I'm most focused on folks like you who are doing that long-term experimentation to build out these more bespoke tools that really dig into what it means to practice as a lawyer. I mean, when I was a law student or before I became a lawyer, I really didn't know what the day-to-day of a lawyer looked like. So it's hard for me to expect some, you know, someone who's not worked in it to understand it either.

SPEAKER_01

Yeah, I agree. Um I think I it's maybe a really bad analogy, but I analogize these um non-lawyer trained, non-lawyer-bedded AI tools to kind of like us in the law schools. We're like, you know, we're trained on the law in theory, but it's not like practicing, right? And we don't we can't make those judgment calls. We can't, we don't have that pattern recognition until we've been in the trenches for a while. And I think it's very similar in that in the um the gap between the tools, right? And then also non-lawyers building legal tech is tricky to me. Um, it doesn't mean that there can't be a really good tool out there. There are, I know there are, but I think it's different when a lawyer builds because engineers are like, oh, I can do this, right? Which is cool. Lawyers are like, okay, what do we need? Okay, can we do that? It's kind of a different in my mind at least. It's a different thought process and a different vetting process and a different building process. And really, even on our um platform, there's two customers, right? We have the attorney and then we have the startup founder. And we can build something a certain way for the startup founder. But I just had this, we're in this now, figuring this out for my engineer. But lawyers need it like this. And yes, that's more difficult, but we have to do it, right? And so it's figuring out those things, and sometimes it's kind of competing interest. Um, can we do this quick enough, cheap enough, fast enough this way? But is it going to work, like you said, for the long term for lawyers? And um, you really have to take the time to figure that out.

SPEAKER_00

Yeah, I I think in addition to the quality of the technology or or the tools, it it's a change management issue on behalf of the lawyers as well. I in my experience, lawyers are are generally kind of slow to adopt new technology. I think even if you put AI aside, probably most law offices across the country were behind on what you know a tech company might be using internally when it comes to tools. AI has been maybe a little bit of a kick in the ass for lawyers, right? And now we're we're digging in uh and leveraging tech more. Um, but getting people comfortable with it is a huge portion of it, and making it so that they don't really have to lawyers tend to be sent in their ways, maybe. Maybe everyone tends to be sent in their ways, but lawyers are people, so they they kind of stick to what they know. And when the tools are designed for them, I think they're more likely to adopt it than when it's kind of like a labeled legal AI tool that's maybe trained on some legal data sets, but isn't as like ergonomic for them.

SPEAKER_01

I agree. Like the goal is for it to fit into their workflow as seamlessly as possible, or maybe change it a little bit, but not totally upend it, right? And I do think um some lawyers take an all-or-nothing approach and are standoffish to AI because they think it's gonna they have to change everything they've wrong, everything they've ever done, but that actually is not true, um, if done correctly.

SPEAKER_00

Yeah, one of one of the criticisms I hear a lot that I think makes lawyers standoffish is well, if I have to review it anyway, why you know why am I doing this? I I don't really buy that argument, but I see it a lot and I hear it a lot. And as a result of that, I I when I'm designing work streams or AI work streams, I like to think about how do we make the output easier for a lawyer to review? How do we make the output something that the lawyer um can check in an efficient way so that they don't feel that? But I'm curious, have you heard that pushback as well? It it's facially inaccurate to me, but have you seen people say that?

SPEAKER_01

Yes, yes, I have. And it it's review or do this piece of it anyway, or what insert whatever clause there. But yes, it's true. And I don't think they realize how much it can speed up processes and help synthesize information that they already have or prepare first of things or do these initial things. And what I really realize, so I guess when we went to market, I thought, oh, they will see that this is lawyer trained and come to us instead of going to like Claude or GPT or whatever. And what I found is they really want instead of a like AI on the front end and lawyers being totally on the back end, it's actually AI in the middle. I don't know if you're finding this too with what you're doing, but they break that down. You need the human lawyer on the front end because they want to ask questions, they want to be reassured. And lawyers, that's where we come in, is because we've seen so much. Well, if you're experienced, you eventually get there. You've seen so much, and you can take what they're telling you and recognize patterns or follow-up questions or things like that, right? And it's all judgment, this human judgment. And the founders, the business owners, they actually want that. And I think it's part just human connection with all the AI out there, and then also maybe just reassurance or hey, they like to ask, Am I not thinking of certain things? Right. Things like kind of open-ended questions and problem solving. And then on the back, and then that's also where strategy comes in. So then it's like, okay, this is what we're gonna do, this is how we're gonna get it done. And then you utilize the AI tools the best you can to accomplish that, or in the you know, whatever tools will work, and then you come in, do final reviews, strategy and on the back end, right? And and then meet with them again. The human interaction is important. Go over what's been done or what needs to be done, and so forth. So I really see that AI is in the messy middle, more so than it is on the front end, or at least should be, right? And so um, yeah, we're pivoting, we've had to pivot a little bit because of that.

SPEAKER_00

Yeah, I think that's a great way to frame it. Uh, this kind of lawyer AI sandwich. I mean, certainly on the back end, a lawyer should just always be reviewing an AI output. I think that that's just an immutable law of legal AI. On the front end, though, you know, even as the AI models get increasingly better, someone still needs to translate the context to the AI. And I, in so many ways, I feel like the lawyer's job, even without AI, is understanding the context and making an assessment of the situation in that context to provide the appropriate advice. I mean, so often a client will come to me and what they think we're going to discuss is not what we actually need to discuss, right? It's like some other issue is the actual lurking thing or how to approach it. So when I think about maybe someone who's not a lawyer using AI to solve their legal problems, I worry that they are starting off just in the totally wrong context or they immediately go down kind of an incorrect lane. Maybe I just tell myself this to give me comfort that like lawyers will always be valuable as AI gets incredibly powerful. But I think you do need kind of a curator of that content, certainly today.

SPEAKER_01

Yeah, I think this is true. I mean, if you think of it even like a teacher, right? Like you can learn a lot from these programs, but a teacher has context, teacher has explanation, a teacher I can give you examples, and that's just on the teacher's educational front. But the lawyer brings the experience too, right? And like you said, like I had I have clients that come to me and they're like, okay, we're gonna do this transaction to accomplish X, whatever X is. And I'm like, but have you thought about Y and Z, you know? And like, yeah, we can get there, but it's gonna cost you this, right? And right now I'm thinking of a tax situation, but it could apply to anything because they don't understand how one thing changes the other. And it's not faulting anyone, it's just they haven't been trained on it and they haven't seen it. And that's been our role, right? And that's where we can really continue to add value.

SPEAKER_00

Yeah, I I noticed that very recently I had an experience with um kind of a client who was using AI to maybe re review my work or to review the contract we were working on. And the AI got it wrong because it was unable to connect the dots between an earlier provision and the indemnity clause. And sometimes, you know, AI at this stage of the game gets lost in the sauce a lot, right? The context window for a human lawyer is just bigger right now, um, or at least more capable in this respect. I don't know about how many tokens a human processes, but um, you know, we we we still need to kind of be there to guide it. So you run into those issues, but I I've increasingly seen AI drafted contracts coming in from my lawyer, uh from my clients, rather, AI drafted questions. AI, you know, I had a call right before this where someone was clearly reading like the AI generated what you should ask your lawyer about this issue. I view this as a good thing, actually. I think in a more educated client is a good thing, and I'm I'm ready to answer those questions. But I'm curious if you're seeing that kind of as well.

SPEAKER_01

Yes, definitely. I mean, a couple of years ago, it was more basic. Like a client would send me something that they developed with Chat GPT. Oh, look at this equity incentive program that our package, you know, this whole thing. And it had like stock in there and all of this in there and LLC. I mean, it's just things like that. And with nine to five, what I'm seeing is they are still utilizing some of these other programs to make sure we're not missing anything, right? And then when we do consultations and go over it, it is very much like. They have something that they want in the contract, and the AI is like, this is how you get it. This is a provision. And then my job is to say why we don't put that in the stage, right? And so, yeah, AI will do what you want it to do. We'll give you the provisions that you want it to give you, but you don't understand why that's actually bad for you, you know, or why it can harm you in six months, a year down the road. So, yeah, a lot of that.

SPEAKER_00

No, the AI's desire to please can oftentimes um, I mean, as lawyers, sometimes we have to say no, or maybe we're a little bit of an impediment to what our client is trying to do that's part of the job. The AI is not really fulfilling that function uh in the same way and can be very direct. I I've definitely seen provisions of just kind of straight up saying what the client must have requested it to.

SPEAKER_02

Yeah, exactly.

SPEAKER_00

On the flip side of that, I also see often, especially in like maybe like the miscellaneous section of a contract where I maybe the AI is running out of tokens or something, and it'll just have kind of like one sentence like governing law, or it'll be like no electronic signatures are permitted under this agreement or something, instead of kind of like our boilerplate that we would include. And I I do wonder if there's maybe a certain beauty to that long term where maybe we can start to cut down on on legalese to these extremely concise sentences, but I I don't think we're we're quite there yet. I don't think the law's maybe uh caught up with what the future of written word might look like in that regard.

SPEAKER_01

Yeah, you know, I don't I've don't know that I've thought about it in that way, but I do think there are ways where we can cut down a lot, right? And it used to be like this competition of like, look how much I can do for you. Look at this beautiful contract, look at this super complicated LLC operating area, right? And they would pay for that because oh wow, that must be amazing. And now it's like, wait, we can get this done in so many less words and so much less competition. Um, so much less construction. But I guess the other side of that is you have to have two people that are willing to do that. If you still have an attorney or a party on the other side that wants all of this, you know, legalese because it makes them feel more secure, then you, you know, how do you cut down on it on the other on your side? So yeah, I think we're going, it would be great to go towards that, but yeah, we have a ways to go, I think.

SPEAKER_00

Yeah, and there's there's a good chance it ends up going totally the other way, too, because it's so much easier to generate text. So it's kind of like, well, why not just generate a huge provision? Because on the other side, we're just gonna have the AI read the output, you know, so we don't really need to go through it ourselves. I mean, maybe this is the future. Certainly to be clear, like no practicing attorney should be relying solely on the AI, kind of reviewing or or writing this stuff now. But you know, long term I could see a world where the AI takes maybe more of a uh a lead role in in the drafting if it becomes materially better in certain respects. But that judgment in context we talked about is always gonna be you know, I think require a human.

SPEAKER_01

Yeah, I agreed. I very much agree. We don't know what it's we don't know, you know.

SPEAKER_00

Yeah, it's kind of the beauty of it in some ways. It's all you know, it's also a reason why you know I love the builder-lawyer attitude you have, because like I said earlier, it's like the whole thing is in beta, right? We we don't know where the profession's going. I think it's just incumbent upon us as lawyers to pay attention right now and see what happens. For a lot of lawyers, the product is changing, you know, what you tell.

SPEAKER_01

Yeah, and I think it's changing so fast, right? Like for nine to five, I would think that we were doing one thing, and then I was like, nope, we need to do this thing. And so when we really launched the platform, I just designed it for flexibility because what we think is this amazing tool right now may not be so amazing in six months or even shorter, right? And it's moving so rapidly now that it's hard to keep up, and so it's really designed such that we can either build the tool or we can partner with other AI legal tech companies, and I think that's really important, the flexibility because we don't know where things are going.

SPEAKER_00

Yeah, absolutely, absolutely, and that that partnership model I think makes good sense too, because there's so many people experimenting and different tools are gonna emerge for different things. And it I it's that's why we're having this conversation, right? Like we're both building legal AI tools, but I don't, you know, it we're all kind of in this together to a certain extent, um, to try to figure out what the law looks like going forward.

SPEAKER_01

Yeah.

SPEAKER_00

Legal practice at least. One thing that surprised me about like other lawyers when I started building legal tech and like talking with them and learning their needs was how different different lawyers practice and how different different institutions are within the law. Is that something that you feel has been noticeable to you, or because of your kind of diverse background across the legal industry, did you feel like you were already prepared for that?

SPEAKER_01

Yeah. I don't think it's been surprising. And I think maybe because I've had I've been in so many different roles. I mean, I've been in big four in-house multinationals. I've worked for a really small firm that was not on the cloud. There's just a big spectrum of things. And I think um, so yeah, that's not really surprising to me at all. And but I do think that that also goes to the flexibility of what we're what you're building, right? And how we need to build it such that it fits different workflows or pieces of it would fit different workflows as much as possible. We're never gonna get it completely right. Um because and then lawyers, like you said, maybe don't love change all the time. Um, very not technologically advanced historically, and we're being forced to with AI, but it's some more experienced lawyers, older generations, maybe may not adopt AI and maybe retire before they do, you know. So we just have to, and then the ones coming out of law school, they won't know a legal world, which is crazy to me without AI, right? So we have very different sides of the spectrum we have to consider.

SPEAKER_00

Yeah, I I thought uh kind of in the post-COVID era things really changed the way we work with Zoom, like conference calls were kind of now Zoom calls and like just the dynamics shifted a lot. But with AI, I could only imagine how different the perspective of a junior associate or a first-year associate at a firm must be and will be for the foreseeable future.

SPEAKER_02

Yeah.

SPEAKER_00

Do you do you think at all about um how your tools or any AI tools will impact kind of the development and training of these lawyers? I a lot of people I think worry about that.

SPEAKER_01

Yeah, so our platform is built to disseminate information and to help people with forms and to remove barriers to access to legal in way in whatever ways it can. And so right now we're partnering with our first accelerator, um, and we're building that out, and it's based on that. And I think that we could go to like law firm small business clinics, right, or entrepreneurship clinics, or whatever the law firm calls it, and really try to help not only just provide access, but train the law students as well on best practices and giving them some technology that they could utilize. And we get feedback from that, right? I think AI, when it's doing its best, can bridge that gap and really leverage the knowledge experience of the old, I don't want to say older, more experienced surgeons and the less experienced regime and help both. And yeah.

SPEAKER_00

I I I think that makes that I think that's right. Like the there is obviously the existential risk maybe to the new lawyer who would normally do the kind of quote unquote grunt work, and AI kind of you know takes that roll over. But I think a a self-starter who's graduating law school right now has an amazing opportunity to do research with AI and learn and dig in on a concept and say, hey, go pull me everything on the internet about this. And you know, in some ways, you might be able to accelerate your career more than ever, although I think it requires like quite a bit of gumption and and and effort to do that.

SPEAKER_01

Yeah. Yeah. I I actually thought about that before this podcast, even how in coming out of law school, you know, when we're younger, we tend to sometimes be dismissive of old ways of doing things and like no, there's this more efficient way, you know. I mean, I'm guilty of that just living. But if they can find a way to make themselves more valuable because of the tech and because of their knowledge and the use of it, and then really helping bridge that gap for more experienced lawyers and leveraging the more experienced lawyers strategy, knowledge, pattern recognition, all the things. I I agree. I think they have a real opportunity to do that.

SPEAKER_00

Yeah. I I I at the same time, I like you know, pattern recognition in particular, kind of the repetitive nature of going or iterating over the same contract and reworking it. I I do think it builds up um your your eye in a way that maybe like raw research and knowledge doesn't do. Uh, but like as a lawyer, you kind of I I don't know how to totally describe it, but over the course of my career, I suddenly look at pieces of paper differently. And I I I wonder how that might change in kind of like an AI-powered review world. But maybe that's for the better.

SPEAKER_01

Yeah. I mean, we kind of had to learn in the wild, you know, like working for other people and really absorbing what they were doing and why they were doing it that way and all of those sorts of things. I think maybe AI has the ability to pick up those patterns and even teach those patterns in a different way.

SPEAKER_00

Yeah. At the same time, when we talk about kind of why they're doing something, I do worry that data sets at a lot of law firms don't necessarily capture that. They will oftentimes capture kind of the negotiated outcome of a deal or will, you know, kind of the compromise at the end. And even if you have multiple versions that context, you know, the client's risk profile or this particular event that was occurring, that, you know, or this meeting that you know had someone ANSI and resulted in X. That's kind of like institutional knowledge that has historically, I think, been passed down in like the mentorship, uh, you know, master apprentice kind of model that I think a lot of lawyers kind of have been part of. Um, do you think that that's something we should be worried about as we kind of build AI systems, maybe relying on the data bases in our law firms too much?

SPEAKER_01

I think yes, especially if those um databases, data sets are like you said, right? If I think it's gonna require lawyers, law firms to really take it a step further and identify why those decisions were made. And really, you know, you can do it in decision tree or some sort of other way mapping it out, right? Like, oh, it was this reasoning, this reasoning. I mean, I don't think everything can be deduced down to AI. There is human judgment in there, right? But it can start to pick up these patterns as to why things were done. And I do think it's possible with drafting, um, even with tax, right? Because we have these like flags or these things, oh, if we're doing this, we need to look at X, Y, and Z and make sure, you know, and triple check and do things. And I think those are all patterns that um, but yes, I think it's gonna require them to take a step further. But I think that's why there's like law firms like yours and these AI-enabled firms that are AI native that are hot right now, they're doing it in a different way. And you really have to go that extra step. But then on what they're yes, it requires that extra effort. And it also the recognition that there's value in that outside of the billable hour, right? To help in the long term. And I there's still a lot of firms that don't recognize that and um acknowledge it as the you know how we're going to have to.

SPEAKER_00

Yeah, reconceptualize value or maybe look at value with a new perspective. And you mentioned kind of some of these AI native firms. I I too have seen the the big trend with a lot of you know venture firms backing these kind of AI native law firms or AI-powered law firms. Curious to get your perspective on on kind of that model versus maybe a more traditional firm that is just leveraging AI and um also on the idea of like a venture-backed uh law firm. I I think that the nature of how a law firm might grow and develop client relationships might be at odds with that.

SPEAKER_01

Yeah, I think so too. And you may have better insights than I do, really. But I think it's a lot to unpack actually. Um should lawyers be raising capital and focused on returns when they are supposed to be providing legal services and counselor and advocate for their clients, you know? Um, I don't know that those two things align necessarily. And I do think that it is cool to see it be you know a law firm built with AI. Um, but does it have to be be venture back to do it? No. Um, are there barriers to access to AI for solo and small firms? Yes, at this current time. Do I think they'll always be there? No. And I do think that, like I said, there are some lawyers that will just be like, no, I'm not doing it. I mean, I've talked to some with they're like, no, I'm just too old, I don't care. I'll be out before I have to deal with it, right? But then the rest of us will have to adapt. And how much more value are they providing by building from scratch AI native, VC backed, as opposed to like putting the AI in the messy middle and really providing that extra value to clients? And I don't know that we can um put that into a you know data sets range.

SPEAKER_00

But I'm uh I I I have similar feelings about it. I mean, I'm sure there is a market for these kind of like you know, AI VC backed bang out NDAs kind of services. Like there's definitely a sector of the economy that will will benefit from that. But I I fundamentally think that if you take a good experienced quality attorney and just give them access to something like nine to five, they probably will perform better than they did because now they have these better tools. And I don't know if you need to kind of do the whole like native from the ground up stack at this point.

SPEAKER_01

Right. Yeah, because if you're doing from native from the stack, what is the AI learning from? You know, maybe it's lawyers that are bringing stuff into the firm, so maybe it's not fresh, but I've seen some where if the lawyers have any experience, it's not much. And so, really, what are they drawing upon when they're building these tools? I'm not saying it won't like you that they're not going to provide a value in a certain case, but I think we all have the ability to become AI enabled as these AI tools become better and works as well.

SPEAKER_00

Yeah, I also like I've certainly been exploring flat rate services and that kind of stuff more, but there's always something more, right? It's I need to do this formation. Oh, but also like my brother-in-law signed a note and I got to deal with that, or there's like always some extra thing that you know, AI can certainly help with, but as it stands today, you kind of need the lawyer to to exercise their their judgment on it. So I I think my point there is I I don't know if the technology is efficient enough yet to fundamentally change the firm structure, although it is certainly more efficient than maybe, you know, 10 years ago.

SPEAKER_01

Yeah, no, I agree. And I've seen that with nine to five with me as the guinea pig. Um, I'm like, okay, well, we have these documents, we're gonna do a consult. Yeah, I would um tailor these a little bit more customize these a little bit more specific to fit your fact pattern, but then there's always follow-up question, follow-up question, follow-up question, right? And so we're so it's even as a platform, we are shifting more to packages that involve lawyers in the process because it it inevitably happens, right? There's always facts or follow-up questions or follow-up, follow-up, follow up questions. Um always.

SPEAKER_00

No, absolutely. Um, and also if you're a founder starting a company or you know, or CEO or whatever, may even if you can do some of this work with AI, it's probably better to have someone dedicated to the legal function or someone who's kind of filtering the stuff so that you're not spending your time on you know that legal question. So I I think that uh lawyers are here to say for a while, which is uh I think a good thing for us. And I think that maybe there will be a time when the technology is dramatically better in a way that we all like the way we all do this is dramatically different. And we are maybe from the ground up AI native. But I think right now we're seeing just efficiency gains on the individual level, certainly, with you know, whatever off-the-shelf tools. And I think more and more folks like you and I are thinking about how you might design a system that's um creating more efficiencies in the whole work stream. Yeah, but that you know, fundamentally, um this is going to change. I mean, you've I'm sure you've seen, I mean, I know you've seen the dramatic changes over the last year in AI. I it's exciting, right? Like things you can now ask it to do makes your life a lot easier.

SPEAKER_01

Right. Yeah, and that's just the past year, and so it's compounding effects, right? So it's just gonna get that much better.

SPEAKER_00

Yeah. Are you guys at nine to five generally uh model agnostic, or do you focus on a particular kind of stack?

SPEAKER_01

Um, we're more anthropic focused, but um we do have multi um model built in um to the system.

SPEAKER_00

I uh I uh I kind of like started with anthropic. I feel like they have like a special place in maybe my heart. I feel like call it just kind of like really started, kickstarted my use of AI, especially in the real context. But um, so I I we'll see kind of where this race goes amongst all of them, too. You know, I mean it's gonna be I I suspect we will have years of this one leaprogging this one, and it will just keep getting better and better over time as they kind of compete. For the folks who are listening and have heard all this and are interested in in using nine to five, can can you tell us maybe where to find you guys?

SPEAKER_01

Yeah, so our website is nine to fivedocs.com. We're on almost all social media. Um, and then my personal social media is Marla Calvert Miller on LinkedIn, and then the lawyer Marla on TikTok and Instagram.

SPEAKER_00

So all right, excellent. Well, everyone be sure to follow those accounts or otherwise visit. Um, and if you're a lawyer who's not using AI, which I I think is probably the minority at this point, um you you should definitely get on it. Maybe that's a place the place to start.