Find All the Citations ChatGPT is Using to Answer Your Target Customer's Questions

It's another GEO or AIO or whatever you want to call it type of episode, but this one is way different. We're actually talking about the tactical way that you actually can show up within the search results. And to help teach us, I'm having on my friend Sean, who is one of the founders of Eldil AI. We're going to talk first about how all observability tools are basically BS in this category. We're then going to talk about what is the most impactful thing that you can do for your brand, which is building citations. And this is what Sean's company does. They basically aggregate all these citations and then handle the outreach negotiation for you. It's actually crazy what they're building. It's like this whole AI agent suite. We'll go into that with more detail. You're also going to learn this like power law of specific citations. So the citations that AI is referencing, it typically references them more often. So it has these like preferences that it goes to over and over again. And so when you think about building these citations for your company, you basically want to focus on those and then work your way down that list. And finally, Sean's going to share just his raw data, like what they're seeing from the companies that they're seeing have success with this and why you should be focusing on this and why you should be even thinking about this spoiler alert, it's because ChatGPT searches are increasing daily by a ridiculous amount. They see this in their data. And with that said, today's episode is brought to you by graft.com. Graft is an AI agent for marketing analytics. You just plug in all of your data sources and you can have it build charts or dashboards or reports. There's zero learning curve. It's just chat functionality and you don't even need to understand the underlying data. You can ask very broad lizard brain style questions like how many people went to my website last week on phones? And it will understand that context to build out whatever it is that you need to basically do the analytics on your business. How you can use Graph for AI SEO is you can actually plug into your Search Console and you can filter keywords that are over 50 characters long that have low clicks but high impressions. Sean and I actually talk about this together. And you can also plug into your Google Analytics with Graph and track the referral traffic from ChatGPT over time, analyze that session data, and see the conversion events from these users that are coming from those sources to your site. I want to give you a free trial of Graph that's worth $500. Just click the link below and I'll also do a personal discovery call with you to set up your first dashboard and get you onto the platform. And with that said, let's get started with today's show. Sean, I'm so stoked to have you, man. I, I say this like every episode now, but I just look forward to these calls anymore because it's like I'm having these conversations offline and I just want to like focus on it with like somebody for an hour and extract, you know, all the knowledge from, from them. What you're doing is super fascinating. Just to give people context, Sean owns a company that is like an AI SEO optimization. We're going to get deep into that and all this stuff that they're seeing working, working. But specifically what they really focus on is the citations element. And this is like why we're here today, because this is actually how you make impact for this. They've got like hundreds of millions of citations in their database that they're mapping to specific queries. But anyways, man, super stoked to have you here. Thanks for, thanks for jumping on. Dude, thanks for having me. Any, I'm going to give you all of my secrets. So just ask anything. Like I'll tell you. No, we love it, man. So I think to start, like, I just want to begin the conversation. There's a million observability tools. We got connected because like, we both, I think we share the same belief that observability is, it means nothing unless you can actually get placements. All the data supports that the only way that you can actually like get more traffic, show up higher within these AI search engines is by basically getting listed on the articles that are on page 1 through 3 for the keywords that are related. So the whole process, I did a, a previous podcast episode talking about AI Fanning. So the audience should have some knowledge on like what that looks like and how that functions. If you haven't listened to that one, go back to the back catalog and you can see it. But just to like, you know, bring the hammer down on this to begin with, like why are all the observability tools just bullshit basically? Yeah, it's fake. It's fake. It, it comes down to like people want to feel good. Um, and when they see like numbers and charts, like they generally feel good. And if you go and put out a chart, if you start tracking essentially your referrals from ChatGPT right now, they are naturally going to go up every month just a little bit, right? There's a trend there because more and more people are using LLMs. And so like there is this like natural feel-good element that like marketers love. And so people are going to sell, sell that. So, um, what we sell a lot of times does not make people feel good. We actually oftentimes tell them like, look, you, you kind of suck. Um, and here's what you need to do, uh, to stop sucking. So there is a layer of observability that's really important, but you know, these tools are thousands and thousands of dollars a month. Um, and they just give you basic, like, you know, think like Google Analytics for LLMs. Totally. No, I, I think, I think the thing that I like want to focus on today too is that all of them are pulling from the same data, like just to, and maybe dive into like how this like functions. Like from my understanding, basically everybody's buying third-party clickstream data and sending synthetic data to these LLMs to see like what the queries like look like., you know, coming back. Is there any other secret sauce that people should know about, like about how this data aggregation is occurring? And like, is there any difference between all of these tools? 'Cause the price points are crazy. You know, I've seen some from like $99, you know, all the way up to, you know, whatever, couple grand a month basically to, to, to, to watch this. So talk to me about that. Like where, where is the actual, like, you know, is there a difference in the data? So I will say there is a difference. So we all— so there's a bunch of geo companies, right? A bunch of observability tools, and we all get data from the same places, right? There's essentially data brokers that we can buy from. It's pretty inexpensive. I would say there actually is a substantial quality difference team by team, essentially. So when you get a response back from an LLM and you're essentially looking at that, you're analyzing it, there is a big difference in quality between these tools on how well they analyze those responses. We had to like really dive into the science of this for a pharmaceutical company of ours where they like the, the minutia was extremely important to their data science team. And so we really had to get into that and we just realized like not all these tools are made the same. They're not all equal, especially in how they extract mentions. So brand mentions and company mentions. And so there can be like huge, huge differences in stats between tools, even within tools between between different tenants within that tool. Let's say you, you bring your company onto the same tool twice, you will probably see two different sets of data that, that are like pretty different. That's super fascinating. So it, is it really, it does actually matter like where they're sourcing this from? So a lot of people leaning on like synthetic data. I've seen this a lot where it's like you connect your brand and then basically you're just like throwing stuff at these, uh, um, these platforms to see what returns back. Do you feel like that is more representative or is this third-party data more representative? Like what do you guys trust more? Like Personally, like I hear this and I think, you know, the third-party data is probably more of the shape of like what's the actual truth. But I'm just curious your thoughts. Yeah, so we, we, so when you're talking about prompts, like what are we actually throwing into the LLM? That's, that's where this becomes interesting. So the third-party vendors, I've talked to them every, every time I see one, I get, get in touch with them. And so far they have either been shady or just like I did, I don't trust what they're doing. And it comes down into where they get their data, which the best explanation I have gotten is they get it from essentially, what are those things called? The, oh man, I'm looking at them. The Chrome extensions. Yeah, totally. Yeah, they get these, they get this anonymized data from Chrome extensions and then they essentially have a data science team just kind of like expand on that. And when I actually talked with one of these vendors, they offered me, I kid you not, one CSV pull each month for $100K a year. And they did not have, they could not explain their methodologies to a sat in a satisfactory way. I'm a longtime software engineer and what they said was just kind of like BS. So here's, here's how we think about it. There's like, there's 3 different ways to do this. The first way, and a lot of people do this, believe it or not, um, they just guess, right? Like, so they basically have a Perplexity ChatGPT prompt. They say, you know, essentially what are the bottom, bottom of funnel keywords for this company, they expand those a little bit. So that's, that's just a total guess. That's what they use for prompts. The next best, and this is what we do, is we integrate with your Google Analytics and your Google Search Console. And we're seeing essentially what is converting for your company historically. So in Google, so right, ChatGPT hasn't told us what people are searching and in what volume like Google did. And so right now we're leaning on Google as a proxy for what people are actually searching. So one good thing about LLMs is that they're generative, right? So there's a question behind every question. If you look at a sentence, right, that could be a 7-word sentence, but that 7-word sentence means the same thing as a 4-word sentence, right? There's a lot of different ways you can mark up a sentence. And thankfully, the LLMs understand the question behind the question. So while it's not a completely satisfactory way to find prompts by using Google as a proxy, it's the best we have right now. I am— To break this down for people too, just again, I just want to drill down on this real fast because I think it's really important. You can actually see these queries, right? Like, and again, I have another video that's talking about this, but you can like do a regex within your Search Console to find queries that are super long. The, what the query, the shape of them will look like is basically imagine like it's a 50-character-plus long query. It's going to have a ton of impressions and zero clicks. And that's a strong signal that that is a query that the AI is doing. So like to, you know, to walk through this, imagine like some lizard brain person puts in a 6-word prompt. The LLM is like, what does this mean? They go and they expand that into like, okay, we think it could mean this thing. And then from that they go and they fan out and write all of these different, you know, prompts that they're basically going and like scouring the internet for aggregating that. And then that's what the response is based off of. You can see this, this information that's relative to your business. And so what Sean is talking about is basically they're using Search Console to identify that. And then that can influence the content, um, or the kind of understanding the, the, the data. That exists in relationship to the specific category that you're in. So anyways, just wanted to add that there and before you move forward. So I could not have said it better myself. Thank you, Cody. There's one painful piece to that, which is you can't exactly tell which LLM it came from. Was it Perplexity, Gemini, who actually made that query? So it is a really, really helpful piece of data right now. And I'm like, I'm really excited. Like the synthetic data, I think there's hope there. I hope someone goes and does that well, does it right, does it really transparently. I think that's the important part. There's like anyone listening to this, like if you're a crack data scientist, like you should go solve that problem, ton of money in it. But like right now with synthetic data, it's not worth it. But we're keeping our eye on the market. No, fascinating, man. Cool. So and just to go deeper too on the Chrome extensions thing, because I have a little bit of knowledge about this, there's a huge marketplace for this. So all the free tools that exist, if they're free, they're basically aggregating your data and selling this. So they're just like spying on you and selling this in the background. Um, we actually had this Chrome extension that my co-founder and I, Max, built. Um, this was years ago. We like threw together in like 3 hours, you know, and just had it on the web store just to see what would happen. Um, it, uh, got to like 10,000, uh, daily active users. It was just free, basically wrote, uh, Excel functions, right? Like was the, the existence for it. Um, got approached by a data broker. They wanted to buy it for like 50 grand basically, cuz they just want access to those 10,000 people and the activities that they're doing and And that's how they go and they aggregate this. It's like very manual, but they basically find these projects that are already doing well and then they go and, you know, buy those assets from them. So, awesome, man. Okay, let's jump into it. How do you actually get like to show up within AI Search or GEO? Everybody talks about like fucking making all of these blog posts on their own sites, spinning up like tons of spam sites where it's like surface area. Tell me what's actually working based off of the real data that you guys are seeing. This is what you guys spend all of your time on, and I think what makes you different is really focusing on like, what are the levers that we can pull both on and offsite? Just gimme the rundown of that. Nice. It's not posting on Reddit. That's the first one. That's a, that's the question I get every day. Like, so does this involve our Reddit strategy? No, it doesn't. We, we have like pretty concrete evidence about that. Um, it doesn't does depend on your industry, but we mostly work in B2B SaaS, healthcare, e-commerce, and like Reddit. Reddit is not it. So what works? I'm going to go through like kind of 3 things we think of in like pillars that are essentially, they increase in importance. So authority is the most important piece of this and we're going to talk about that. That's citations. But to make this really human, right? I'm talking to you right now, Cody, and I'm technically communicating with you. So the first and most important thing is that your website can technically communicate with an LLM. Can an LLM use it? Can an LLM navigate around it? Can it find content? Right? That's table stakes. Funny enough, when we look at clients' sites, oftentimes they have, they've been paying an agency for 10, 15 years and they have actually not solved this problem. An LLM actually can't use their website well. And it shows in their geo results. So the first thing is like technically you have to have a website that works. The second and obvious is like you have to have content, you have to have a web presence. Like there has to be actually something for the LLM to like interact with. And I hear a lot of people, they'll say, oh, you know, how should we format our, you know, how should we format our landing pages? And you know, what shape does the LLM like? What kind of content on page? And honestly, like we haven't seen any conclusive evidence that like like any of that really works. Like, again, it's like a feel-good thing, right? Like people feel good when they're like changing around the elements on a page because they're AI optimizing it. But there's really like not that much evidence that that does anything. So, but content is important and you have to have it. But the third and most important, so we hit technical, can I tell you something? Content, which is, you know, the value of the words coming out of my mouth, and then authority, which is, do you believe what I'm saying? Does the LLM believe what you are saying? And so authority is the biggest lever we have access to. It is, I would also say, the most tried and true, at least among my customers. And so essentially authority means having other companies talk about you like you are the expert in your industry. So when you look at LLM responses, there's often these things called citations. And a citation is basically a link out to another piece of content. So this could be your content, this could be your competitor's content, this could be third-party content. And so that is a piece of the observability stack, which is like what citations are being essentially pinged the most frequently. And that, that gives us some good hints as far as like what the LLM likes in response to a certain set of questions. And so we're tracking those and we can see essentially what the winners are. So within a single client, you might have 12,000 to 15,000 unique citations over a 6-month period.. And when we look at those citations, there's only about 200 that matter, 200 to 300, right? So there's a power law here. Not all citations were created equally. And so when we look at those top 200 to 300, we're trying to break it down in terms of essentially what's your content, once you— what's your competitor's content and what kind of content can we influence, right? You can't influence your competitor's content, right? They, it's working like they have no reason to change it. You can change your own content. So maybe you can take some hints off of what's working for you currently. Currently. And then the last and best lever we have is third-party content. So these are guest posts and listicle placements. So we see that listicles specifically, third-party listicles, I like to call them like mommy bloggers, right? Oftentimes these are actually pretty small companies that do content like industry-specific content. And essentially these things, these these industry blogs are being cited by the LLMs very, very frequently. And my guesstimation is that it's because they're very, very narrow and very niche. You know, an example of a listicle that does well, there's this company called Marketer's Milk. It's a 3-person company. The CEO's name's Omid, and he has this one article. It's like 25 best AI marketing tools. Well, turns out that That single citation gets picked up almost every time you ask any question related to AI marketing in Perplexity. So it is, it is the most trusted source in Perplexity. So if you have a company that sells marketing tools, you better be at the top of that listicle. So that's kind of where we go from observability into like action and implementation, which is like, okay, there's this listicle out there. Like how do we actually get on that? And that is much easier said than done. So PR agencies have done this for decades. SEO agencies, they'll sell you, you know, networks of backlinks. And we're actually building something pretty similar to that internally. So a citation is different than a backlink. That a backlink was just a link. A citation is an actual block of text that has actual value to it. So when an LLM goes to that page that, you know, 25 best marketing tools, it sees your citation that explains what your company does and why it's so good. And so essentially the hard part there is going and actually convincing Omid G. in this case to, hey, put my company at the top of that. So Omid's going to want some money. So we are building the tools to do that, to actually go reach out to that company, go and broker a relationship to actually get a placement, because that's what's actually going to move the needle for people. It's not just going to make them feel good. Hopefully it does that, but it's actually going to get them more money. And that's, that's what we care about. Okay. So just to mirror this back to you, it's basically like find the keyword phrases related to your brand, uh, which is like, we've been doing that since the dawn of time. If you've done any SEO work, uh, once you identify those, you find the citations that are being referenced within those queries, uh, then you identify the owners of those blogs and you do cold outreach to negotiate with them. Them. What you're saying, just to be like very specific, is what you guys have built is basically like an AI agent that handles that negotiation process for the actual placements. Um, just to contrast this with like, you know, this is actually how we got introduced and originally was like, oh, I, I'm gonna pull these citations and go do this process. And then, um, you guys are just like building this now, which is amazing. Um, the, uh, I, I think the thing that people need to understand is that like, this is actually how you can like really quickly pull a lever to create a lot of impact. Um, can you give like, just like a, you know, give me some numbers on like, what does the first 60 days look like when you actually implement a campaign like this? Um, cause I think people don't understand like, like SEO traditionally is a very long, like long form time horizon, right? Um, this is the con— big contrast when you do this activity. Um, it's like relatively fast. I've seen 30 days, I've seen 60, I've seen 90 where it's just ridiculous. So talk to me about that. Like, what is it like? How many, and gimme, yeah, I guess what I'm curious about is like concrete numbers of like how many placements are you looking for? Like what's the tactical components of that and the actual process to like, you know, the numbers that you're trying to hit for this? Yeah, so this is really interesting that a lot, like in marketing services, people want retainers, long-term retainers. They want to have clients for a year, 2 years, 5 years. And what I have realized about GEO is that it is more of a project. So the actual implementation is more like a project. And then of course they observe ability like goes, goes longer than the project does. Um, but like you can do a lot of this very, very quickly. We were an agency, we actually used to do this, um, for our first client, um, a law firm essentially. Over 2 and a half months we had about a hundred and I believe it was 192% increase in referrals. So this was from about, you know, a handful of referrals a month. So about 100 referrals a month from ChatGPT. So, so our, our North Star is ChatGPT. I should say that like we, we care way more about ChatGPT than Perplexity, than Gemini. Is there rough estimates on like how much more search volume's happening there? Again, you probably see this data, so, you know, like I've heard it's like, you know, 1 to 10 from, you know, Perplexity versus like for ChatGPT. I've heard higher numbers. What do you guys kind of see in your data? Yeah, so, so I'm in my head right now, I'm looking at a real client dashboard for referrals per month. And this client has about 200 ChatGPT referrals with 3 Perplexity referrals. Nice. Right? So maybe that's a fluke, but like consistently when I'm looking at client data, like it's the same thing. It's like really the other LLMs are, you know, tiny little stakes in comparison to it. Yeah. People stakes, man. Not saying that it's not worth it specifically. Like if you're selling a product that's worth hundreds of thousands of dollars each year, well, heck yeah, you should optimize for Perplexity. But if you're selling widgets Hits, then go and optimize for ChatGPT. So we have seen consistently, like, this is, this is a very short process. This is like a month, 2-month, 3-month process. We've seen this. So we sell primarily to agencies. So we have the benefit of looking over their shoulder when they're doing implementations. And, you know, we've seen, you know, essentially agencies who can get 5 to 10 of these citations and they'll do that for several months. They can, they can like essentially double their referral numbers. So I'm being a little bit vague on purpose because I know this information from looking over someone's shoulder. I don't want to like— No, no, no, no. Yeah. And I'm just looking for like kind of like rough estimates. Yeah, yeah, yeah. Like, so say for example, you know, you obviously then there's, there's a lot of, you know, factors that come into this, right? Like how competitive is the industry, like, etc. Exactly. I'm just thinking about like, okay, you're a software company, you're trying to rank for some target keyword phrase that, you know, is super high value to your brand. Brand, do I need to build 10 citations? Like, do I need to get on 10 listicles? Do I need to get on 50? Do I need to get up on 100? Like, what, what does that scale look like? One. One. You need to get on the one. You need to get on the one right one, right? Like, it goes down, it goes back to the power law, it goes back to it, it goes back to investing, right? Like, like VCs fling 99 checks for the one that lands, and that's exactly what's going on here, there are 1, 2, maybe 3 citations that are the really big levers. Now what I've seen with— so they're weighted is a way to think about it. It's like a weighted— they're weighted. Each, like, so I'm just, just again, to put numbers on this, you got page 1 through 3, um, of those like 30, right? Um, that are, uh, basically being sourced or something like, we'll just, yeah, again, just trying to put a date on it. Like 2 of those are the ones that have the most weight, the most surface area in your specific category that you're in. Like to go back to your point. Like for some reason it really likes this website, so it references that website. So being placed on that website is super critical. It's very important. So, so I guess the question for you is like, do you only focus on that single one or do you build out as much surface area as you can? Like, do I go and try to get, you know, 1,000 placements, etc.? Yeah. So, so what we're seeing like with, with companies who take this seriously is they go and get those, those power law winners, right? They secure those at all costs. And this could be like, I mean, with clients, we've literally looked at placements that cost $10,000 a month. Month, which is just like insane. That is for the most competitive, essentially the most competitive industry in the most competitive location. So like, of course, it's like crazy expensive. But again, it's like those power law winners are worth getting those 1, 2, or 3 placements. And then there's probably 50 others that are still worth it. And so what we see with, with like really serious clients, so they go and they get those 3 and then they're going and they're chipping away at the rest of them. And then now here's the crazy part. They start moving into other verticals. So they went and got the top 3 for their main vertical. Now they'll go move into a cousin vertical and they'll do the same thing again. And so they're, they're essentially buying up verticals one at a time. It's very, very— the people who do this are very serious. They have lots of money and like it is the company mission to go and make sure they dominate in, in Geo. That's so fascinating. So it's, it's, it's really basically the process here ends up being, okay, I find these power laws. Do you guys map that or how do you, how do you identify that data? Basically you're like, cool, like you're, this is these, uh, you know, these links are what are being referenced most. We, we see that this has the highest weight, so like this should be your, your biggest focus basically. Is that how you're going about the process? It's like you make almost like, I'm imagining like you know, a mind map, whatever you want to like use to visualize this. And then from that you go and, you know, rank stack these on which are going to be most impactful for the brand. Is that the right way to think about it? Exactly. So, so again, we start with, with your brand, essentially your strike zone. And then where we're going with this is, right, we have millions and millions and millions of these citations. And so we know essentially what the, the next set of citations you should get. So what are the cousin citations you can go and get after you go and you knock out those first ones? So we're actually building is so, like I said, you know, SEO agencies used to have this network that they would sell you for backlinks. We're doing that with GEO. So we're building that network and it's going to be crazy powerful. I'm so excited. Dude, that's awesome, man. Okay, so citations, most important. I mean, I, you know, I'm sold on this. I think everybody that's listening to this is now sold. What are the other levers that you can pull beside, like, should you be spinning up separate sites once you identify these? Um, should you be, is there any, is there, uh, have you seen anything move the needle that is outside of just doing basically this outreach process? I'm also curious just to like, you know, this is literally a job function or a job role, and I'm curious about the business model here too. Like, this is something that we think about a lot, like at Graft, right? Is basically, you know, can I take a job to be done and automate that? Or like literally a job job function and like, you know, that salary, can I eliminate that? Right. Um, I, I, so talk to me about those two. So again, just to, uh, restate it, what it, what, what are the other things that you're seeing actually work? Like, should they be building other sites? Should they be astroturfing this shit? Or is, or is it just literally focus on getting in citations? That's the best thing they can do. Yeah. I mean, I mean, there's a critical path and, and that critical path is citations, right? But if you're going after citations, that assumes you already have essentially good technical SEO. Like good fundamentals, you have good content, right? If you don't have that, then like, you know, like, yeah, you could go after those citations, but like, is it— it will move the needle, but maybe not as much as you would expect. So like, we're looking first at the easy things to fix, right? Are your schemas really descriptive and really good? That's crazy important. And like a lot of clients actually don't have that. So like we're trying to find the easiest things to fix first. We kind of like, you know, it's like boiling a frog. Like let's start with like baby steps before we get to like full-speed running. Um, and like, in my experience, very few companies are very, very like extremely serious about this, ready to go and take the leap. And so starting with those little things where they can see a little bit of increase is important, more for making them feel good. Um, but, but really the way, the way to win is those citations. So you were asking about essentially building out separate sites. This is an interesting— this, so, uh, it's an interesting, uh, method that we have not done anything with, to be totally honest with you. We've looked at it. I'm just thinking about like private blog networks, right? Like we used to do this shit in the past. Right. Um, and I'm just imagining that before this, but I guess my question is like, will the LLM even see that? Like, is that, is just making surface area like valuable? Is it, or is it more like I should focus on the surface area that's already there and basically cannibalize that? Yeah, so the first we're looking at is can you piggyback off of essentially content that's already working, right? Can you piggyback off of someone else's site? Can you just pay them money for it? Can you pay them $1,000 a month, right? It's going to cost you a lot more than that to essentially build that out yourself. But where we tell clients otherwise is when there's not any opportunity to piggyback, right? So if we're looking at like, you know, if we're helping like an HVAC company in Missoula, Montana, right, there actually might not be a listicle in, in that area, in that geo, um, that, that they could go and hijack essentially. And so that could be an area for them to go and make that. Specifically us, we haven't got to that point yet. Typically we do actually find like there's, there's like pretty, pretty gran, uh, granular like content placement opportunities. Um, and so yeah, you could even like, if you were in a geo that didn't have that, you could build that yourself. And that goes back to like my un— my understanding of LLMs picking the most niche the most local content, there's actually, there is a lever to be pulled there. We just typically don't get there. Yeah, no, that makes sense. It's funny, I'm just thinking about, I had, uh, uh, my friend Jesper on the show like a couple weeks ago and he basically was sharing, he does a lot of parasite SEO, but specifically for local businesses. So it's like, how do I get, you know, a Claude, um, output to rank? How do I get an Instagram post to rank, et cetera? So I'm just imagining that where it's like, cool, yeah, we can't find anything to actually influence the citations. Here, so let's just like go and yeet out all this content onto these parasite platforms. We'll do like a drip ping link to this. Uh, you know, so we have just like drip traffic going to it to actually get it indexed and seen by Google. And then that suddenly influences the, like the, you know, the geo for those responses. I'm curious, like the companies that you're seeing this be the most successful with, like what I've heard and, you know, we, I've had a ton of conversations around this is like, like actually home services is ripping, e-com not as much, but like starting to become more like in the, like if you're looking like supplement research is happening there, which can, you know, be high profit, like long, you know, customer lifetime value, like larger customer lifetime value. And then software where there's like any purchasing decision that has like a longer time horizon. Is there anything else that you would add to that that you think is that this is a good result for? Because how I'm thinking about this just for the audience, like a lot of our, the people that listen to the show show, like their founders, their heads of growth, their heads of marketing, like they want to just go and go ham on this, right? Like what, what are those, I guess the, like how, how do they be as aggressive as they possibly can, I guess, which is probably the contrast to like a lot of the agency clients that you guys work with where, and also again, they'd like, you know, come to you guys and just be like, yo, turn this on for me basically. Right. Which I know you've been talking about this, like making this, this offering, but anyways. Throwing those, those pieces out there. I'm curious to hear your thoughts. So you, you had a question, uh, a couple minutes ago, which was like the job, the job to be done. Um, I want, I wanted to like kind of hit on that, which is like, we're, we're hoping Geo is not a job. Like it should, like, we're hoping to make such a leap. Yeah. Yeah. Just, uh, evolve past all, all the grunt work. Right. Like we, we, we think humans are more valuable than that. That. And we think we can just automate it all, essentially integrate with our systems and we'll take it from there. And so that's, that's really our north stars. We actually don't want it to be a job to them. We don't want you to think about it. We just want you to win. We want essentially the people who pay us lots of money to just hella kill. So, um, that, that, that's, that's essentially, um, that, that, that's kind of where we're going. But as far as like what, what kind of industries are actually worth it, I have actually turned down so many startups who have asked me to do this for them. I get pinged all the time, please, please, please do this for me. And the first question I ask is, hey, have you ever tried Facebook ads? Because you should start there. It's a lot easier and you can get a lot more signal a lot faster. And so this is definitely not something for everyone. Now, what the people— The people who want arbitrage, So like, is that like, are they, 'cause again, like what I'm always looking for, right, is there's always a moment in time when there's a platform that's basically like underpriced, right? From a distribution standpoint. Are we in a moment where that exists? And like, should companies even be fucking thinking about this? And like, this is what I'm wrestling with too, right? Like I have resources, everybody has limited resources. It's time, money, mental energy. Like where can I basically allocate those to get them, you know, the biggest return returns. And you know, this is what I'm trying to get to with a lot of these conversations is like, is this, you know, should this be even be focused on, right? So anyways, again, curious your thoughts on that. No. Yeah. So it really goes down to what you're selling, how much money your customers pay you. And then how, you know, what are your other acquisition channels? So the companies that love geo, that it's worth it, it's worth it for them. Or typically like B2B SaaS, right? Or high, high pace, high pay e-commerce goods like you were saying, like, you know, skincare supplements, like things with high lifetime value, right? So like a company I used to work at, like their, their average contract size, it's like minimum quarter mil a year and they have a very crazy niche audience. And so their, their audience is is economists at basically publicly traded companies. So like very, very hard to get in the room with. Well, these people are early adopters of LLMs. They're on ChatGPT, right? They have very, very complex questions. And so that is the obvious first place for, for this, um, data company, uh, a big, uh, an obvious lever for them to pull to get in front of their audience. So a big question like, is your audience there? Can— if you like talk to 100 of your customers and you ask them, hey, how often do you use ChatGPT? If they started saying like, what's ChatGPT? Probably shouldn't do this, right? There's still a lot of industries where that's the case, right? And I would go back to like HVAC Missoula, Montana. I'd be like, I don't know, do people in Missoula, Montana use LLMs? Now if you said that in New York, I'd be like, absolutely. Um, so it really depends on the company and their ticket size, their appetite for risk, right? There's still risk in this, right? So like some corporations are just like, nope, sorry, like jury's still out and we're going to, we're going to, we're going to sit this one out, right? We don't want to take any risk. There's plenty of those. So yeah, do you have the appetite for risk? Do you have the money to do it? Are your customers there and are they worth it? I think what you said is so valuable, like just customer conversations, like how, like, are you using this? Like, are you using ChatGPT, right? For research and relationship to whatever it is that you do. Like if the answer is yes, then it's like, cool, I need to spend time there. And this is like such a simple principle. It's like, where are my customers spending time? How do I get in front of them with a story that like they're gonna find interesting that's talking about their pain points or an outcome that they're trying to accomplish? And then, you know, figure out how to do that in a, in a novel way that doesn't feel like like the rest of the market. So man, what am I not asking you? Right? So like, I feel like this space, there's so much to go in on, but what are the questions that I should be asking you or the things that like are keeping you up at night or the things that you're excited about now with like, you know, being on the forefront of this? I'm like, I'm very excited for ChatGPT ads. Like no official announcement. I mean, I know people who know people kind of thing. Um, but I, I am excited for, uh, the godfather of advertising platforms. I really hope they pull it off. Um, it will be very good for my business as well, right? Like, we want to do the one-two punch of— like, the, the one-two punch is like essentially organic and paid, right? And we want to do that in LLMs. So we're, we're like very excited for that, uh, for that future. Um, and just being able to get a hold of these, essentially these hard to get a hold of people, right? Economists in like publicly traded companies like that is an insanely hard ICP to get in front of. Virtually very hard to get in front of virtually. And I think ChatGPT has, has a good promise. So I'm specifically excited for that. I think some companies, some geo companies are like kind of like hesitant. They're like, you know, this will take away from us. But I just look at this as like the whole TAM's expanding, like the whole market's expanding. It gets, it gets, it gets bigger, better, faster. And like, that's a good place for us to be in. Totally. It's really interesting on the ad side. I have a friend who's building this company. I keep telling him like, come on, come on, come on the show. And he's like, not yet. We're not there yet. Like that type of thing. But anyways, what they're doing is they basically are like the third-party ad seller for all of these, like, you know, like free chat tools. So imagine like, you know, there's thousands, you know, hundreds of thousands of these like ChatGPT clone wrappers that use like open-source model LLMs and then like, you know, some type of scraping technology to like give Results and Answers. So basically what they're doing is, um, they create, like, they're understanding the sentiment of a person's research that's happening on these, and then they go to an advertiser and they're like, cool, what's your, like, you know, your CPA number that you want? It's like, oh, we need, like, we'll pay X for a new customer. And he's like, awesome, we'll always get you new customer for X price. Wow. And then just serve it, which is crazy. And so I think that's like, I don't think people understand like how like good this is going to be. Like the, the, the understanding of the user's intent and where they are in the buyer's journey. You're gonna have all that contextual knowledge to be able to like get them to make an action that provides value to the brands and to the companies. So it's just, it's crazy where this is going to go, right? From like an ad platform standpoint. Also like Google has all of this information, so it's like there's 100% chance that they're gonna probably do this and win in, in some capacity. I think that's the funny thing too. Everybody talks about ChatGPT and it's like literally like Google is absolutely still in this race. You should be thinking about them. How do you communicate that actually to people? Or like, you know, are you only focused on ChatGPT or are you also thinking about like any of these other, you know, like any of the Google-owned properties? Yeah. So, so I want to say like competition's good. Like Google will have a better offering being essentially challenged by ChatGPT and, you know, Perplexity and Claude also will, will essentially like all standards are about to rise. And that makes me very excited. Excited. Um, so like how I think about, uh, Google, Google AI Overview, Google AI Mode, um, is that it's, it's like increasingly becoming— they're rolling out to more and more people. Um, that's very interesting. So there's a lot more eyes on it. Personally, I have not found it to be very valuable. Um, I have not found Google— I, I read a Google AI Overview maybe for the first time a few days ago, and I just thought like, man, this kind of sucks. Like, I don't know, it's not, it's not tuned in yet. It's not, it's not that good. Um, that doesn't mean it won't be good. It may be the version that I'm looking at. I'm looking at like the crappy part of an A/B test or something, but like, it wasn't very good. Um, I hope it gets better. I'm sure it will. No, totally. Yeah. I kind of have the same feeling about it as well, where it's like it doesn't provide the same amount of value. I also think like a piece of it too is like I just use it in a different way. Like when I'm, because of the, like the app and as it learns my preferences, like that's this whole other piece where like conserve that information. Yeah, man. Any other, any other things you're excited about in this space or this category that you think like the audience should be thinking about? Again, these are all like people that have decision-making power. Like what are, what are the, what are the, what are the things that they should be like contemplating in relationship to this, this category, this space? Yeah, I mean, we've been doubling like essentially daily use has been doubling in ChatGPT for the last 2 years. Um, so when we go from 12% to 24%, um, all of a sudden we're at a quarter of US, uh, essentially US searches. That is going to be crazy. Um, and that is going to make everything very competitive and very expensive, right? These listicles that, that we're getting customers for, you know, $500 a month, it might be $5,000 a month in a year, in 2 years, right? Because they will be, have been proven to be the winners, um, and everyone will want on them. So, um, it's it's a great time to be in content actually. If you can spin up some of these very niche blog articles and improve— I have a friend doing this, by the way. He was basically like all creating these, getting citations, and then like selling ad slots on it. Just printing money. It's just being basically like a degenerate affiliate. So it's so funny to hear that. I want to meet him. So let's connect on that afterwards. But like, I would love to meet him. No, for sure. For sure, dude. Thank you so much for coming on. Sean, where can people reach out to if they want to find you? We'll include all the links below, but what's the best place to get in contact with you? Yeah, you can find me on LinkedIn. You can visit our website, Aldill AI. You can find me on Twitter. I'm Sean spelled the right way, S-H-A-W-N underscore TXT. Yeah, feel free, shoot me an email, you know, follow me on Twitter, LinkedIn, whatever. Happy to talk. Talk. Awesome, man. Dude, always a pleasure to talk to another person from Idaho. I feel like there's this, like, it's a rare— Idaho is a small town, so it's always funny to know, like, see who everybody knows in relationship to this. So we'll have to get you to meet Kobe Conrad and the Conrad brothers because they're out in SF and doing gnarly shit. And anyway, so are they from Idaho? Yeah, they are. They're both from Boise originally. So anyways, yeah, man, thanks for coming on. We'll have you again, back again in like 3 to 6 months as the, you know, stuff gets along. We love having friends of the show just come tell us about everything that they've learned in that period. So anyways, love to see you soon. Yeah, that gets spicy. All right, thanks man. Cheers. Yeah, cheers, brother.