What GEO audits are revealing about manufacturers in AI search w/ Maya Harrington and Chelsey Trevino | IMC Live

Speaker A: M. Foreign. Welcome back to IMC Live interactive series hosted by the Industrial Marketing Collective. I'm Alan Finnawald, associate content director at Guerrilla 76, the industrial marketing agency, and one of your IMC Live hosts, along with Peyton Warren. Um, so some little, uh, details about what we'll be talking about today. Uh, GEO audits. You might have heard of them. Geo, AEO ago, aio. All the acronyms involving AI and how it is affecting our buying processes. We might be, um, a little obsessed with that here at gorilla76 right now, and we don't think that we are the only ones, or at least the only ones concerned about it. Our clients definitely seem to be, um, everyone's excited or worried about what's going to happen next, and you know, how they can either stay ahead or catch up. And that is why we have two of our strategy extraordinaires here at Gorilla 76. Maya Harrington and Chelsea Trevino. Thank you for joining us. Uh, could you, uh, introduce yourselves to the folks here before we dive into the topic today, Chelsea, why don't you go first?

Speaker B: Uh, sure. Yeah. Like Alan said, I am a strategist here at, uh, gorilla 76. Uh, prior to working at Gorilla, I actually was in house marketing for Manu company, uh, that did a lot of medical soft goods. Uh, so really excited to be able to bring this new topic. Uh, well, not new, but, um, you know, it's getting bigger and it's worth discussing.

Speaker A: Absolutely.

Speaker C: Awesome. And I'm Maya. It's nice to connect with you all this morning. I'm a senior strategist at Gorilla, and I've been here for about two and a half years now, um, working with a lot of different industrial manufacturers. Um, and I love what I do. Um, and prior to this, I was at NB2B SaaS marketing. So a little bit of a pivot, but, um, have been loving the industrial world.

Speaker A: That's convenient for today's conversation, for sure. The SaaS audit or the SaaS angle. So, Maya, why don't you just, uh, dive on in here, um, and tell us why is now the time that we should be worried about running a GEO audit?

Speaker C: Yeah, I mean, I don't. It's funny you say worry because, I mean, I think it is a bit of a big kind of ass, kind of a bit of an unknown for some people, but, you know, at the very core, like, we're all using ChatGPT, Gemini, Perplexity, Claude, whatever it is to, you know, be a research assistant, be a thought partner, just talk to about Whatever it is, personal and professional. So it's just fair to say like our prospects and our um, customers, anyone in our audience really who we want to reach, are probably also using an LLM to talk about their work, their life, whatever it is. Um, so you know, it's just going to be. Doing an audit is a simple practice of like understanding how those platforms are representing and like talking about our brands or our clients.

Speaker A: Chelsea, anything to add there?

Speaker B: Um, no, I don't think so. I think, I think Maya covered it. Um, you know, I think it's important that, to note though that the audit is a great starting point. Um, but it's not necessarily the hard part of the process. It's prioritizing what's going to come next after the audit.

Speaker A: Absolutely. The follow through.

Speaker B: Yeah.

Speaker A: And so uh, what are the common pitfalls that people need to be aware of before they dive into this process? Like what are the things that they need to know there is it, you know, things about their competitors, their keywords.

Speaker B: Yeah, I feel like I kind of just mentioned one. Maya, uh, do you want to tag on?

Speaker C: Yeah, yeah, I mean I think just having a good sense of like the business context, what you're trying to get out of the audit before you go in and it just spits out a whole bunch of like, you know, all about your content, your technical like, aspects of your website. Like it's a lot that it uncovers. So just having a clear vision of like, you know, what is the most important question I'm trying to answer about this? Is it like brand visibility? Is it, you know, how are people feeling about our products or you know, what's the sentiment about our brand? Um, kind of having an idea before you go in is good. Uh, so that you, you are not just overwhelmed with the results.

Speaker A: Mhm. And speaking of results, now you, you went through this with a industrial uh, actuator brand. Um, so could you, you tell us about what this process revealed? Um, and you know, like what was valuable about it so that people, you know, why they, why want this on their end?

Speaker C: Yeah. This example of um, one of, one of the audits I did was for ah, a newer client. They came in, they were an industrial actuator like Alan said. Um, it's a bit of an unknown for them. You know, they had a lot of, they have a lot of scattered marketing efforts and they're just not sure like how is that all coming across in these LLMs. So maybe they had some, they were worried about that, you know, coming in. Um, um, but the GEO audit was, and this has been kind of a common thing across my clients too is that it's actually not, they're not as bad as they think they're doing. You know, everything they do and put out there is, is contributing in some way to, you know, what the LLMs are saying about them. So it, it revealed they're not like completely invisible. There's just, you know, you know, they were showing up for a couple things they wanted to be known for like specifically their motor technology. Um, but they weren't being associated with like the industries, some of the newer industries that they wanted to, that they want to want to reach, like power gen for example. So we'll go into this more on like what the GEO audit entails. But um, something that it creates is like a prompt universe. Like what are some prompts that our audience might be asking LLMs? So any of those prompts associated with the power gen industry, for example, they weren't showing up for. So that just helped us be like, okay, let's prioritize content for that industry. Um, and it specifically narrowed in on that industry, not the five other industries that they, they were doing okay with. So helped us prioritize a long form, a long term and long form content plan.

Speaker A: Great. Yeah. And like you said, we will talk about that more in a little bit. But first Chelsea, why don't you tell us a bit about your experiences with a polyurethane ah, foam manufacturer.

Speaker B: Yeah, so I've run a few audits and had similar experiences to Maya with a lot of them. But there was one that really stood out to me, um, because it was different. And with this client, um, if you type their actual brand name in like who is X brand? All the LLMs could tell you this is who they are, this is who they serve. But if you asked any real questions a buyer might ask, like what's the best polyurethane foam for orthopedics or the vibration dampening polyurethane foam. Um, they were nowhere there. It wasn't so much that LLMs didn't know who they were, but there was no association to the problems they solved, their industries. There was no connection there. Um, another interesting insight of that one that I thought was very intriguing was that the whole category for polyurethane foam was defined by one competitor brand. And all LLMs, like AI, treated that competitor as the generic term like Kleenex is to tissue. And, and every other player got described as either a converter of that brand or an alternative to that brand. Except our Client. Our client does make a legitimate direct alternative, but it was never framed that way by any LLMs, partly because there was no content or authority that framed them that way. So there was a huge lack of content and topic authority, um, across them. It also surfaced, um, a kind of assets hiding in plain sight. Their website had a lot of authoritative technical content. It was just buried or unstructured and also trapped behind PDFs that AI crawlers are told to ignore. Um, so again, like, like Maya said, we'll get into kind of the steps that we see as like, the priority. But this was really great starting point for this client. On here are some steps that are very clear, actionable things that we need to work on fixing.

Speaker A: Yeah. And, um, speaking of. Peyton, you got something?

Speaker D: No, I just, I just wanted to jump in because, uh, so far, uh, the themes that I'm hearing are

Speaker A: what

Speaker D: content do we need to create? And I guess I just want to give everyone a vote of, like, you can do this. You can handle, um, you know, geo and optimizing your company to show up better in these platforms. Because so much of it really comes back to just the foundational stuff that all of us in industrial marketing really understand about our, um, buyers. And it's that they want helpful, valuable, detailed content. And, um, the, the. The. One of the really nice, like, benefits of all of that is that, um, that's what the LLMs want to, uh, you know, they want these, um, these specific, um, details that all of our, you know, all of our customers are craving too. So it's like just more of the same. But these audits, what I'm hearing is like, they're giving you a lot of direction, um, and insight into what those questions would be. Whereas before maybe you would have to get it from a sales conversation or like listening in on a sales conversation, uh, or customer interviews, you know, the social listening. It's like another channel where you can be listening and testing, um, to see how you show up in queries that you think you should be showing up for.

Speaker E: So.

Speaker D: Pretty cool.

Speaker A: Yeah. And speaking of showing up for queries, uh, that is where your content tends to really come in. Uh, so Maya, uh, in your experience, what are the, uh, most important things about having quality content to be able to show up in these LLMs?

Speaker C: Yeah, I would say one thing is using, like, the bluff method. So bottom line up front, this is opening with a direct summary to a core question that the page is going to answer, rather than burying it at the bottom of the page. And having that title really Match what, what they're going to get out of that content. And then also many of the audits are usually all the ones I've run. Always recommends foundational educational content. Like Peyton said, um, like going back to my industrial actuator client, like an electric versus pneumatic actuator, a rotary versus a linear actuator. Like they're very technical. Like there's a lot of those direct comparison questions that are very obvious. But the audits also generate other prompts that are a little bit deeper, you know, less surface level. But um, this foundational content really helps models like build confidence in what your company knows. So even if it sounds very like top, top level, very high level or basic content, it's going to help the LLMs, um, understand like what it should associate you with.

Speaker E: Mhm.

Speaker A: And Chelsea, how do you make sure that everything you have on there is accessible? We just had a question in the chat about PDFs. You mentioned that. Um, so yeah, what are the best practices to make sure that all this work you put into having this content doesn't go to waste because the LLMs just can't read it?

Speaker B: Yeah, so I was actually just responding to the comment. So it's, it's not that PDFs aren't readable for this specific client. Um, they had back end technical data that was telling AI crawlers basically to ignore those PDFs and it is worth like also poor formatted PDFs are highly difficult to read. And so if you're following accessibility standards for like PDFs on the web and making sure they're accessible, they're going to be accessible as well for LLMs. And then it's just making sure that you have the technical markup on the back end that is clear that to tell the robot, the AI. Yes, to scan these as well. Um, for this specific client, their technical markup was just telling crawlers to ignore basically the PDFs that were on the page.

Speaker F: Mm.

Speaker A: And speaking of markups, I hear about these schema things. Uh, like they're like things that are computer language or something. I don't know. Chelsea, what, what, what's the schema markup stuff about?

Speaker B: Yeah, so that's nothing new. And some of you may know, know what it is. I actually did not really understand it before I started these audits. So I spent some time talking with our uh, Creative Digital Director. But schema markup is something that is a structured code in the background of a page that labels what the content actually is on the page. So it Tells a machine or a browser. This is a product page, this is an FAQ page. This is a how to guide instead of making the crawler guess from the raw text and scanning through it all. In almost every audit we've run, we found that the technique, like the proper schema markup is basically non existent. Um, an analogy that Randall, our Creative Digital Director, gave me is that it's like you label boxes when you move. All the stuff is in the boxes either way. But if a box is labeled as kitchen dishes, it allows you to put them in the right place faster and not have to start opening everything and go through everything to find what you're looking for. So schema labels your content so the LLMs don't have to try to infer and guess what's on the page where, you know. We already know that a lot of LLMs have had difficulties with hallucinating things. Schema markup just makes sure that it understands from the get go exactly, uh, what content is on your page. Um, oh, I just kind of mentioned like, so what this means from you guys for anybody specifically is that if you don't have proper schema markup, LLMs can misinterpret what you do without those labels. The models try to read your page cold and guess the intent of the page within the two seconds that it scans the page. Um, schema markup helps make your really good content extractable. It's easier for AI to lift a clean, correct passage and cite it for a proper query that comes in. Um, I do want to make the note though that schema is not a magic wand. You can't tag your way out of bad content. So if the underlying page structure is weak, schema is not actually going to help save it.

Speaker A: Yeah, and you know, I don't want to tell people that schema is not important, but in our experience there is, there is also success without schema with our own website. Um, because we have been appearing fairly well on the LLMs and such. But we have what, no schema markup as of now on our site. Is that correct?

Speaker D: Yeah, yeah. I mean, we're working on, uh, redoing our website, which, you know, we shared a little bit about earlier this year. But, um, we've never prioritized schema on like the gorilla76.com um, website. And does that mean, is that good? It just was never really that important. Um, and folks have said in the chat like, schema is not new. It's not. Especially if like you were trying to sell products, um, show up in like Google Shopping or like trying to really optimize in like a very technical way for SEO. Or you've worked with like an SEO agency in the past. Like you might have some of this already like built in. And just how depends on how your developer kind of set up, um, your website. Um, but it is becoming like Chelsea was saying, it's just, it makes it easier on the LLM. And I think everyone uh, is becoming more and more aware of how um, like tokens run out and you want to be token friendly, um, when you're talking or uh, when you're, you're kind of getting crawled by these LLMs because they'll just stop at a certain point. So that's why you hear advice like you know, do a blurb at the very top like too long, didn't read sort of thing. Or now folks are calling it like the bottom line up front. Um, but um, yeah, so it's like you want to make it easier for the LLM to crawl and that's where schema really shines. Um, yeah, we have had a lot of questions about PDFs. Um, I want to make sure that folks are um, getting, getting answers about that because, you know, PDF is a, uh, very common, uh, I think like asset that comes up in the industrial world. Um, and uh, I think folks, uh, we had a question, um, about like if folks have to submit their information to gain access to a PDF, is that going to be readable in an LLM? Um, being one of them, I'm just going to kind of volley that out, um, to the crew here.

Speaker B: Yeah, so if you have content that is gated behind having to fill something out that is something that LLMs most like the M most commonly cannot access, that is one of the crawlers on the back end that would tell it, this is inaccessible content to you. So it's LLMs, you know, running these sort of things. Um, actually is a proponent for ungating content and having it accessible to everyone because then it becomes accessible to the LLMs as well.

Speaker A: So how do you think about increasing your visibility in the LLMs or your market share or defining your voice to be more, um, accessible or more referenced by the LLMs. Is there anything specific that people can think about that when they're creating their content?

Speaker B: Uh, get specific. Um, I mean if you're searching for tennis shoes, you're going to get hundreds of thousands of responses. But if you're searching for tennis shoes that help with arch support, it narrows it down. And so don't focus on broad terms, um, get really specific on the, the niche topics that you want to own. Um, and that's going to move the needle a little bit better.

Speaker E: Mm.

Speaker A: Mhm. So instead of screwdriver, maybe it's screwdriver for underwater welder who also needs a magnet tip.

Speaker B: Yes. On that topic.

Speaker A: Mhm. Maya, what do you think? Do you have anything to add there?

Speaker C: Yeah, I was going to say something that um, these audits will reveal too is like the uh, like whether your brand is associated positively or negatively with certain topics. If they're asking about, you know, reliability, are they asking about lead times? You know, so it'll reveal like is it a positive or negative sentiment? Right. So you can create content to help, you know, break some myths about your brand, things like that. So, um, another kind of like Bain was saying, just audience research, um, what are people thinking about you? Um, and what are the LLMs thinking about you?

Speaker A: Mhm. Yeah. And what questions could you answer better than anyone else that is currently using the LLMs?

Speaker C: Exactly.

Speaker A: Yeah. So we, we talked about these audits for a while now, but I realize we have not even told people how to audit. So Maya, uh, how, how do you even do this? What, what do you use?

Speaker C: Yeah, yeah. So I'm going to kind of talk about really like out of the box solutions. You know, there's so many platforms out there, um, that are, you know, popping up because of the, the AI bubble. Right. You know, so we have a couple of tools that we use at Gorilla. By no, by no means do you have to use these tools there, out there. Um, HubSpot also has an AEO tool if you're on HubSpot. Um, but the ones we personally use at Grill at least are Visto and Semrush. Um, Semrush has like a, we already have it for SEO, so they have a plan that gives you visibility to AI, your AI score as well. Um, so that's not really where we want to focus is tool specific, but there is that route to go and then Chelsea is going to share more about um, like a more LLM based route. Um, but I'll kind of go into those platform specific things and what you might want to like have handy to make your life easy and make the audit go quickly. Um, because it's not the audit itself doesn't take too long to set up or anything in these tools. It's really the outputs that are important in kind of putting that into context for your business. Um, and the action steps coming out of it. So to have Handy to set up these audits. And if you have access to one of those tools, and it's always going to ask you for, um, at least these tools will ask you for your brand name and alternative names, acronyms for your brand nicknames or common misspellings. So it can track, like when you show up for that visibility score. Um, it'll ask you for competitors, so the other competitors you want to want the tool to scan that also show up for like similar prompts. Um, and then it'll also ask you, each one does it a little bit differently, but basically product category, context. So like Vista will ask you for a bunch of keywords so you can pull like your Google search keywords or like the organic terms you're trying to rank for maybe, um, just so it has more context into what category you play in. Um, and then like HubSpot already has a lot of your, a lot of your context in there. So it asks for like the product, the products and the ICP that you're, you're targeting. Um, and it asks you all that. So it can generate prompts. It generates. HubSpot does like 25 prompts for you. It, um, basically uses AI to generate prompts for you, um, based on those inputs that you just gave it. So it can get a sense for what you're trying to compete on and what you do and how well you show up for those things. Um, and then it'll show you all the prompts that it's going to run for you. So you can delete some or add your own things like that. Um, and then basically you hit, hit, go and it spits out, you know, a whole bunch of, you know, visibility metrics. Um, when you show up versus your competitors, like citations, mentions, just a bunch of like, depends on the platform. Again, um, but yeah, the best tool I think is just going to be one that, you know, you have access to, one that you're going to use. Um, and as long as you're confident, like, I use a couple of different ones just to see, like, does one tool say one thing versus another one? But it should just give you the confidence to answer, like, what areas of expertise you're trusted in which competitors are cited over you. Like, who are your top competitors in terms of like, maybe go check out their content since they're performing better than you. Um, and then like, what associations are missing. Like we've talked about, um, if you want to be showing up for your motor technology, but you're not in those answers when people ask about motor technology Then you know, you have to create more content around that. It's pretty, pretty straightforward and you know your business best. So, um, adding all of that context to the audit when you think, you're thinking about what to do next is something that only you can do. Right. The audit's not going to give you all of that. Um, so that's really the more of the work is after the audit, um, and which first steps you're going to take.

Speaker A: All right, thank you. And before we move on, I believe we have an audience chat that we want to bring up. Right, Kelly? Kelly, you want to talk to us a little bit?

Speaker D: Yeah, Kelly, I was seeing your chat, um, as you, as you kind of get yourself uh, unmuted and stuff, just about your experience creating an AI info page specifically with schema markup. And I just love, would love to hear the story. I know a lot of folks here know Kelly. Um, but Kelly, if you can tell us a little bit more about where you work, what you're doing, the page, all that stuff.

Speaker E: Hi, can you hear me?

Speaker D: Yes. So good to hear you, Kelly. I miss you.

Speaker E: So good to hear you. What's up? I'm working for Cadenus usa currently doing their social media, content marketing, that sort of thing. We're a two person team, but we just happened to see uh, like a LinkedIn post about creating an AI info page because that was. We were kind of doing AI audits through tools like HubSpot and we were having some um, problems with all these different, you know, products and brand name changes, that sort of thing. So the AI really wasn't sure who we were, what we did exactly, especially compared to our competitors. And so we just went old school and just kind of googled what to look for in an AI info page. We looked at a couple of examples and we built it with the schema in just like a couple days. And our Semrush rankings went up fairly quickly actually.

Speaker D: So when you were building out the page with the schema, um, had that been something you'd done before or were you kind of learning as you.

Speaker E: No, I had never done that before. I didn't know what that meant. But um, we're, I'm still trying to kind of figure it out. But ironically we asked AI to help us build the AI info page and explain what kind of a schema was. And we use WordPress and so we, it was fairly simple to just go on the back end and then put in all the text that we just wrote in like a Word document that sort of Thing.

Speaker D: That's great. Yeah. Kelly and I saw you dropped a blog post about this, a link in the chat. So I'm going to grab that link and drop it again at the bottom. Just so if anyone wants to read a little bit more.

Speaker E: That was kind of our, like, main guide. And then HubSpot has like, an AEO audit tool and like a full video tutorial. So we've just kind of been going off of that so far. But tbd, stay tuned.

Speaker D: But you have seen your rating, like your rankings go up when you look back at semrush. So, I mean, yes, so far, so good.

Speaker E: Just, actually, just yesterday we were talking about our Semrush rankings went up. Like you could even just physically see on the graph the date that we posted the AI info page and where it. It started to go up.

Speaker D: So, yeah, we've definitely been getting a lot of inquiries from, um, you know, when we, we asked that golden question on our form, uh, you know, our consult form, like, how did you hear about us? And, uh, we definitely see a lot more of ChatGPT or Claude or whatever, um, uh, coming through. Um, and so, I mean, like, qualitatively too, we've seen like a lift, but we've also seen some misdirection in the beginning that has inspired us to create different content. Um, you know, Guerrilla is not like an a la carte agency like I would imagine. A lot of your companies don't just sell a la carte products. It's like a service, a, um, a program, a full package, um, of working with you. It's not just like they can come and buy one little widget. Um, ah, I saw Rich Ward on here. So it's like, could someone come and just buy one mat? Maybe. But also there's like more of like a service component that needs to go into, um, selling, you know, the system of how your product's going to be working really well. So we were having a lot of folks just reach out to us on like, this kind of like baseline level, looking for all CART services. And that's not really what we do. And so that's not an ideal fit lead for us. Um, and we were getting almost like spammed, you know, with just like all of a sudden, like, people were finding us through these LLMs and they were saying, oh, gorilla can do SEO or email or whatever. And it's like, yes, we can do all of these things, but like, it's not how we do them. You know, it's not just like you can just, oh, I want to hire you for two weeks to run our Google Ads. Like, that's just not how we work sort of thing. So we had to shift some of our content and edit some of our content to be more clear, um, so that the LLM wasn't misinterpreting what we actually do, um, if that makes sense. So a little qualitative side to Kelly's quantitative side about, like, the. The lifts.

Speaker A: Yeah. This reminds me of when we, you know, were setting up campaigns and LinkedIn and you didn't set your audience right. And then all of a sudden you get a million leads for something that isn't a completely wr.

Speaker C: It.

Speaker D: It really does go back. And this has come up. Dale mentioned it long tail, keywords. It's like. It's just like, are you talking to the right people? Are you using the right language? Like, you know, audience is everything. And, um, it's the same. This is just a new channel. And these audits are going to give you so much detail about, like, just things you could do. Um, so. Yeah, but you don't just have to use tools. You can use the LLMs yourself. Uh, just like what Kelly was saying about how she built the schema page. So that segues us really nicely into, um. You know, you have another option to run the geo audit. You don't have to necessarily buy a new tool. Maybe you just use something, um, like the LLM itself. Right, Chelsea?

Speaker F: Yeah.

Speaker B: Yeah. So the audits that I have run, uh, lately have been using CLAUDE skills. Um, a caveat. I did not build the skill myself. Grace Wright, our VP of innovation here at Guerrilla, is the one who built the skill. Uh, so that's not something I actually know how to do, but maybe if people are interested, maybe that's something that we put the zipped file of the skill into the collective. Um, if people are using Claude. Uh, but I added the skill to my claude, and then it runs the audit on a defined methodology, uh, based on the skill itself. Uh, you do input things like your website, um, competitors. So you give it very similar information that you would give any of these other tools, like Visto, HubSpot, or SEMrush. One big difference with Claude, though, is that you also give it access to every file you have. So for us, when we do it for clients, I'm putting in discovery surveys, workshops, transcripts, um, revenue and stats. If you're doing it for yourself, you could still pull sales call transcripts, ah, PowerPoint presentations, anything like that. So it's not auditing in a vacuum. It knows the business goals that you're trying to accomplish when you go into it, um, the more information you give it, the better the output it's gonna get. It's gonna give you. It runs topic clusters very similar to how like Vista would run prompts, but instead of a percentage, it tells you just yes or no. Yes, you were cited. No, you were cited. Who was cited instead? Um, and then what you get back is the full audit. Plus what I like about it is it gives a prioritized action plan. Like it says, here are your top three fixes that you need to focus on. These are the critical ones. And then it gives it like a medium low priority for the rest of the list of the things all reasoned against, what it knows about your own business. Um, so like Peyton was saying, what's a perk of this one is that if you already have Claude or even ChatGPT, which can do very similar, you're not having to pay for a new tool. Um, and it's going to give you, um, a really good output. One of the, a couple of the notes I'd make about not necessarily pitfalls, but differences is that with Claude, the output is very, a lot more technical and dense. Visto, uh, uh, Semrush HubSpot. It's very pretty. It gives you these nice easy to read dashboards. Uh, so if you're wanting to hand your audit over to a stakeholder and you use Claude, you may need to translate some of it first and make it look a little bit nicer. Um, another thing that it doesn't do is that it doesn't give you any standardized, standardized scores. Like it doesn't give you a two point. Like it doesn't give you like oh, you have a 30% brand visibility or a 62% share of voice. So it's not the tool you'd want to use if you want those like really defined metrics to be able to track time over time. Three months, six months. Like if share of voice is something you want to be able to see is moving, uh, CLAUDE is not a tool that, that gives you that sort of score. It only just tells you where you cited or were you not? Um, yeah, and then kind of like I already mentioned again, this is with all AI, we already know this, it's only as good as the information that you give it. So the more you can provide to it, the more context it has and the better the audit's going to be.

Speaker A: Mhm. So that all sounds like it requires a decent amount of understanding in these Platforms and maybe some code. Say I am a very simple marketer and I would like to make the light version of an audit possible. Is could I use just prompts and ask basic questions and would it. Would I be able to still get something happening?

Speaker B: Absolutely, yeah. So there are light versions that don't do like this, what we call like the deep research function in these LLMs where you just use your more basic prompts. You can maybe provide your website, some competitor websites and ask it. Like I think someone mentioned that comment they've been doing competitor research kind of in a very similar way. It's going to give you those more easy, uh, yeah, those more like, uh, easy to fix overviews. Like here's what your competitors are doing. Well, here's the content that is being cited from them. So you can do a light version that requires a lot less, uh, you know, um, in advance, like building all of it out.

Speaker D: But I do think with that, like, I guess an added pitfall that I would say is like anytime that I've started, I love when I have a full day and I can just get into Claude and like just go uh, to town on like something like this. Like we're. But like it is a lot of back and forth and yes, claw does like work very quickly and it will respond to you quickly. Um, you know, sometimes, whatever, 10 minutes you have to wait for, like if you say, okay, I want deep research or whatever, but like I find myself spending hours, you know, like conversating with this LLM to get to a point where I'm like, yes, that really got me where I needed to go. So I think like, that could be a benefit because you're really directing it and like steering the ship in a way where if you use an out of the box tool, like what Maya was talking about, it's kind of, it's got steering like built in. Right. It's kind of telling you these are the buckets that are important. Whereas with something where you're just doing a prompt base, whether you're using uh, Gemini or Claude, whatever, and you're doing deep research, you are guiding it a lot more. But it is more labor intensive for you and you're gonna have to just continue to like follow the thread and think through it together. And I think that's where like Chelsea, you mentioned it's dense. Um, um. And so it's like there's a lot to suss out there. Um, um, could be a lot more value maybe, but also like, I think it depends on like how much time do you have to do this? Is this something you're trying to move fast on? Is this like m. Maybe just we try. We started free trials, um, with a lot of these tools and we did some demos and we said, okay, I think this one's going to work for us for now. And we're not married to anything. I'll tell you right now. I asked in the chat if anyone, um, has tools that they like. I'd love to hear more about it. We can have conversations about it in the collective too. But I, um, think. Yeah, there's just a lot.

Speaker F: Mhm.

Speaker D: There's a lot there, um, in terms of like, how valuable is your time at this point in time. Um, and uh, maybe take advantage of some free trials.

Speaker A: So we're, we think we've got enough time for one more audience question. Melissa, do you want to come on and ask real quick?

Speaker F: Yeah, yeah. This has been awesome. So thanks everyone for your experiences. You've shared with that. I think the question I just have is, um, is there like a template? Like I'm sure I can ask an LLM M, do you have a template to do a GEO audit? I'm sure I could ask that, but that's what I was curious, right, because so much of it is the input, what the output is all dependent on the questions we ask, right? If they're, if they're specific enough and that sort of thing. So I was just curious, is there a template or a guide that you guys have or use or suggest to even get us started with the prompts?

Speaker D: I know Chelsea mentioned, mentioned the skill. Um, we'll follow up on that and see if that's something we can share in the collective. And then, um, but I would say, you know, we gave a lot of like, I guess bullets of like, what you would want to include as far as resources or context for an LLM. So like, what do you know about your competitors? What do you know? Um, uh, and maybe Maya, you're a better person to answer this, but you know, your product category, like the context, what industries you're really, um, trying to break into. Um, things like that, like, give the tools the context of your business. Um, upload, like what Chelsea was saying, discovery surveys. Um, but I would say like, make your short list of app. This is how I would do it. Make your short list of assets, collect those things, write a very basic prompt to get started. And then like something like, I want to complete a GEO audit. And like, literally could be that. And then, um, you know, here are some resources that I'm going to give you about my company. Here's what I'm trying to learn, here's what I'm trying to improve. Spell those things out if you know them and then before you say run the audit, what I typically do with these LLMs is say what information like or what clarifications do you need from me? Ask me five questions one at a time and then I let it dig and prod at me and challenge me to give it more details that would help it and direct it. Um, and that's how I would start with like a, a prompt based one. But yeah. Maya, uh, um, any additional.

Speaker C: I feel like that's a really good start. We could even maybe I don't think one exists today like a prompt for like a chatgpt or something. Like Chelsea said, we have a skill for Claude, but if you're not on Claude, then we could probably extract a simple, you know, prompt to get started in a chatgpt and like Peyton said, upload all of the same inputs that all of these tools use because that's how they're using it to do their research. So you get a similar result. Like you're going to have to guide it more. You're going to have to. It's not going to be a very clean click a button and here's all your response, all your data. But you could you know, talk with it and um, get a lot of what you need probably to with ah, a prompt that we just kind of extract from that skill, the cloud skill.

Speaker D: All right Chelsea, anything you'd add as far as color there?

Speaker B: No, I don't think so. Yeah, I totally agree. Um, and I agree with Maya that I think we could probably extract some sort of like prompt template uh, based on all the information that we have. But yeah, telling the G, telling the LLM, I would like to you to run a geo audit. Here's the information. And Peyton, I loved your note about what clarification do you need like have it come back to you about what it's missing, uh, to just really fine tune that.

Speaker D: That's my favorite LLM hack when I'm working with these tools as like a thought partner, um, is I. And I hate it sometimes because it asks me really hard questions and I'm like oh man, you're making me work so hard, making me think.

Speaker A: But yeah, like it keeps asking what things cost,

Speaker C: pricing.

Speaker D: Yeah, get out of here. Well, this has been so fun. Thank you Chelsea. Thank you Maya for coming on the show.

Speaker B: Yeah, thanks for having us. This was fun.

Speaker C: This was fun.

Speaker A: Yeah. Thanks so much for joining us. It's been a great time. Um, and if folks want to keep the good time going and talk to you, uh, later on, what would be the best way to connect with you?

Speaker B: Oh, uh, yeah, I'm on LinkedIn. Um, I'm pretty active there and trying to be more active, you know, building the personal brand. And, um, I can throw my email, too, in the chat if anybody's interested.

Speaker A: All right, good deal. Maya, what about you?

Speaker E: Same.

Speaker C: Same for me. You can always reach me on LinkedIn. I'm always on LinkedIn. And, um, yeah, I'll drop my email as well, so feel free to reach out.

Speaker A: All right, good deal. Thank you so much. So our next IMC Live will be a little bit later on than usual. We're not going to be meeting again until July 17, so almost a month from now. But we're going to have a very good episode, so don't miss it. Keep it on your calendar. We're going to have Taz Bobber joining us to talk about landing pages. Um, she is the expert. Like, literally all she does is make landing pages for B2B companies. And she's going to be helping us unpack a challenge that many of us face. Trying to be able to use such a condensed format as the landing page when you're trying to sell really complicated products. Like, how can you do so much with so little? Uh, we're going to be really diving into that. So if you're interested, please tune in, folks in the room. We love seeing you. Uh, we hope that you'll be back for that. And for folks listening, thank you for tuning in. We'll talk to you soon. Thanks, everybody.

Speaker E: See you later.

Speaker D: Bye, folks.

Speaker B: Sam.