GEO vs. SEO: What Cybersecurity Marketers Need To Know With Clinton Karr
Learn how GEO and SEO work together for cybersecurity marketing, from AI search visibility and E-E-A-T to original research, governance, and leads.
Learn how GEO and SEO work together for cybersecurity marketing, from AI search visibility and E-E-A-T to original research, governance, and leads.
Founder and Principal Consultant at Ramsey Communications
Clinton Karr is the Founder and Principal Consultant at Ramsey Communications, a strategic communications consultancy specializing in cybersecurity and emerging technology. He advises clients on corporate communications, media and analyst relations, integrated content marketing, and executive thought leadership. Clinton is also the Founder and CMO of ProGEO.ai, where he applies 20 years of communications experience to generative engine optimization, AI search visibility, and original research on how brands are discovered and cited by large language models.
AI search is changing how cybersecurity buyers find information, compare solutions, and form opinions about brands before they ever visit a website. Traditional SEO still plays a major role, but GEO adds a new layer for marketers thinking about visibility in tools like ChatGPT, Claude, and Google AI. So what should cybersecurity marketers actually do differently as search continues to evolve?
Cybersecurity communications and GEO expert Clinton Karr argues that the smartest approach is not choosing GEO over SEO, but understanding how the two work together. Strong E-E-A-T signals, useful structure, original research, credible statistics, and expert commentary can all improve the chances that content is found, trusted, and referenced across search and AI platforms. Clinton also highlights the importance of writing for natural-language queries, keeping humans involved in AI-assisted content, and focusing on business outcomes rather than AI visibility alone. The goal is simple: create authoritative content that helps buyers and gives search engines and LLMs something worth surfacing.
In this episode of Cyber Funnel, Thorn Compton talks with Clinton Karr, Founder and Principal Consultant at Ramsey Communications, about GEO versus SEO and AI search for cybersecurity marketers. Clinton covers LLM visibility, changing buyer behavior, and content strategies for generative search. He also explores AI governance, original research, and why human oversight still matters.
This episode is brought to you by Amplifyed.
Amplifyed is a specialized SEO agency, and we partner with in-house marketing teams to boost search visibility, drive qualified traffic, and turn organic search into real pipeline growth.
From technical SEO to content strategy, Amplifyed handles the heavy lifting so your team can stay focused on building and selling. If you want to show up when your buyers are searching, Amplifyed helps you get there.
To learn more, visit us at amplifyed.io.
Powered by Rise25 Podcast Production Company
Intro: 00:02
Welcome to the Cyber Funnel podcast, brought to you by Amplifyed. Where we feature the top minds in cyber security.
Thorn Compton: 00:13
Welcome to the Cyber Funnel podcast, where we're lifting up the people behind the brands across the cyber security industry. I'm your host, Thorn Compton, and some of my great past chats include amazing people like PR extraordinaire Renatta Siewert from Black Bear, from Backblaze. She's going to hate me that I totally mispronounced that. I'm sorry Renatta. Today I have the pleasure to welcome Clinton Karr from Ramsey Communications into the cyber funnel.
How are you today, Clinton?
Clinton Karr: 00:38
I'm great. Thanks for having me on. Thorn.
Thorn Compton: 00:40
Absolutely. I'm very excited about this conversation. We're going to get real nerdy. It's going to be awesome. Before we get started, of course, this episode is brought to you by Amplifyed, the SEO and GEO Agency for cybersecurity companies.
Amplifyed serves cybersecurity marketing teams by driving more leads from Google and AI search. We handle all the heavy lifting and know the exact levers to pull to get you more skills. But most importantly, we speak the language of cybersecurity. Learn more at amplifyed.io. That's amplifyed.io.
As I said, I've got Clinton Karr joining me today, and we're going to get into the nitty gritty about GEO. We're going to try not to scare anybody off today. All right.
Clinton Karr: 01:26
Yeah. I mean, I know it's one of these things that a lot of people are interested in it. Their CEOs are asking, how do I get, you know, how do we get our brands mentioned in AI? How do we get visibility into AI? So I bet you're going to have a pretty strong appetite for this subject.
Thorn Compton: 01:44
Always, always a strong appetite. I also have a strong appetite for getting to know you a little bit more. So before we get into it, I have one question for you. And then we're going to ask another fun question to get to know you at the end. What did you want to be when you grew up? Clinton.
Clinton Karr: 01:58
Well, yeah, I know we're both gonna nerd out today, so I'll admit, in high school, one of my favorite extracurriculars was model United Nations. And so I went to Boston University because they had a good international relations program. And unfortunately, I'm just not very good at speaking foreign languages. And it also turns out that international relations is a lot more of writing policy papers and analyzing policy papers than it is meeting with people and, and having that personal connection. So I shifted into public relations.
And I remember asking, you know, older classmates, friends of mine, like, what's a job where you just get to talk to people all day long? And he said, oh, that's PR. So, you know, here we are now. And, and it's working out. I'm getting to talk to you today.
Thorn Compton: 02:58
It's so funny. I'm a I was a journalist in my past life before getting into SEO and cybersecurity, and I had a very similar thing happen where like, I'm a sports guy, but I was like, I'm going to be a journalist and I'm going to be like a real hard hitting, like cop journalist and that kind of stuff. And then I did like a cop beat for a little while and I went, this is so boring. This is so much more boring than I ever thought it was going to be. And then I pivoted back into sports.
And that was and that was a great time. And hey, God bless the broken roads. And here we are. We're, we're both cybersecurity marketing and, and, and.
Clinton Karr: 03:27
Yeah, there's enough true crime podcasts now as it is, right?
Thorn Compton: 03:31
Hey, don't throw me out with a good time. We can, we can film another episode.
Clinton Karr: 03:33
You know, I got into cybersecurity and, and tech PR, I was interested in the more traditional broadcast TV or radio or that sort of thing. And, and people sort of said partly because I lived in Boston and there's a strong tech corridor there. You know, if you can make it in tech PR, if you can make that interesting, then you can make anything interesting. And now I've just been in tech for 20 years, so it just is what it is.
Thorn Compton: 04:04
That, you know, that pivots right into my first, you know, real question here and talk about your journey kind of into cyber security. I know you've been around in the industry for, for, I don't want to say quite a while because, you know, we're, we're not aging anybody on here, but like you've, you've been through a lot of.
Clinton Karr: 04:20
I'll age myself because I like to call it the Reader's Digest version. But now I realize I don't even know if Reader's Digest exists.
Thorn Compton: 04:28
It's gotta be around. I've seen them at the grocery store. Yeah.
Clinton Karr: 04:32
It was on my grandma's coffee table when I was a kid. But yeah, I, I started off in Boston working for PR agencies that were generally high tech healthcare and that sort of thing, and did kind of specialize in the high tech side of things. And they transferred me out to San Francisco, which was always a goal of mine to live on the East Coast and the West Coast and really kind of just fell backwards into cybersecurity, where they kept adding me into more and more cybersecurity accounts. But this was at a time almost 20 years ago that RSAC, I remember because it's in San Francisco, people in 2008 or so were like, is RSAC going to survive? That's kind of how low attendance was.
And now you look at it and it fills up both sides of Moscone. You know, you can't get a hotel room for less than 500, $800 a night. It's just there's 1700 cyber security companies at least. And it's just, you know, so I definitely got in at the ground floor at the right time.
Thorn Compton: 05:44
And now you're dealing with what everyone else is dealing with. And that's these pesky LMS and, and how to how to rank for them. Yeah.
Clinton Karr: 05:52
And I will mention I, I worked both in-house. I worked with Zscaler early on and Bromium Neustar, which got acquired by TransUnion. And now I've been consulting for about the past ten years almost, in fact.
Thorn Compton: 06:09
Fabulous, I love it. Yeah. So yeah, you really have seen quite a bit of growth. We're going to talk mostly about GEO in this. And so just for people that have either been living under a rock or have heard the terminology, but never, you know, wanted to ask exactly what it means.
What exactly is GEO?
Clinton Karr: 06:26
Sure. Well, GEO is generative engine optimization and I'm on team GEO also. I like that we're calling. Yeah, some people are calling it, you know, an E for answer engine optimization. To me.
The thing is that you're not always asking a question. It's not always producing an answer, right? It's generating content. And so that's what it is that generative engine optimization is, you know, how do you sort of optimize your content for the retrieval and recommendation of your brand in large language models like ChatGPT, Claude or, you know, even Google's AI answer overviews? Yeah.
But also, I'm not precious, if you want to call it, you know, a GEO or AI search optimization. It's, it's, and frankly, I think I know that Amplifyed feels about the same way. At the end of the day, it's largely SEO.
Thorn Compton: 07:27
Yeah, yeah. The first I do draw the line at a I e. Oh no, no, you, you get three. You don't get four. And yeah, you know, it is a little bit how we though there.
I do believe there are fundamental differences between SEO and GEO. What are some differences that you've kind of seen throughout your research?
Clinton Karr: 07:49
Like you and I have talked about before? Right? At the end of the day, it's all about sort of optimizing your eating signals, right? The experience, expertise, authority and trust. Right.
And that's both, you know, on your own page and the links that you generate across the web. I do think that there are particulars that you see. There's this fundamental study that we can dive more into, which is literally called generative engine optimization out of Princeton. And they found that in particular, things like statistics and quotes provide a lift in GEO that I think you can't necessarily say has that same sort of lift in SEO. And likewise that sort of research Muck Rack has produced shows that press releases have more pick up if they have quotes and statistics, and that I think those are kind of two big ones.
But, you know, it's more the commonalities, right? If you want the structured format, the H2, H3 headers, you want to sort of write for the natural language search intents and answers the questions that users are likely to ask there. There's more technical stuff, right, that we can start to get into. Like in particular this idea of the LMS dot txt. Yeah.
And I think that that's where there is. A lot of debate in the industry, right. And, and it's even the sort of thing where Google's left hand doesn't know what its right hand is doing, where, you know, John Mueller is saying that you don't need LMS dot txt. It doesn't help anything. And then they have lms dot txt on dev.google.com.
Right? And I think that part of it too is that there's kind of like levels or layers to what we are talking about when we talk about an LM, right? And so some of these optimizations are maybe more for coding assistant LMS than they are for the chat bot.
Thorn Compton: 10:22
LM you, you just touched on first you answered a question that I had about how they're similar. So how dare you get ahead of me on my question. No, you actually, you really just ran into a very interesting thing to me about LMS. People just kind of use the terminology AI or LMS, and they're thinking that they're talking about one thing. They're not. I mean, even when you're talking about the differences between perplexity and Claude and ChatGPT and Google's AI.
Overview. We're not talking about one monolith. It is different elements that are going to be affecting different things.
Clinton Karr: 11:00
And even if you are using an API connection versus, you know, the web portal or desktop agent, you're going to get different results.
Thorn Compton: 11:11
Yeah, it's, how difficult does that make even trying to like improve your visibility or even measuring when, when so many different factors can affect it.
Clinton Karr: 11:24
I mean, it really, to me, it's, it, it reminds me a bit of Twitter, right? Of like how the early days of Twitter was just this fire hose of how do you make sense of any of this? And I think that trying to make sense of AI visibility is a bit the same way, especially as tokens are getting more and more expensive. Oh, yes. You know, it's the cost of trying to gain any visibility into.
Into what? And it's almost a vanity metric too. Yeah. Right. Like, I think that there are these best practices that organizations should be aware of and should be following.
But at the end of the day, it's really about your lead gen and your website traffic and how many people are converting on your offer. And that's what really matters to the executive team and the board at the end of the day. Right. So they might be asking how we get visibility in AI? But what they really want to know is, you know, how can we keep driving sales results?
Because I think one thing that companies do know is that they're getting less search referrals, right? And that is because of AI. And that is because, you know, the sort of zero click phenomenon, right, which is this propensity for users to just stay within Claude or within ChatGPT and, and just have their conversations that way, right? So yeah, it is changing the whole dynamic around the buyer's journey and. More so, and I, I think that $0.06 has some really good research around the way that AI has changed the buyer's journey.
And it's really just sort of changing the timelines. Like the buyer's journey still looks the same. It's just that people are spending more time conducting their research on their own. And when they're coming to your website, they're already pretty well informed now, right? So the point is to still produce that educational and awareness building content and, and still produce that demand generating content.
But just to be aware that that content is, perhaps , getting consumed and synthesized and, and spit back out through the large language models.
Thorn Compton: 14:09
I like to say a lot of the time that it used to be that the first touch on your website was very rarely going to be like a conversion. Now it's to the point where the first touch is still a little early, but the first touch with your brand is very rarely going to be with your website. You know, even going back five years, it was the first touch when someone searched you up or something similar. They find you on Google, they get a very small amount of information through the meta in the title, and they have to go into your site and go figure you out. Now they're getting so much more information with you that the first touch with your brand really is going to be coming from AI.
And that really is how I think it's really important to, to kind of own your message between your owned media, your earned media and your shared media to kind of have that surround sound approach to your content, you know?
Clinton Karr: 14:56
Yeah. And there is, of course, the issue of sort of hallucinations, right? And. I think interestingly, I've seen that there is this tendency or whether it's intentional or not, I, I'm guessing it's intentional or it could be laziness on, on sort of other content creators parts. But I will see Google AI overview or, you know, Claude research mode, maybe quote a statistic that it pulled from another blog or from a LinkedIn post, right?
And that statistic is on that page. But then if you go to that page and try to find, okay, well, what's the source of their statistics? They just made it up, right? And I mean, it's the sort of thing where I see people on LinkedIn essentially inventing statistics and attributing them to Gartner, which, you know, this is Gartner's world and, and we're just playing in it that are then getting cited in, in research that I'm conducting. And it's like, I think it's important that as I kind of shift gears and talk about AI maturity, that when we are using AI as marketers in our workflow, and I do every day, and it makes me very efficient that we're not just like offloading all of our cognitive processing to these models, right?
You got to treat it like an intern that is very good at doing research, but will still sometimes bring back something really questionable. And it's when you get too comfortable that you let your guard down. And that's when that, you know, false statistic slips through. But guess what? It's your name on the line.
It's your brand on the line. And in the short term still, it might not even hurt your eating signals, right? But I think that we're still pretty early days and the hallucinations are always going to be a problem. But I think that the sort of evaluation of these citations and the links and all of that, I mean, Google and Microsoft and so forth, but they've been doing this for 20 plus, you know, years and they're only getting better at it, right? They're also experts at AI and machine learning.
And all of that is very good at pattern recognition. So, you know, you can see some of it where even in the various updates, Google kind of has provided some advance warning about, you know, AI scaled content abuse and that sort of thing. You don't necessarily see it playing out in the search results immediately, although I've seen some stuff from the start.
Thorn Compton: 18:05
To.
Clinton Karr: 18:05
Catch up, folks like Lily Ray that have analyzed some of like the listicle spam or, you know, comparison stuff that a year ago was like rocket fuel and, and is now like, you know, space X two weeks after launch.
Thorn Compton: 18:28
It didn't, it didn't quite break the break the atmosphere, but it is quite there. Yeah. And actually, you know, that's a great little transition into let's talk about bots For again, real quick breakdown for the layman specifically. All right. So like you've probably heard of Google's bots, if you are, you know, any anywhere in on SEO, you understand robots.txt and the importance of having your site be crawlable.
Where does that come into play with LLMs dot txt and how is a GEO bot different from a Google bot, and is it at any point?
Clinton Karr: 19:08
Well, it is definitely different in the sense that there's, you know, different bots, right? There's both the sort of training bot where the LLM itself is trained on the entire corpus of the internet and there's, you know, a sort of hard cutoff date to that which prior to retrieval, augmented generation, you know, about 18 months ago or so, you would see that, oh, you would get a response. I don't know the answer to that because it was prior to my training data. Right. So that's sort of the training bots are more similar to a spider that's crawling and indexing a site.
But then there's also the user bot, which is when you are using one of these models and, and you ask it something that it does need to go search the web, it's going to use that user bot and in particular, consequently, at the end of 2025, ChatGPT has said that their user bot no longer respects robots.txt because they see it as a proxy for the user. Because if you and I go to a website, we don't go to the robots.txt and see, oh, what, you know, pathways am I allowed to look at and what's disallowed? Right. And so that does make sense to me. But kind of back to the big picture, if you look at and this is tangential to robots.txt.
Cloudflare has a dashboard with so much interesting information. And one of the things that they track is the crawl to referral ratio, right? And you can see that like Google search will index your site ten times for every one visitor that it sends to your site. And it is like an entire, it's multiple orders of magnitude different for ChatGPT is something like a thousand times and Claude is something like 5000 times, right? So these bots are coming to your website, but they're not sending traffic to your website.
So that is, you know, kind of one big difference in that regard. But in terms of how they behave aside, aside from the ChatGPT user, you know, they should sort of theoretically follow robots.txt. And I think, what does that mean for you and me and the companies that we work with? Well, really not a whole lot, right. Because ideally you have your robots.txt set up like a wildcard allow or, you know, a few specific disallows around content management systems or that sort of thing.
You don't need to go out there and say, allow ChatGPT, allow this, that, and the other. On the other hand, there are companies like the New York Times that have sued ChatGPT and, and even about a week ago, Cloudflare has announced more news around their content independence program, right where there is. A class of content creators that don't necessarily want a large language model to come and consume their content. I don't think that applies to the companies that we're working with or, you know, cybersecurity brands want their company to be mentioned in these models. But if you are concerned about, well, let's say like a, there's like a Decipher type of thing, right?
Where it's a media site that's owned by a cybersecurity company, and maybe they don't want their content getting consumed by a large language model, then. Yeah, then you might want to say, block these, these models from training on my data and, and serving it up without sending any traffic to me.
Thorn Compton: 23:32
Yeah, I, I tell people all the time, one of the best things you can do for LL or for LMS is make yourself plagiarize. Make yourself very easily plagiarized. And if you don't want to be plagiarized, then yeah, you probably should at the very least, block it. But right now you want to be because you have to think about it from the user standpoint too, of the different user. You have a user for the LM who is looking for an answer versus the user for Google who is oftentimes looking for some options or they're looking for, hey, give me the things.
So the way that the bot wants to serve that user is just going to be fundamentally different. You know what, this has all been great. We got real nerdy on that. So we're going to take a hard pivot. We're not even going to talk about this.
The next question is going to be another fun one. What was your first concert?
Clinton Karr: 24:21
Yeah. My I mean that yeah. Big hard pivot, like clear the context window and just start a new conversation. My first concert was with Aerosmith. I don't remember exactly when or how old I was, but yeah, that was like the big arena tour.
And I also started off my career working at the Rock and Roll Hall of Fame. And so I got to go to some really cool events that way. I sat backstage with the Pixies, and I got to go to a seminar presented by Grandmaster Flash where he was both sharing stories about the evolution of, you know, they call it turntablism, right? He was like the first DJ to mix records together and, and gave a lot of sort of hands on demonstration. That was really cool to see.
Thorn Compton: 25:22
I've always found it fun that like the I'm a big Johnny Cash fan and I, and I loved that one of the early inductees to the Rock and Roll Hall of Fame was him, because a lot of people don't necessarily say it's like he, they, they, they don't have a, a huge binary. This is rock and roll for the Rock and Roll Hall of Fame. And so therefore.
Clinton Karr: 25:40
Yeah, yeah. I mean, this year they're inducting the Wu-Tang clan.
Thorn Compton: 25:43
I, which, oh, like that was a big Wu-Tang. They did a Tribe Called Quest a couple of years ago. Like. And the fact that those people, because it's really more just a history of American American music kind of thing. Love that.
Fantastic. All right. We're going to hard pivot way back because we're toward the end here. Clinton, this has been fantastic. Before I give you your final question, talk to the people.
Tell them about where they can find you. Tell them about anything that you have coming up or any fun things you have going on.
Clinton Karr: 26:10
Sure. I am the one and only Clinton Karr, which is great for being an entity in search engines and large language models. You know, it's funny though, because I want to touch on one more thing as we talk about sort of what works for large language models is, like I said, you know, statistics and so forth. I produced some research which people can find on an associated site that I launched called ProGEO.ai. And I literally, as I was studying generative engine optimization, kind of wondered to myself, oh, what percentage of the fortune 500 have implemented LMS dot txt right now?
There's a whole conversation we could have about whether or not LMS dot txt is all that effective at what it claims to do. But I was just interested in it as a sort of adoption signal, like how much of the fortune 500 is interested in this as a subject? And when I searched for that, there were no results for it. And so I actually vibe coded myself a website scanner and had to go through all the iterations of it serving up false positives and, you know, had my own AI psychosis moments and all of that. But, you know, it took a few weeks, but I got a really good data set and analyzed it and all that.
And now if you ask Google, what percentage of the fortune 500 have LLMs.txt, you're going to get my research. You're going to get the press release that I published. You're going to get the Reddit post that I made. Right. And so it's that sort of thing.
And that's why I think in cybersecurity in particular, it's such a great industry because of all of the threat research and, and statistics and all of that, that there's opportunities for cybersecurity companies to just constantly create original content.
Thorn Compton: 28:21
Yes. Oh, 100% and that OG that new content. I mean, that's the stuff that LLMs are feeding on. They have to have that to continue to grow. So by creating good quality content on that, that's a great example of a good thing that you can do to kind of take advantage of this AI era.
My last question for you is, what are some of the biggest pitfalls marketing teams and cybersecurity are seeing when implementing and adopting AI into their processes?
Clinton Karr: 28:50
I think again, if you take a look at my website, I did some research around AI maturity within the cybersecurity space in particular. And it was really just intended more as, as initial benchmarking, especially because I conducted this back in February or March, I guess was RSAC. And it shows sort of the obvious result being people using it to create content and repurpose content and stuff like that. I was surprised to see that not as many people are using it for role playing, right? Where you can stress test your messages, say against adopting the persona of a CISO or that sort of thing, or even, you know, prepare for a sales call or customer meeting or that sort of thing.
Right? But in terms of the biggest pitfall I think is, it is not something that I measured. It was something that came up in conversations that I had at RSAC . Let's say that you and I are both working on the same marketing team, or maybe I'm a contractor working with the vice president of marketing, and we're all using Claude, and we're all using Claude projects. But each of our Claude projects has their own context, right? And they're probably in a similar sort of context, but there might be one conversation that the VP had.
That doesn't make it over to my model or vice versa. Right. And so I think that. Probably the biggest pitfall is around usage and governance to begin with. But sort of that next level of maturity is then how do we make sure that we're operating with a consistent brand voice, a consistent sort of context?
Thorn Compton: 30:49
Absolutely. Clinton. This was incredible. I already have you written down for. We're going to be launching a new podcast.
That's, that's the Cyber Funnel After Dark where we have cocktails and talk about Google conspiracy theories. So we're going to have you on that. I appreciate you coming on, man. Today, you have a wonderful rest of your day and we'll talk to you soon.
Clinton Karr: 31:11
Great. Thanks, Thorn.
Outro: 31:12
Thank you for listening to the Cyber Funnel podcast. Tune in next time for more tips and insights, and be sure to click subscribe for future episode updates.
