In this episode, the focus shifts from SEO gaming to fundamental brand building as AI search optimization proves unreliable. Companies pursuing “generative engine optimization” face unstable algorithms, shifting data sources (Reddit to Wikipedia arbitrarily), and unpredictable LLM responses that make GEO hacks unsustainable compared to traditional brand equity.
Key Takeaways:
- GEO Hacks Are Failing Fast Section AI interviewed a profitable GEO startup founder who shut down because ranking shortcuts proved unsustainable—brands with stronger long-term content and reputation consistently outrank tactical plays.
- The Real Moat Is Boring Work Brands showing up in AI results earned it through high-quality content, consistent citations, and public testimonials over time, not Reddit spam or vector embedding tricks.
- Three Tactical Hedges Worth Trying Use vector embeddings for concept optimization, entity graph gap analysis via ChatGPT, and front-load critical info (information gain rate)—but only as supplements to core content strategy.
Listen to the full episode to hear how Serious Land Capital is balancing AI optimization experiments with their newsletter-driven brand strategy that’s already generating inbound from multimillion-dollar operators.
(Podcast transcript below)
Welcome to Get Serious today. wanted to go over some updates regarding organic marketing techniques. I know this can get kind of in the weeds or over a lot of folks heads, but I think a lot of the core basics really any business should be implementing just because if you really get organic,
type of lead generation going like that is probably your most stable and consistent route for leads and lowest cost. If you can do it properly, it just is a, you know, again, a slower role compared to just shoveling money at paid ads or sponsorships for instance. So something in mind here, I know I’ve talked previously on this pod about
you know, within the AI side, know, everybody’s kind of familiar with SEO, search engine optimization, how you’re ranking in Google, or sometimes Bing or the other search engines. And, you know, with the advent of AI, people talk about, and it doesn’t seem like it’s fully settled on it, but it, know, GEO as in generative engine optimization, or some people call it AEO, AI Engine Optimization. I think GEO is becoming a little bit more.
stable in those two, but you know, it’s like how do you get your brand to show up in chat GPT your perplexity Claude what have you since you know so many or even just like the Google Gemini blurbs I mean when you use Google nowadays, I mean you can just look at the stats But you know even consider your own behavior how often are you just like not even clicking on websites anymore because those AI Overviews are becoming, you know, so solid even if
You know, some of the times they’re still clearly hallucinating or gaslighting. So just something to keep in mind as you look at your own behavior. And so now us as companies, it’s like, OK, with more folks using AI, it’s like, how do we properly show up within those queries that users are making? So a few avenues here.
One is there’s kind of a growing number of companies out there who are, you know, purporting that, you know, they can boost companies rankings within GEO. Just like, you know, how many SEO companies were, you know, built into, you know, nine, nine figure, even billion dollar companies. Because if you really can nail it, you know, that just, yeah, hitting
great organic lead generation for certain companies is just like wildly monetizable. So like, yeah, that next avenue was like, okay, can these companies start figuring out how to optimize for showing up in AI engines? So the key thing here is there’s kind of debates within two camps. Some are like, yeah, it’s hard to even say what…
you know, can show up within GEO because everybody’s, well, the AI models are like constantly being updated and it’s more of a black box plus like everybody’s memory might be different or it’s more contextual per user. So you can, and you know, even if people use the exact same keyword or, you know, not really keywords, but well, it’s kind of keywords, mixed within a certain question that’s being asked to an LLM.
Like if I asked that, you know, exactly same worded question as you might across, you know, one AI model, we almost assuredly will not get the exact same answer. there might be similar elements to it, but because it’s so user dependent and just like dependent on the underlying LLM model and how it just chooses to answer at any given time, it’s impossible to fully predict and like back up some of these companies claims of saying, yeah, you’re showing up higher on.
Chetchiputri or what have you plus the Models are so dependent on Where they’re sourcing information so like a lot of? Alarms will pull from reddit Some from Wikipedia a good amount but like that share of where they’re Pulling info from can just change You know a kind of on on a whim here. So where it’s like if you’re doing SEO, it’s much more stable. Thank you working
to set yourself up in the eyes of Google. Like there’s a consistent pathway with metrics that you can follow to determine how well you’re doing with an SEO. Whereas GEO, it’s just, harder to do from that standpoint alone. And so like, if you really optimize for having your brand to show up everywhere across Reddit.
you know, maybe that algorithm changes and then it’s just like purely more on wiki or other sources instead. That could make, you know, one of these or multiple of these possible channels for your optimization, like not work nearly as well. And so,
you know, what I wanted to bring up is, there was this article from Section AI over the weekend. I think Section is, you know, one of the best, if not the best when it comes to just writing about AI specific to how companies and enterprises work with LLM models. And they were, you know, interviewing a startup founder who had built a cash flowing business centered around GEO. And he decided to shut it down because he’s like, hey, this
you know, we have paying customers, we think we’re doing the right thing. Like it just. These hacks, don’t feel long-term sustainable and, they kind of determined, that, you know, even though there are certain claims they could make that geo optimization was working for the most part.
The brands and companies that are showing up, you know more highly ranked within geo questions and results tend to be the ones that just had better longer-term brand building in in the first place, so What that means is that yeah I mean if you’re putting like highly valuable content that it’s getting routine clicks or conversations and citations You know across
different social media mediums or different news organizations, what have you. Like that ultimately is what is still going to drive the best results. So, you I took that away. It’s like to me, I actually prefer that as a moat here where the hacks, you know, we’re still hedging our bets. Like we’re still, you know, efficiently.
producing content that appears to assist with bumping results within GEO. But at the same time, with our content production, especially with the newsletter and podcasts, et cetera, it’s steady education, consistent, really, really focused on high quality value add that has built up.
our brand to be what it is. know, we have certain testimonials, people for the most part talk well about us publicly. So, you know, it’s almost like what would normally work for SEO is kind of almost a one-to-one to GEO as well with, you know, a few key nuances associated with it. So again, this is just demonstrating another factor where again, like shortcuts are only short.
shortcuts until, you know, everybody learns about them and then it’s no longer a shortcut, right? Or as one of my mentors routinely will say is like all shortcuts are just long cuts in disguise. Cause once you actually dive into it, yeah, like ultimately the hard work is like the only thing that’s reliably gonna get you to that next level and get you the results that you want, which is clearly showing up within GEO. Everybody just wants a hack. Hey, I’m just gonna like spread my brand all around Reddit and I’m just gonna show up like crazy.
But first, that’s not actually proven. And that doesn’t appear to be long-term sustainable because the brands that are doing the work to provide an excellent service and product and get talked about and are providing high value ad content for additional organic reach, that is the moat you need to be building to show up within GEO. So, but like no one wants to do that. It’s hard to do, it’s long to do, it’s boring.
oftentimes. So that’s just something to keep in mind. But again, I would prefer it that way. And any, I think more serious business owner would also prefer that because it, it, you know, separates the wheat from the chaff as they say, since again, a lot of folks don’t want to do the boring work. All that being said, some key avenues that I think are continuing worth optimizing on as again, like kind of a bet.
a hedge, a hedge bet, because again, AI is still just so new. You like, you, you don’t know definitively how everything works necessarily. So, you know, as long as you can do things efficiently, like why not still try to utilize some, you know, hacks for geo positioning. so long as it’s not like a, you know, core focus of your business, but you, if you can do it alongside of the
all the content that you’re already producing, then you know, may as well. So you can just hedge a bit and like, you know, utilize like there, there are some, you know, very intelligent folks that are putting a lot of effort into this. So, you know, just choose your sources wisely here, but you know, we’re getting one of my mentors who was attending a talk that, you know, was around a guy who was really, really focused on how do you show up within, LLM results here.
And some of these, are, you’ll probably want to research this out separately because it gets like pretty mathematically deep. mean, no, no surprise when we’re talking about AI and LLMs here. But there’s something called vector embeddings. And at its core, like when SEO will optimize websites for keywords, like, you know, coffee website, you know, coffee.
throughout most of your pages and again ideally around attractive and high-value content for the user like that’s just gonna position you better for that particular keyword. But vector embedding is more optimizing the content around concepts. So it’s like an extended form of keywords here. It’s like if people are gonna be asking a certain query to an LLM, you wanna position yourself to where
you know, your site and your content is accurately answering that particular concept. So like similar, but not identical, something to keep in mind. Again, that’s like just brushing the surface of the complexity there. And then another piece is entity graphs. So this one, like, again, there’s a lot of mathematical reasoning. So you would actually just like want to use an LLM.
by like, yeah, can you just do an entity graph gap analysis within ChatGPT, like post some of your content, blog post, whatever that you’ve done, and then just have it do an analysis, and then you can plug that into Google’s natural language API, and then you can kind of figure out, where is there a chance to improve my ability to show up?
or to match the entity graph for certain queries that people are asking it into LLM models. borrowing some of the phrasing that my mentor had posted is like, if somebody was asking, hey, what’s the best toothpaste for children with sensitive teeth due to a palatal expander and which doesn’t contain fluoride, like a very specific question.
like you would want your content oriented around properly answering something like that. But again, unless you really know how to do that, it’s, probably not gonna know offhand, like I don’t know how to do that offhand, like that’s where you wanna utilize the tools to help you assess that properly and adjust accordingly. And then I think this one’s probably the easiest to implement.
And I’ve seen it done on certain blogs here. I’m almost certain like Sranthravaatsa’s Next Billion blog does this, but it’s called the information gain rate and time to answer. Because like we all know this like, you know, whether concretely or just in practice is that like the LLMs will be lazy. Like if you give them a huge PDF, they’re going to start scanning, you know, quickly after they start assessing it. And they might gaslight it. Again, the models are getting better to where it shouldn’t happen as often, but.
a lot of LLMs just when being tested is that if they’re given a piece of content, they’ll kind of start skimming or they’ll stop reading initially after the first hundred lines. So the information that will sit at the top of your project knowledge or whatever your reference content is has a better chance of being referenced to the LLM anyway. So that can be included of anything in your project knowledge or your
content on your website, you again, for geo purposes. like kind of the bottom line here is that you should probably have your most important info earlier on in the content. So and critically, it should be new and valuable information to to the reader. And so you don’t really want semantic redundancy, but they should be like
clear cut answering something quickly without filler. It’s like pretty much everybody wants to consume content nowadays. Like give it to me just quick and fast. I don’t need the fluff here. So that is something to keep in mind with your content. like, yeah, it might even be worth doing a TLDR summary right at the top. So then you can make sure that those key concepts are like.
bullet point summaries and you’ll see that on certain new sites and so forth now too. again, some of these sites are definitely using it if you pay attention. Again, whether that’s actually panning out to be better in GEO, questionable, but it doesn’t hurt given, again, these at least three examples of what you can do are like pretty basic and you’re already gonna be producing content like this anyway, so why not just adjust it slightly on the off chance.
or even the possible decent on chance that you can hedge your bets and get better results there. So just some key food for thought and also we’re investing in this other company. It’s pretty expensive to work with, we call Concierge AI. So we’re onboarding with them now, but we can just tell, again, looking at your own…
usage across websites and so forth. Like we’re just going to have some, and we already see the browsers like the Comets and AIs, Atlas, and I use Dia browser. Like it’s just combining LLM models with your browsing experience now. And I think that’s going to be so much more consistently implemented across, you know,
every website effectively. And you get like a ton of different chat bots now. It might have AI like, you know, look at the big corporations. But I really like how Concierge set up because they just ingest all of the information across your site and they build effectively a FAQ that, you know, can very appropriately and, you know, both concisely as well as comprehensively answer user questions on your website. And so
Even though we have a solid FAQ, we have solid content lighting and people are just fast. You know, they want to get their answers are lazy. So we answer so many questions about how we operate. Like what are your profit splits? What type of deals you’re looking at through email? So we’re really excited to try implementing this on our serious land capital website as well as land funding partners too. It’s like a different setup. Kind of more of our arm’s length, you know, sister website.
kind of managing all the different funding options across the industry and to just, you know, provide further insight to users on those websites to get them the answers they need and, you know, prevent additional other questions unnecessarily being asked to us more directly. So you’ll see that live here shortly. Very excited to try that out. But again, I think
Some of these, like there’s a lot of like fluff AI companies out there, but I was like, okay, this team really understands the core problem here. And I was very impressed by their initial product demo. So again, expensive to try out, but you know, if they hit it right, I think it’s a very, very interesting approach to lead generation and conversions. well, not really lead generation, more, well, actually it’s probably both.
lead gen and conversions, probably more on the conversion side, but you can still get, if somebody doesn’t raise their hand as a potential lead here, when they visit your website, then they’re not necessarily a lead anyway. if you’re not actually capturing them in some type of way. So that’s just something to keep in mind here. SeriousLand.Capital for any of your funding needs, 50K plus purchase price. Land Daily Diligence is gonna be hopping on that.
number of deals to review later. If you haven’t seen, check out my Serious News subdivide outline on how to price subdivides since pretty much the entire industry is bad at it. Hopefully you got a kick out of that one. And with that, subscribe and share. Take care until next time. Bye.
