In this episode, the focus is on drafting complex legal documents using AI after a deal structure required amending an existing LLC operating agreement to accommodate a self-directed IRA operator. The process involved transforming an attorney-prepared 10-page operating agreement into a 26-page document with tighter language and additional ancillary documents, all built through iterative AI collaboration rather than a single-pass output.
Key Takeaways:
- Section-by-Section Drafting Beats Full-Document Generation. Instead of asking AI to produce a 20-page legal agreement at once, breaking it into focused sections and comparing against existing attorney-prepared language produces dramatically tighter output.
- AI Legal Docs Without a Baseline Are Dangerous. Having an existing attorney-prepared operating agreement as a reference model was critical; without that tested foundation, confidence in the AI-generated output drops significantly.
- Context Retention Has Leveled Up. Current LLM models maintained focus across thousands of words of legal drafting in a single thread, eliminating the constant need to restart conversations and re-establish context that plagued workflows even six months ago.
- The $5K-$10K Attorney Savings Comes With a Catch. AI can replicate the broad legal provisions, but if you don’t understand the underlying dispute resolution implications and core questions you’re trying to answer, the output will have gaps that could cost far more later.
- First Lift, Heaviest Lift. Investing heavily in building reusable legal templates now (even across multiple days of work) creates compounding time savings on every future deal using the same structure.
Listen to the full episode for the detailed walkthrough of how iterative AI drafting, attorney-tested baselines, and internal review processes combine to produce sophisticated legal agreements at a fraction of traditional cost.
(Podcast transcript below)
Welcome to Get Serious, where at Serious Land Capital, we have funded over $6 million worth of vacant land deals with industry leading 41 % operating margins. So today I wanted to carry over the discussion from last week where I went over the ability to really think critically about deal structures and
you know, the requirements to overcome your own inertia and, you know, get past your own ego and so forth. when new data or, expertise that, is beyond your own, comes to light. so, you know, this discussion really centered around, our pursuit of entitlement related deals.
specifically in regard to funding soft costs and how this particular deal or a couple deals that we’re looking at, it got complicated because our typical debt operating loan structure backed by a limited recourse personal guarantee would not be effective if the operator who brought us this deal was using a self-directed IRA because you
have certain prohibitions against cross-collateralization of assets or personal guarantees. So it created a whole mess. so, the summary of that video is that we decided to utilize our existing special purpose vehicle equity structure that we’ve utilized in the past for one-off deals where we’ve brought in other investors. But this time we would join an existing LLC as an investor.
in, order to have more inherent protections in the case that the operator was using a self-directed IRA. So, you know, probably a number of you, including myself, even though as I talked through this, like just gets wildly complicated really quickly. but the takeaway is, you know, some of these situations require really diving deep and figuring out, Hey, is there potential for us to strike a deal here? We have a solid operator. We have a solid deal.
what are the requirements for getting this deal over the finish line? What does the structure have to look like? what can we be comfortable with from a risk alignment? and, potential return profile perspective. So I talked about this as well too, within, this week’s serious news, the newsletter, free, you can check it out on our website if you want to sign up, if you haven’t already. but.
to update this when we actually got into the weeds of adopting the legal drafts for transferring these membership interests and amending existing operating agreement and so forth. It’s a lot more complex in practice. And I think most entrepreneurs or small businesses at this point have at least explored
utilizing, one of the AI LLM models for legal assistance, whether it is help with contract prep or, you know, at least reviewing contracts that are sent over to them. That that’s become, much more of a standard practice. and we’re certainly, you know, not, not new to that as well too. That’s probably one of the biggest use cases we’ve utilized AI for.
at this point, but like I called out within the newsletter is that, you know, if you don’t understand the, core questions you’re trying to ask and what type of, structure you’re trying to arrange a certain agreement around, the chances that AI is going to lead you astray or, leave you hanging on really critical language that could come back to bite you, later on.
uh, is, you know, and absolutely critical, uh, um, note that, uh, a lot of people bypass. And so, you know, even though it’s extremely impressive how, how much, um, you know, AI can put together, uh, and the context it can keep in mind now, like there’s just less hallucinations and so forth. Um, you know, the chance that you’re going to get a really robust document without
significant back and forth and an understanding of what’s actually required for underlying dispute resolutions. That’s not going to put existing lawyers out of business anytime soon because you’re going to get so many of these bare bones, looser, CHAT GPT created legal agreements that inevitably are going to have to be tested by the legal system. And we’ll prove that
They’re not quite up to snuff here. That being when you do understand how to work with AI as a partner and have the proper context and know how to go back and forth with it, mean, the outputs are just astonishing. So, I spent admittedly much of the last two days working on these drafts.
Again, it was more complex than I imagined to have to go through this, but all this is worthwhile because it it cleans up a lot of work that we would have to do in the future anyway. These are all going to be reusable. So oftentimes the first lift is going to be the heaviest. But I kind of woke up earlier the night after I’d prepped these stocks and realized, Hey, some of this still seems thin. Like I need to go back to this and just like double check to make sure that
Um, our language isn’t too loose and ultimately it was, um, so, you know, to, to kind of run through the process here in a little bit more detail. Um, you know, we had in existing, uh, equity special purpose vehicle at L, uh, LLC agreement, otherwise known as an operating agreement that we’ve utilized that was attorney prepared, um, in the past. our goal was to.
prepare a mere version of that to amend the existing LLC that’s managing this particular entitlement deal. So that was the biggest lift. It’s like, okay, can we transfer all of these various protections and different member strategies and so forth over to this other document? So that was the heaviest lift here, but rightfully so when I was utilizing a specific
legal brain that we built within CHPT. It was suggesting other documents to be supplementary to that or, you know, okay, the original manager of the LLC has to, you know, waive their or effectively assign new membership interests and so forth. So you had a whole bunch of accessory docs that needed to be added in. And the AI was able to prep all of this, you know,
prep seven or eight different documents all at once and in downloadable format. And astonishingly, compared to six to 12 months ago, so often when I would have much longer dialogues within one particular chat thread with one of the AI models, it would get lost. It wouldn’t be able to remember the context from earlier in the conversation or just
you know, the responses would start to devolve and it could just get worse. And so you’d have to restart different chat threads. I’m sure anybody who’s you just, Hey, I like that. It can, relate, but now just the amount of context, these, um, cutting edge models can, you know, kind of retain, uh, within their memory and, um, you know, not, uh, lose focus. Uh, even, know, when you’re saying, Hey, double check, double check, and you’re expecting things to happen.
it is able to retain that, that context. Like, mean, I must have thousands and thousands of words, within this one specific, legal draft that we were doing. And, the AI was just, Hey, you know, Chet, GPT just never, it never lost focus, which is just so helpful. without having to like readjust all your context windows and so forth and new chat threads, to update on the context. Like it’s just ridiculously impressive for, the amount of firepower.
you can get in just more of the trust you can place that, the AI is not losing its place in, the, the line of thinking you’re, trying to progress on. that being said, you know, the outputs, if you’re still trying to get, know, these full legal agreement, you know, outputs, especially if it’s trying to do seven or eight docs, at once, like it, it, just doesn’t have the capacity to, artfully do that in a detailed enough manner.
Like it can get you the core sections down and broadly cover the protections you need, especially the ones really focused on adjusting throughout the threads or the existing chat thread. so like ultimately the legal provisions and the implications that you’re trying to arrive at remain the same.
compared to if you just had a much more robust, agreement, prepared on, on your behalf. but the language is just not tight enough. at least that that’s been my experience and I think, you know, most people’s experience probably. and so the key fix to have to address this at the moment, which is still fine. it just takes a little bit longer and this is something I’ve learned from, Callan,
you know, probably around a year ago and still applies today is that when you’re trying to do these longer drafts is that you should go section by section that you’re trying to address. Cause the AI can then just really focus in. And instead of trying to prep, you know, a 20 page document all at once, it can focus on a few paragraphs, really nail those. Then you move on to the next one. And so fortunately, because we had an existing, LLC agreement, that was fully attorney prepared, could
go section by section and just say, okay, here’s the copy that we’re very confident in. Here’s what your output was for the AI prepared sections. What do we need to change and adapt here? And then, you know, I was able to go back and forth and that was much more effective. So like the original document might’ve been, I don’t know, around 10 pages or so, this new operating agreement. After going through all this, spending several hours on it, a lot of back and forth.
You know, it got up to like 26 pages, like a really meaty, um, legal agreement, uh, that, um, you know, anybody who has ever been in a legal dispute, you know, it helps both parties, um, because the more you can reduce gray area, um, for any potential dispute, like that’s just going to lower costs for everybody. If, if you are encountering certain scenarios and you’re already addressing them within the writing that.
both parties signed, like it just makes things a lot simpler, a lot more cost effective because everybody agreed to how certain situations, no matter how like black swan type that they might be, like you’ve already signed in writing for them. So even though like the broad implications, again, provisions that we’re trying to achieve within the agreement, like remain largely the same, it’s just the language is now a lot tighter and a lot stricter.
so there’s just far less wiggle room to, to work with them. And that’s always going to be your friend when you’re considering legal agreements. and so then I went back to the other, smaller documents that, were basically ancillary to the main operating agreement had those updated as well. so, you know, understanding this process, like it’s only going to help going forward here. And I was just like glued at trying to figure this out.
you know, frankly, just really, again, impressive and exciting to work on this because, you know, setting up agreements like this. mean, if you went to an attorney and just try to do this from scratch, I mean, this is at least five to 10 grand minimum to to get this done. But again, like contextually, what’s super important here is that we already had an attorney prepped document. Like if I didn’t have that, if I didn’t have a.
guideline on how we should be modeling our new amended operating agreement. Like this would have been way harder to do. It’s not impossible, but it gave us a lot more confidence that our output is going to be legally viable given a test to document that we already have in the wild here. And, know, the AI was just like even pointing out some issues
with the attorney document and how we can shift it to be more protective for us and our counterparty and so forth in this newer document and adjusting it to the relevant state law and all that. That’s just really, really critical information. It’s, again, amazing that it just has all of this. the entire internet that it has at its disposal to be able to check in on and can do it all so quickly.
You know, that was key from our perspective. Uh, and then additionally beyond that, um, uh, is, know, kind of carefully looking over and making sure all the changes that the AI is suggesting is consistent. And luckily it was able to do that. Like it was maintaining, okay, this is the correct section that it’s referencing like 20 plus different sections, um, and not making any mistakes there. Uh, and then, know, my core business partner who’s, you know,
effectively a lawyer outside of holding the degree. Like he’s reviewing that as well and checking, okay, is there any other provisions that we need to include? And so far it’s looking very good. like internally, we have a lot of double checks in process that help us get this right. And even so my partner is like pointing out, yeah, we probably need this other document and signed.
you know, initially first like this power of attorney document, because in case there is any default, you know, if we need to work with third parties, like no one’s going to work with us unless we have this letter. you know, things that can be improved upon. But, you know, overall, very sophisticated in terms of getting the correct output.
And, initially these deals that we’re going into, they’re like a little bit smaller check sizes, you know, sub 200 K even sub a hundred K, uh, initially here. But, know, if we were going to do a half million dollar, one million dollar deal, um, or so forth, like we would certainly double check then with another attorney to like, make sure our docs are super solid. Um, but again, it saves us so much groundwork because we already have tested attorney docs that we built, um, AI docs.
AI generated docs from, model them off of, and then we can go back and like confirm, okay, we’re all squared away here. And that just, you know, makes the attorney’s job a lot easier as well too, and costs us a lot less. So hopefully this gives you an idea on how, you know, to work with AI just in general, again, context and back and forth matter. And if you don’t understand, again,
what you’re trying to achieve. Like you’re going to be led astray so quickly, especially if it’s in an area that you don’t quite understand, which a lot of folks don’t understand like detailed case law and legal provisions outside of lawyers. So I would count myself that as well here. So it’s easy to just like for your eyes to just cloud over and just believe whatever the AI is putting in front of you. And I see this all the time.
You can start to see like ChapGPD’s output, what their legal docs look like. And, you know, I just say like most entrepreneurs seem to be defaulting to that at the moment. So, you know, just a word to the wise on how you should really adjust your process so that it won’t come back to bite you later on. This goes for just using AI in general. So really, really cool features. One of the biggest lifts, biggest ROI piece is possible.
for anybody personal or business owner to work with AI from a legal perspective, but use these same principles for really anything that you’re trying to do with AI. Do the deeper work, use it as a deeper thought partner, but just understand the question that you’re actually trying to answer and do the hard work to actually understand what the output is and what the implications are.
Hopefully this is helpful here. I’m very excited by the continued potential. Again, these LLM models are just remarkably better, remarkably better than even several months ago. And just very excited to continue to dive in and keep growing accordingly. This just feels almost like a superpower. It’s just ridiculous the potential that we have at our fingertips as small business owners and how much we can just
get done. So I will talk with you all next week here. Subscribe and share. Remember SiriusLand.Capital if you want to submit any land deals, 50K minimum purchase price. Looking forward to next time. Take care everybody. Bye.
