In this episode, Claude Cowork takes center stage as a tool that compressed what used to be multi-day legal and operational lifts into hours. A 40-page crypto theft audit report, 10 legal documents totaling 150-plus pages for the LandPricer LLC restructuring, an investor deck for a $1.1M Northeast deal, and a full newsletter prompt optimization were all produced within a single week. SEO article production time dropped from 58 minutes to 17 minutes on the base model alone, prompting a decision to 5X weekly content output rather than pocket the savings.
Key Takeaways:
- Cowork Requires Attorney-Grade Inputs Garbage-in still applies: the LandPricer operating agreement only worked because existing attorney-drafted LLCs served as boilerplate context.
- Time Savings Should Fund Output Increases, Not Idle Time When SEO production dropped 75% without any added automation, the correct move was scaling from 3 articles per week to 15, not celebrating the efficiency gain.
- Parallel Processing Eliminates Backlogs Cowork handles multiple tasks simultaneously, which converted a chronic patient-note backlog for a clinical therapist into a single final-review workflow.
- Investor Deck Prep Went from Days to Hours Feeding prior raise decks as templates alongside new deal parameters lets Cowork produce a capital raise summary fast enough to pursue soft commits on a $1.1M deal without a team scramble.
- The $100/Month Max Plan Is Non-Negotiable The $20 plan burns out quickly under real production loads; the upgrade is the price of admission for meaningful ROI.
Tune in for the full breakdown of exactly how each of these workflows was structured and why this single week may represent a before-and-after inflection point for the business.
(Podcast transcript below)
Welcome to Get Serious, we’re at Serious Land Capital. We have funded over $6 million worth of land deals successfully with industry leading 41 % operating margins. So today I really wanted to go into some more tangible ROI driving AI wins here. I know I’ve been remarking more on these on the last few episodes, but honestly, within like the last month or so,
the capabilities of some of these LLMs, especially, Claude co-work is just leaps and bounds better than anything we were, we were experiencing, like honestly, even three or four months ago. you know, some of the tactics and strategies I was using even two weeks ago are just outdated. Like I think one of the most recent podcasts I did was talking about, building out a, a highly,
a structure, you know, 30 plus page operating agreement for an entitlement deal that we were pursuing. And, you know, that took me, you know, a couple, almost a couple full days to really nail in. You know, obviously, if we do more of those, it would get a bit easier here. Co-work now just absolutely blew that output out of the water here. So
If you aren’t familiar yet with Claude Cowork, it’s almost combining the core functionality of Claude, which is just a standard LLM. You can kind of speak to it or chat with it, and it’ll produce the outputs that you want with the more powerful Claude code, which…
is a bit more of a learning curve. Like if you have some software development experience, it would be a bit easier to work with the terminal and so forth. But that’s a way to vibe code and be able to produce real programs and applications. Claude Cowork is almost merging the two to where, you know, anybody without any development experience can kind of come in and start building some really powerful things or at least
you know, not even necessarily repeatable programs, but, um, you know, the, the, the opportunities are almost infinite here, but like for a super bare bones one, if you just wanted to start testing this out, um, is, know, if you just have a million screenshots on your desktop or, you know, file, like who, who, who’s file system, even if you don’t, uh, even if you do organize it, like doesn’t start getting messy again.
within two to four weeks or something like that. Like, Co-Work is set up perfectly just to organize your digital life as a really low-hanging fruit to start with. And critically, it lives locally on your desktop. It now works for both Mac and Windows. More recently, it was just for Mac. And…
it definitely requires higher output. we upgraded our Claude subscription to the max plan. So it’s a hundred bucks a month. You you’re going to burn through it pretty quickly if you’re using the 20 bucks a month plan, but you know, even for a hundred bucks, like it is a bargain to be able to utilize this tool just given like the almost unlimited capabilities. So I just wanted to set the stage there.
Um, and then go over, uh, some key examples of what we write, like literally just over this week alone. Um, like sub seven days here, this is, this is just the output. So I just want to lay it out for you. I’ll provide a few bits of context and, just how dramatically this is, you know, changing the trajectory of where I think our business can go. Um, just cause it honestly, it honestly feels like having a magic wand in hand. Um,
We’ve been using AI for a while, like this, these capabilities here are just, you know, again, leaps and bounds beyond anything else. And I’ll probably write about this again soon, but it’s like, you know, you, for anybody who used to be a, you know, a gamer, maybe you were like, used to play a of video games have written about that before, but you know, it would always be like, I don’t know, three or four years between different, you know, generations for consoles, you know, think you’re
Nintendo 64 or GameCube now like the PlayStation 4 or 5, whatever. Like it would take significant time between iterations here with AI. it is so much like you’re just skipping generations within such short period of time. Like if anybody can remember like the original NES, the Nintendo Entertainment System came out like the late eighties. Like that’s what AI felt like even, I don’t know, late summer of last year, like even
GPT-5 and so forth. Still so many hallucinations you still couldn’t do nearly as kind of robust, just parallel processing and long chat chains to where now it feels like a PS5. You’re just collapsing the time down with the…
increase of capability of these models. That’s absolutely mind blowing. So again, if you haven’t utilized this, and we were even a little late to the game by a couple of weeks or so to utilizing cowork and starting to apply it here, but it will absolutely change what you think is possible for your own business, your own life going forward here and just making it so excited. Like this is all I really want to do within the business here.
Or offload even more operational tasks just so can focus on my zone of genius Which is like largely capital relationship work and underwriting So again all of that preamble mind here I’m gonna run through some of the examples So one of these which I wrote about recently really unfortunate one of my earlier crypto investments in helium Was stolen entirely by a really sophisticated thief almost a year ago right underneath my nose didn’t notice
super unfortunate and like that this was the final nail in the coffin for, you know, getting involved in, in unregulated, exotic investments. I detailed that on a newsletter if you want to read more about it. and even though it was like a sub five ish percent, you know, estimate on my side that I would get any recovery at all here. the exchange where the thief offloaded the tokens to
was at least responsive to me when I reached out, however they required law enforcement to send in an official request. So I had to file a police report. But given the complexity here, like all the different address changes, all the crypto and so forth, what I went ahead and did was I just, and it took, I don’t know, less than 30 minutes, I screenshotted the flow of funds.
you know, again, fortunately, one of the benefits of crypto is you can follow where tokens go, you know, across the blockchain and follow the different addresses. and so I, just followed the chain of custody there, put it all within a folder. Like I didn’t even bother labeling this. Like it was all just timestamped by my screenshots. I gave Claude cowork access to that folder and, you know, I basically told it, Hey, you know, your top,
crypto investigator when it comes to stopping sophisticated thievery, et cetera. Like again, you always want to still set the role. And I’m just going off the cuff here. I’m not looking at what my actual prompt was. And I just told it, okay, I gave you the screenshots here. We got all these tokens stolen. More context than that, but still pretty basic here. And I just told it, hey, if I need law enforcement to…
reach out to this exchange and I need time to be able to follow a full report of this, you know, theft of my crypto here. Can you build me out a report for this? And like it built out in almost 40 page audit report, just, you know, super easy to follow all color coded and calling out, you know, exec summary and stating, okay, here’s the, you know,
precisely when the theft happened, it ended up here, here are the key addresses, here’s all the associated screenshots and evidence that you could easily follow. Just stunning, 95 % of it was done properly on the first go. There were some formatting issues and then I wanted to adjust like it made an error as far as what the total amount.
was sent over to the exchange. like, yeah, you always want to double check AI’s work here. But, you know, overall, like it was just incredible output. Like for me to have done that by myself would have been at least a full day. Like we’re talking eight to 10 plus hours to have produced something like that and probably wouldn’t have looked as nice either, like even utilizing other, you know, AI assisted tools and so forth. So it was just absolutely incredible. And that was done within
sub 30 minutes. And again, like I’m so busy with my own serious land capital and all this other work is like, gosh, you know, even though I lost money here, like I can’t be too distracted by this. I just need to get this off my mind. Last it, Jeffer, reach out to law enforcement, see if they can do it. If they can’t, okay, at least we explored it. So I gave it, gave that test to cowork, got it all set up, you know, sent in the police report, sent into the FBI, and Austin police department, they got back to me right away and they said, yeah, this is, know,
We see this more frequently than you’d imagine, but a lot of people are way less organized. yeah, thanks for preparing all this. It was easy to follow. it’s like, because again, you think about your counterparty on the other end, having to sort through all of this. And if you give people a clear roadmap on what to follow, they’re just way more likely to help you out. So they complimented that work there too. And it was all done through cowork. So still very unlikely I get anything back, but at least
I explored every possible avenue to attempt to achieve it. And as a side effect, I got a lot more practice working with cowork. Next piece is just writing the newsletter. You know, the serious news newsletter free every Monday. If you haven’t signed up for that is, you know, we’ve had almost a year of these go out roughly. And I was trying to determine, OK, open rates,
based on the subject lines and the preview texts that are going into emails, like what’s the best way to piece that out? I was able to get, and some of these like, I don’t even know exactly what I should extract out and like which are the right type of statistics to look at. So like, this is what you just ask AI, like, you know, position yourself, what’s the best way to accomplish this problem and then go get the data and give it access so that it can, you know, help you achieve the goal that you’re actually trying to get to.
And so, you know, we extracted the relevant, you know, email statistics from MailChimp, got it back and was trying to determine, okay, over, you know, 40 plus of these newsletters so far, almost, you know, we are just shy of a year here. You know, like, how do we determine, you know, which subject lines work better and so forth? And, know, we just went back and forth with cowork a little bit, like not that much time.
And it was able to determine, here’s the key winners here being mindful. Yeah. There’s like not hundreds, thousands of data points, but it’s enough to start building a trend. and let’s update the underlying newsletter prompt to account for the findings that we’ve gotten. so it improves the, subject line suggestions as well as the preview text.
And as an extra side effect, it also determined, hey, there’s just, you know, repetitive or superfluous information within your newsletter prompt. Do you want me to optimize that and clean it up as well? Because, know, the shorter and more compact those are, the less token wastage you’re going to get when you’re utilizing the AIs. Like, yeah, do all of that for me. Boom. It pumps it out. I update the, you know, copy paste the actual prompt output.
into the existing project and done with that. Just absolutely remarkable. just Opus 4.6 on average, it’s all my thoughts. You can tell when they’re writing the newsletter, it’s all me writing it or I’ll just dictate into Claude and then I’ll just have that prompt kind of clean things up in the formatting because like…
That’s the annoying part of writing is, okay, where am I gonna put a bold or where are my paragraph breaks going? And so forth here, that’s the inertia part. But when I can just feed my thoughts and that makes me want to write even more because then I don’t need to worry about formatting and I gotta get this over into the actual delivery mechanism. No one really wants to do that. That’s not the exciting part. But now it’s like, okay, Opus 4.6 is even better at just.
formatting and maintaining my voice when it’s directly coming from me and it just needs less direct editing and now it makes me want to write even more. So that’s another piece. Now moving on to the legal documentation piece. So, you know, I know I mentioned in the previous podcasts, you know, a couple full days drafting all this LLC for
the sake of this entitlement deal. And I thought, yeah, this is like really sophisticated, know, still really impressive what we can do with AI and just maintaining all this context. Now, as we are finalizing the handoff of the, land Pricer LLC to a new operating team, which is, you a new CEO and my, you know, current and former CTO for the business and, know, strategically targeting
enterprise sales to appraiser firms instead. And again, just I applaud this team. They have a better vision, more experience with enterprise software than I ever did to take over this landpricer project. you know, recent demos and meetings have been super, super encouraging with appraiser firms. So it’s like, okay, let’s go ahead and finalize the docs here. Now it’s worth the effort. We have the ball rolling here that this is beyond just like, okay, kind of testing this out.
And I thought I was going to have to go again back to like chat GPT and just line up all of these stocks. And like, it was a significant lift, lift because there was an existing LLC that was going to have to assign the IP to a new LLC. was different investing rights. you know, how does my existing equity transfer into the new company and how, like, what are the correct incentives and, you know, structure for the new CEO and CTO being taken on.
Plus we’re anticipating getting design partners who are gonna be both customers as well as potential equity holders within the LLC. Like it gets complicated pretty quickly to get all of that drafted properly here. Plus we have to shut or basically terminate the previous LLC language, which rewarded the CTO in a different format versus the new LLCs. Like all of that was gonna have to be addressed.
And I thought that was going to be, okay, this is going to be a really heavy lift. Even utilizing that, you know, chat GPT framework, but now I just positioned it all to cowork. Critically, I had a lot of very solid legal docs that could serve as templates, not precisely for a software company, but I had an existing LLC for Lampricer. We had an agreed upon new company redlined rubric of what we were.
you know, roughly agreeing to for the new operating agreement. Plus I had some other existing LLCs that were, you know, like full attorney written and approved and so forth. Um, that could serve as, you know, boilerplate language for several of the sections, even if it didn’t directly apply, um, you know, one to one to a software company. But I fed all of that context and gave cowork access to that folder for it to build out, Hey, I need a new operating agreement. Like don’t.
pull any punches here we need to ensure, like yeah, this is probably gonna be a 25, 30 plus page document, make sure it’s as thorough at least as the existing docs within this folder, accounting for all of our red line changes here, any other key documentations that we need, we had very loose language as far as what the design partner agreement was gonna look like, which again, like I’ve talked about before,
utilizing AI and you don’t give it proper context and really dive in, like your legal stuff is just going to be junk. So I, again, made sure that I gave it deep, deep, deep context, plenty to work off of. And, you know, this particular task like pretty much broke cowork, like it stalled out a handful of times. It was a huge ask. And it told me it was like, yeah, this is major to do like an entirely new company legal structure.
but I had to restart it a couple times as it bugged out on me, but it was still able to produce 10 separate legal documents, including stuff I didn’t even ask for. One was like, hey, here’s just a summary as if you were receiving all these docs from a legal firm, pointing out these are the absolute most critical points. These are most time sensitive. Make sure to look at these if you have any negotiations.
This is easier to give up than this and so forth and all the rationale like super, super helpful. And, you know, the actual docs, I mean, it produced like over 150 pages across all 10 of these docs. Some of these were like 25 to 30 plus pages a piece all did it in one shot. You know, there were a couple that ran into errors when I was like trying to open them up and into words. So, you know, a couple of formatting things that had to be fixed, but
and it asks them clarifying questions. Hey, we got to go fix this and, um, you know, adjust the framework here, tax considerations, et cetera. But a massive lift, like again, if you were trying to produce this, just utilizing a regular legal firm or, you know, what have you, like this is easily into the five figures easily. Um, and I was able to get this all done with cowork and again, moved back into my normal line of work at Sirius land capital. Um, instead of like,
just getting boiled down into the legalese. And now like I can have a real discussion with these new operating team for landpricer taking care of business, like stunning output, crazy.
What time are we at here? I want to speed this up just a little bit. Yeah. 20 minutes in, but like again, these are just absolutely killer. So, three other key points, one, and I believe I had chatted about this again recently too. I wrote about it. We were doing a time audit within our business, every single repeatable task. How do we optimize, anything that is not done once with AI workflows. And yeah, we built out a prompt that determined, okay, you know, based on the
loom review of this particular task, we could chop down this amount of time. And these are the highest impact implementations for AI to target parts of the workflow. So one of our biggest pieces was the SEO article preparation. And on a weekly basis, we were producing one SEO article and two GEO articles.
per week, GEO is to show up higher within AI, like Chetchupt or Perplexity or Gemini, whatever. And really time consuming, especially the SEO. was taking almost an hour per week just to produce one SEO article. And so I was consulting with Carter and Callan at the Oncoming Business and they were giving a number of suggestions on how to orient around this.
But even though we had done this time audit roughly a month ago or just over, Claude as a baseline improved so much that even without us implementing some of these workflow automations or Zapier and so forth, the time went from 58 minutes per SEO article to 17 minutes in like a month. And that’s just because Claude got better inherently.
at its production, less error management and so forth and adjustments. And so then it would have been diminishing returns to like, okay, we’re gonna chop off like another two or three minutes here, like seven, like it’s already almost 75 % improvement nearly just off the base model. So instead what I oriented my team around is like, okay, let’s take those time savings. Plus we know the automation and scheduling capabilities that
Claude Cowork is capable of, and let’s make sure we just have optimized keywords to go after for SEO and geo. And now, since we understand the process and have the process more boiled down and automated in the first place, let’s do more. It’s the Hormozi thing. Why would we limit ourselves to just a few articles per week? Let’s take all those time savings and efficiency and now 5X our output.
let’s do five SEO articles per week, 10 geo articles per week. that should take the same amount of time or less than what it was previously taking to do, you know, 20 % of that capacity. So that is being done as we speak and will be in place by the end of this week here, like real game-changing ROI driving activity, like SEO and geo organic search. That’s been by far the biggest boon for our marketing.
of serious land capital. it’s like that’s where we double down to get more out of. So another massive example on where Cowork and Claude by itself are coming into play here. Next piece here is the whole another discussion too, massive deal in the Northeast US from our number one operator in the land space. Like this guy is just absolute killer. He sent us over like $2 million worth of deals over the past week, all three of them.
past initial DD, like just in a tremendous hit rate when we say no to most things that come our way. And this one’s like just over a million dollar buy, know, closer to like 1.1. Just killer, killer area. Like again, even in rough markets like this, you know, this is something that I’m basically willing to bet the farm on, but we need a lot of external capital and need to come up with it quickly, at least for soft commits. I need to raise like roughly 800K.
in capital for this particular deal and backstop it personally if needed. Previously for investor raises, we would build these deck summaries for the deal and that might take three to four days of iteration back and forth, making sure we’re all aligned. But because we have examples from previous raises, we can feed it into cowork, give the new info related to the deal to cowork and it can just spit out
you effectively a perfect deck summary. So that now all of sudden we can get a brand new huge deal that we got to raise quickly for and get out a deck to investors within a couple hours. again, massive time savings, you know, reduces the burden on our team as well here too. We can give our investors even better.
you know, anticipated returns because we can share more of the economics with them without us having to bear the burden as much. So like it’s win-win across the board there. Um, and then finally, even outside of the land piece, you know, my wife, a clinical therapist, um, you know, she has excellent cloud project that would help her build outpatient notes, um, and format them and so forth here, but it still had to be one at a time. So if she was seeing multiple patients per day, like just like every other, like healthcare provider out there, just
gets behind on their notes, like it’s just a never ending game. And she would routinely have a backlog and feel stressed about that and have to catch up. Now with co-work, she can now parallel process and have all of those notes be fed into co-work, all of them being done at once. And so then she just does a final review versus having to go one by one by one by one. And again, massive game changer on her end of things too. So.
To finalize here again, like this was all done within the last week. We are just getting started here. This might be the most excited I’ve ever been as a professional, regardless of market, regardless of this turmoil, what have you. Like it is unreal what is at our fingertips here. So more to share as we continue to learn here. And you know, if you want to learn more on how to do all this stuff, mean, the Callan’s Accelerate to Automate course is going to be starting up.
again next month at end of April here. So you’ll hear more from me about that. Again, they are my go-to AI coaches driving real results, real dollars into our business here. And I just gave you almost 30 minutes of examples on exactly what we’ve been doing that speak for themselves. subscribe and share everybody. Hopefully this one was helpful today. Looking forward to next time. Take care now.
Bye.
