EP 292 - How to Get Cited in AI Search (Not Just Ranked)

Fab: hi, marketing rebels. Welcome back or welcome to our marketing school. I'm here. I'm my name is Fab.

I'm the head teacher at all marketing school and your marketing fair model. And hopefully, today we can give you a few things that you can do to make your marketing more impactful, more fun, but also more streamlined for yourself. Ben is going to be joining us today. I'm going to be talking about building a discoverable brand in the AI search era, but also in a way that is accessible for us and that it's easy for us to track to.

He's the chief content officer at Inbound Junction, a PR, SEO, and GEO agency. for Growth Stage Tech Companies and Beyond. Ben, I would love to hear more about you, but not as much about your work just yet. More about some truths and some lies.

Could you please, hello and welcome, let us know your three facts. Ben Jacobson: Sure, all right. Yeah. Well, one of ⁓ isn't.

all right, number one, Ben has helped to research blog posts for Rand Fishkin of Spark Toro. Number two, Ben studied mass communications for his undergraduate degree, making him one of the few people in his circles who actually work in the industry they studied. Fab: Mm-hmm. Ben Jacobson: And number three, Ben once led a team of community moderators for a series of MTV branded social networks.

Fab: I want to hope that you actually studied something completely unrelated, like a crazy weird and wonderful thing, and then that might have led you to where you are today regardless. That's my lie. You did not study mass communication for your undergrad, but we'll see which one it is. Ben Jacobson: The lie is number one.

I have helped some thought leaders to produce content before, but never ran Fishkin. Fab: This is a call out. This is the universe kind of manifesting for us. I love that.

We'll take that. I'm gonna follow up very quickly on the MTV branded social networks. ⁓ is there anything that you learned from that experience? Ben Jacobson: It's interesting question.

⁓ Not so much that has a direct impact on, you know, actual operationally what I'm doing. Maybe the importance of brand safety, ⁓ you know, getting communities involved, but in a way that you can maintain a little bit of control over it as well, finding that right balance. ⁓ But overall, I think the biggest things ⁓ I learned there is just the importance of. ⁓ relationships with different people you know, making different connections and ⁓ you never know what business opportunities can come out of those ⁓ over the years.

Fab: I think sometimes we forget actually of the bigger picture of all the other things that hopefully building said systems can help us too then going forward, like for example, actually nurturing our networks and focusing on community piece as well. Now, some of these numbers will probably shift again. And my shifted already is kind of interesting when you look at research, things go really, really fast. But since they are big numbers, we're looking at them as a range.

So there's actually a percentage ⁓ that is predicted, so that's also in itself a of an interesting one, to drop, right? As consumers shift to, for example, ChatGPD, perplexity in AI platforms, there's actually a percentage of organic search traffic that it might be predicted to drop. What do we think that falls under when it comes to the space of research that this comes from? Is it something like 25%?

Is it 50%? Is it closer to 75%? I just want you to think about, based also maybe on your knowledge as well, as a bit of guessing, how much do you think this is going to impact the organic search according to this prediction? Can have a little think, you can always let us know as well.

Go A and B. We've got a bit of a combination here as well. Live. ⁓ people are giving us some of these options.

I'm gonna unveil that for you right now. B actually. So we are in that middle ground of 50% would be predicted to drop, which I think is really interesting. It's ⁓ Gardner's prediction.

So obviously it's it's within a press release and there's other little media that also falls around that 50% drop, which is interesting. as consumers are shifting towards these kind of models, the shift of this mentality as well, and ⁓ I'm gonna say behavior of zero click content and zero click search, which I kind of find it interesting because ⁓ Ben talked about Spark Torah before, and that's something that Amanda Natividad mentioned and coined like this or click content as well.

I think this mentality is gonna definitely impact how we nurture audiences as well. And I was kind of wondering that leads into the visibility, right, that we wanna build by actually being present, being mentioned. And it comes obviously from what we're going to be talking about today, which is the PR itself. So before we even look at systems and things that we can do to make it more manageable for us, can you tell us if there's any mistakes that you see that marketers actually do when they think about introducing AR search visibility and tracking it and monitoring it for themselves?

Ben Jacobson: Yeah, for sure. before we get to that, I I do want to comment a little bit more on the ⁓ the zero click search thing, ⁓ and also these these trends, ⁓ you know, and how how people are learning ⁓ about different solutions. ⁓ basically I think it's important to know that ⁓ zero click search, this is a term ⁓ I think actually coined by Rand Fishkin once again. ⁓ It refers to what happens when someone ⁓ searches for something and then doesn't click through to any of the search results.

⁓ and it's something that's been tracked for several years now. ⁓ and there have been all kinds of studies about ⁓ the share of zero click search and how it's been rising. ⁓ and it it dates to well before ⁓ the rise of Chat GPT and these other AI services. ⁓ and basically Google has been ⁓ has has has had a strategy of keeping people more and more within the Google ecosystem.

And and that means they're surfacing more and more on the search results screens ⁓ in terms of other types of information, whether it's aggregating ⁓ video thumbnails, ⁓ social discussions, ⁓ information boxes of different kinds or answer boxes where even long before AI, you if you asked it a question, it would sometimes give you the answer right there. ⁓ So people clicking through from search results to ⁓ web pages, it's been shrinking for several years. ⁓ now with AI ⁓ it's been kind of catalyzed.

⁓ that fewer and fewer people ⁓ are are clicking through. an important thing to keep in mind also is that even as more and people more and more people adopt these AI tools, ⁓ it's not only people using Chat GPT, ⁓ Claude, whatever else, that's taking share away from ⁓ organic search. It's also Google itself. ⁓ so you know now Google search results, ⁓ they're pushing you towards AI mode.

⁓ more and more different types of searches are are ⁓ surfacing AI overviews as answers, which summarize the search results ⁓ and repackage that information. So ⁓ you know, e even the people who are using traditional Google are being funneled into ⁓ these these AI chatbot experiences. and I think the other important piece to know here is is ⁓ like Fab said that that it's changing how and where ⁓ the nurture process happens for marketing. ⁓ so you know it used to be that we as marketers would try to publish a lot of ⁓ keyword optimized ⁓ web page assets to attract ⁓ traditional Google referral traffic.

and even if it was for low intent keywords, top of the funnel keywords. And then okay, once they arrive on the website, then we can nurture them further by capturing leads, retargeting, building trust, whatever it is. ⁓ now that's totally changed. Now more and more of the journey is happening within ⁓ these AI conversations.

So someone might start with talking about a pain point they're having, or they might be a little bit more aware of ⁓ the landscape and be asking f if a certain solution is ⁓ a good solution, if a brand is trustworthy, how it compares with other brands in the space. ⁓ and then they're continuing the conversation ⁓ in this kind of self self-guided nurture process ⁓ with the AI giving answers ⁓ based on what it's seeing around the web. ⁓ and we'll get back to that in a minute. and and then once you've been shortlisted as a potential solution, that's when they might visit your website.

⁓ and usually the way that happens is. ⁓ they're gonna Google your brand name and then click through to your home page. So ⁓ I I always recommend ⁓ branded search organic search traffic ⁓ as kind of a proxy metric for how effective you are at ⁓ GEO, which is is what we call ⁓ AI search optimization. Fab: Do I need to get another tool?

Do I need to find something else? Do I need to make space or time and sometimes it's finding the minimal viable way that we can start bringing awareness to this, right? Instead of trying to almost run before we can walk. I just wanted to add that.

Ben Jacobson: Yeah. So, you know, getting back to your other question about ⁓ people's mistakes, there's so many. Maybe the biggest one I'm seeing is ⁓ when someone asks Chat GPT a question about their own brand, ⁓ and then based on the answers that they get, you know, they assume that that's what it the answers that everyone else gets and they want to optimize according to that. ⁓ when really what's happening is ⁓ these AI tools are are very good at learning ⁓ from who they're talking to and giving responses based on what they think you want to hear.

So if you're constantly working ⁓ as a marketer on your own brand ⁓ and asking questions about your own brand and others in your landscape, ⁓ you know for all that time, all those conversations, the LLM is learning about you and learning about ⁓ what you care about. And so ⁓ whatever questions you ask are are gonna be you're gonna you're gonna get highly personalized ⁓ results. So it's not necessarily representative of ⁓ what other people are gonna see. Fab: I love that and I think and I'm hoping as well that as we're gonna go through some of the steps as well that we can think about and some of the things that we can set up for ourselves, like going all the way from front research to understanding how to look at citation analysis as a practice and also really understanding some of these words, like for example, earn media asset is something we're gonna talk about, really understand what that means and how can we put some of this into practice, ⁓ I think is really important.

One of the things that Ben suggested we focused on, which is my favorite thing. gonna just gonna shout out to it is feedback loops i'm obsessed with them in other areas of the work that we suggest marketers do and then we teach as well so seeing them coming back here made me really really really happy but obviously it all starts with some of that first step which is thinking and understanding how to do prompt research and how to do it for ourselves. So what are some of your thoughts when it comes to this if we want obviously people to go away from this and start actually building some of these prompts and doing this research effectively?

What are some of the whys or the how of prompt research when it comes to this topic? Ben Jacobson: Yep. So prompt research is one of the most challenging parts ⁓ I find in in GEO campaigns. ⁓ basically ⁓ it's a huge blind spot.

⁓ unlike keyword research for SEO, we don't have reliable data on prompt volume. ⁓ it's also prompts involve a lot of kind of fuzzy logic. ⁓ you know, people might word the same question in 10 different ways. ⁓ and and the LLMs are good enough to understand the same intent behind it.

Then there's also another aspect of this, which is the fan outs. What's happening kind of ⁓ underneath the hood. When you ask a question to an LLM, you know, and it and it's figuring out what's what's the logic and the intent behind the question. So then it creates a bunch of fan out ⁓ prompts that are related.

And then based on that, it's querying its systems, it's querying the the open web to find ⁓ the best answers. So even if you ask a question in a in a specific way, the LLM is gonna kind of translate that into other prompts in other ways. ⁓ I I generally recommend starting, like I said earlier, with what are the pain points that your prospective customers have that you're the solution for. ⁓ and also what are the ⁓ you know looking at your your product category.

⁓ so you know let's say you're offering a ⁓ web application firewall, a WAF ⁓ solution for cybersecurity. So if someone is a little bit more solution aware, they know they need the best WAF. So ⁓ those are the prompts ⁓ you need to appear for if it's You know, what are the best WAFs for mid-sized companies in my vertical or whatever it is, look, you know, comparing the best ones, what are the best most trustworthy ones, you know, basically all the prompts around your product category.

⁓ and then based on what you're seeing using analytics tools, ⁓ you can decide if those look like vi viable prompts ⁓ to aim for. ⁓ with your publishing efforts. So that's that's gonna be also its own kind of challenge. You're looking at how you surface in the answers compared to how others surface in the answers, others in your space.

You might find that ⁓ a prompt is too broad and it's bringing in these huge companies that you can never compete with. Or you might find that it's It's too granular. It's not the kind of thing anyone would actually look for. even if it does seem highly competitive, you might see in the answers and in the citations within the answers, ⁓ okay, these aren't such precise answers.

So if we can publish the right content and land the right assets across the web, then we might be able to kind of teach the LLMs that these other players that are surfacing in the answers don't really belong there. ⁓ and that's something that I've done ⁓ for a couple of my clients lately that ⁓ you know the again with this kind of fuzzy logic ⁓ that that we can teach the LLMs to be a little more precise and to understand ⁓ the competitive landscape a little bit better. ⁓ so yeah, it starts with thinking about what are the prompts that do make sense that I want to be in the answers for, for ⁓ you know, high intent.

qualified ⁓ searches when people are looking for solutions ⁓ and then seeing what's in the results and seeing, okay, this is the kind of thing I feel like I have an opportunity ⁓ to appear in the answers more. Fab: There's one of these two steps that then feels like a natural progression of just making that decision and going to analytics and looking at what I like some of the signals that tell us what's what's relevant and what's not. However, I'm already thinking about some our rebels, and you know who you are.

You might be like, I love this, but How do I even come up with let's say an initial list of prompts that I feel makes sense to me? I feel some of us might more instinctively knowing what to do, but this could be a big block. It's just figuring out where to look for those prompts in the first place, who who to ask, or or where how to gather some of those. So is there one thing that they could do or one tool that they could look at, just one thing to get started if they're doing it in a smaller way and they're not looking to outsource just yet to be able to do it at scale.

Ben Jacobson: Yeah, I think if you know what your product category is, then you can generate prompts around, you know, what's the best solution for fill in your product category. ⁓ you know the pain points. These are all things that ⁓ hopefully with a good marketing strategy, ⁓ you you've already mapped out. So whatever resources you have there to think about, you know, what the the people who stand to benefit most, your personas ⁓ who stand to benefit most.

From your solution, what are you solving for them? What hurts them that you're, and then what are the questions around those pain points that those people are going to be asking? ⁓ and then other places you can look is in your SEO plans. ⁓ if you're doing ad campaigns, looking at those keywords for pay-per-click, and then just kind of formulating questions around those keywords.

⁓ those are also good starting points. Fab: Can you actually start mapping, knowing what Ben explained as well, how then that can inform everything else? Where are you at right now with this? So that you can figure out the next steps you want to take.

Are you looking at short-term goals right now, just more traffic, more authority? Are you actually looking at visibility that is longer term? So, as much as doing the things now, you also want this to become a system that can then start working for you. So, I want to make sure that you ask yourself, why does this matter to you right now?

And where you're at in that process. And I'm gonna and make this one a quick follow-up question for Ben. ⁓ but I was wondering whether you have one or two examples actual clients Ben Jacobson: Sure. ⁓ and to me it it comes down to measurement ⁓ and and getting aligned with my clients on ⁓ what are the prompts that we're that we're going after because some of the it might be more ⁓ realistic objectives and some of it might be ⁓ pie in the sky.

it can be a work in prog process also work in progress also to figure out what what are the best prompts to be targeting. ⁓ for me, I recommend looking at your share of voice in the answers. ⁓ and so you know, once you've closed on at least the prompts that you're gonna start ⁓ aiming for to ⁓ to appear in the answers. So ⁓ how often do you appear in the answers versus your competitors?

⁓ and then also in what sequence do you appear ⁓ in the answers? Because often I think maybe chat BT GPT more than the other LLMs, but A lot of them will rec if you ask for a recommendation of a solution, they'll give you a bunch. ⁓ so if you're towards the top of that list, that's also better. So looking at share a voice ⁓ and also share a voice growth.

Fab: Amazing, thank you so much. And I think yeah, that's just that starting point that I wanted to explore as well. It just said though, there's feedback clue, but before that, we got something else. We got obviously an understanding of the quality of the voices.

Verona was saying in the chat, obviously, like it matters what other people are saying about us. And I will add also who are the people that's saying the things that they're saying about us. That that was that was an interesting ruffle there. So, you know, examining and understanding who are the platforms of publications, who do we want to be.

you know tied to really matters but not just for us but also for our friendly little bots in there. So how would we explain the process and the concept and this step of citation analysis if we were just starting out with this? Ben Jacobson: Yeah, so you know, there are all kinds of AI answer tracking tools ⁓ that give different data points. ⁓ no matter what you're using, it should be able to package for you ⁓ what are the different ⁓ citations in the AI answers.

So ⁓ when someone asks, you know, when someone enters the prompts or similar prompts to the ones you're you've decided you're gonna be tracking, what's in the answer? ⁓ there's been there was a recent study, I think it was from HREFs, that pointed out that ⁓ citations have a negative correlation with recommendations, meaning that if I'm looking for ⁓ a certain solution in a certain category, the LLM will recommend me, but Separate from the sources. So basically, what does this mean?

It means that owned media is less important when it comes to influencing, ⁓ to getting your brand to appear as a a a recommendation. You're the and that means the content that you're publishing on your own website is not going to be as influential to getting you to be the recommended solution. What you need is to appear on these other platforms. If it's if it's trade publications, if it's influential business publications, ⁓ if it's if it's a ⁓ micro influencer's personal blog, whatever it is that's relevant to your space, ⁓ and and also different platforms.

It's very important to have consistent messaging on LinkedIn, on YouTube, ⁓ and Medium different places where you can actually publish also long form content. ⁓ and so looking at what the citations are, what are the sources that are informing the AI answers for the prompts that you're tracking? That will reveal a lot about where you want to appear in order to influence those answers. So you want to look at what are the domains, what are the different publications?

⁓ and even if that means you can't target the same publication, maybe there's a similar one that you that you can reach out to. ⁓ you know, is is Reddit a huge thing? Is it a review platform that's constantly being cited? so you know that that's something to look at also.

And then there's the formats. If it's a review article, if it's a listicle, listicles are huge. These, you know, the they speak to who are the leaders in your category. Those are the prompts that you're going after, then those are the the comparison pages that are going to really ⁓ influence the answers a lot.

So looking at what are what are the different web assets that are being cited most, that will show you ⁓ or or will at least set you in the right direction for what you should be targeting ⁓ with your earned media efforts. Fab: That is amazing. And I'm hoping that as you're hearing this, you might be thinking, ⁓ yeah, I've got a couple of publications already in mind, or I've got a couple of places that I can start with, even as has Ben said, to identify, if not them, then who actually reflects a bit of this as well.

And I love the idea that we took 12 like listicles as I understand why that would be really good for leadership as well. It makes a lot of sense as well. And so which ones are those for you? Which one would be these five outlets of publications that you'd love to be featured in, or that you would love to start exploring to see who should you actually kind of go for and how can you find that list that then matures from these five.

So I always like to start realistic so and reasonable. So five is a great start because as Ben said, you might want to expand from that based on Forbes is amazing, but maybe I should aim for something else as well. And the follow-up of that before we go into our next step would be very briefly. Is there any other format or you know I'm thinking about podcast guesting versus like short articles versus contributions or any type of activity that is a great one to start with when we're thinking about okay, I've got some math lists and publications, but I still am not sure what should I pitch or what should I look for as an angle or how should I position myself?

Because I think once again it's very different to ask for a full feature interview and trying to be spotlighted that way than it might be instead to contribute to something as well. What would be some of these examples that you think? Is it something you would recommend case by case, go for podcasts instead of listicles, or is it more of a case of like just starting small? What would that be?

Ben Jacobson: Yeah, it really it it depends on on where you are ⁓ in your in your journey with PR and with ⁓ AI search optimization. I think you know if if you have these connections, if you have a known brand, if you have research that people like ⁓ that w would be interesting. So I'd I'd I'd move forward based on that. If you're at the beginning of your journey, if you're an unknown brand, if you don't have these assets, then you might be best off looking at kind of thought leadership opportunities where you can publish ⁓ with minimal friction.

You can get involved in Medium. You can you can publish your own stuff on on LinkedIn and and those are platforms that are cited a lot in AI answers. ⁓ so yeah, if if you're just starting on your journey, see what you can do on your own. See what kind of co marketing partners might be low hanging fruit.

⁓ if you have you know, someone who's not a direct competitor but ⁓ you know has a similar target market that you might be able to partner with. ⁓ you can you can maybe publish on each other's blogs. ⁓ doing things like that could be ⁓ very impactful starting Fab: Yeah, I I agree with that. Is that low hanging through all the things that you might be already doing that you don't realize you can start partnering with people with as well.

And also I like the idea of just kind of getting a bit better understanding of what actually are the platforms that have higher authority or impact too, because sometimes you might expect it to be only the big ones, but it's not necessarily always that. So I love the little LinkedIn nugget and tibby that too. So now I wanted to ask about. what you mentioned with earned media assets and this whole idea of relationship building and everything along those lines.

This is a great step that I think sometimes we can skip when it comes to building powerful systems as well. And so I think if we're looking at you know publishing obviously media ⁓ earned media assets and building this and building this as a system and not just like a one-off. I was wondering what comes to mind when it comes to this specific stage. Because obviously relationships take time.

I think it's not a one-off. And also things like building pictures and all these things, they are a lot of moving parts that when you read it, you feel it's one thing, but it actually is not. So obviously I want you to elaborate a bit more on the concept, but I'm always bringing it back to the idea that how can we also make it easy if we were just gonna take that first step. Ben Jacobson: Yeah, so I think ⁓ when we talk about earned media, ⁓ people used to think in terms of ⁓ you know publications and and pitching journalists and landing stories.

⁓ I think in the context of GEO, the definition of earned media has gotten kind of blurred. ⁓ there are even paid channels that are influential ⁓ for the LLMs, ⁓ and also these kind of shared channels like these social channels. ⁓ so basically anything that's not owned media, ⁓ the LLMs are going to look at as sort of a fuzzy type of earned media. So like we've been saying, you know, doing more on YouTube, doing more on LinkedIn, on Medium, channels like that that you can fully control.

You can publish your own stuff, that that's a great starting point. But yeah, at the same time, Building these relationships as a stepping stone, starting with your own ecosystem of kind of co-marketing partners, ⁓ and then and then building up from there. ⁓ as you land more of these assets, you can start including them in your pitches to the media, ⁓ explaining why you're a worthy thought leader. ⁓ and then, you know, the next step might be op-eds and interviews.

⁓ in kind of niche trade publications and then you can build from there and get more mainstream as you go. Fab: I love that. I love that. And hopefully that made you think about some of the things that you can do to start off.

And I think that's me kind of reiterating maybe on this, maybe is the earned media pieces the one do you feel is one of the weakest links. Maybe you're actually thinking the weakest link right now for me is even just figuring out what my goal is. Like what do I actually want to achieve, even if it's just an understanding of the role of AI search within what I do, right? So it might be that the goal.

is the the sticking point. It might be that the one thing that you can do to make outreach easier is actually start writing down who do you want to approach and how do you want to go about that. So it might be step two. Obviously there's also step four which is feedback loops and that might be your weakest link, but I find that a lot of the time if we haven't even started, just a better understanding or a better kind of Overview and bird's eye view of what I want to do with this.

How do I want to approach AI search? How do I want to build it as a practice and a system? Obviously, that ties into that PR piece. It might be that it goes down to goals, but it might be that it comes down to ⁓ tools or just who do I want to talk to, who do I want to reach out to.

So that's the only other thing that I wanted to think about, and either jot it down right now, or you can at the end just answer this question. We wanted to give you a couple of things to think about so that you basically are it's easier for you to take action. But because it's such a big topic and the system that Ben is going through actually has got lots of moving parts, I thought it was only fair to leave this question and bring it in, which is what is the one thing that you can do just to make it one percent easier.

Just to get started with it or to get better at it. Which better might mean faster, easier ⁓ less involved, whatever that might be. As I said though, Ben, this might be actually tied in with the fourth step. It might be that feedback loops is something that you are missing on.

And I think it's important to always make space to build feedback loops and to make them better. Obviously if you don't even if you haven't even started out, you just want to understand the reason why you're doing it right. So feedback loops are my favorite thing. Why is it important to have a great system to track results.

We know why it's important to track results, right? But why do you think and then we're gonna go a bit more into what feedback loops are like, but why do you think we can actually make like tracking better? I think that's the thing. Like a loop is more than just tracking, is actually then iterating and making it better.

Is there a reason why this actually matters? Ben Jacobson: Yeah, I'm glad you asked that. ⁓ you know, you might have noticed a lot of times along the way, ⁓ in the other stages of this process, we've talked about ⁓ how blind we are in terms of actual metrics, what's actually happening ⁓ under the hood, what's actually happening when people are entering prompts, ⁓ what their intent is, what they're asking. ⁓ you know.

What are the different publications, platforms, assets that are going to be most impactful? There's a lot of guesswork involved, a lot of conjecture, and you're gonna have to run experiments. So that's why I think maybe more than in other disciplines, ⁓ with AI search visibility, the feedback loop is so important. You need to see as you you land these assets on earned media platforms, ⁓ are They actually causing ⁓ your share of voice to lift.

A these assets being cited in the answers? And then a week later, you go and you look in your AI visibility tool, and you look at the different citations that are being cited as sources by the LLMs for the prompts that you're tracking, and you see that article doesn't appear anywhere, then you know, okay, that you know, maybe there was some value in that, but it it didn't really move the needle when it comes to my AI visibility. ⁓ so it's really looking and seeing: are the assets that I'm landing being cited as sources?

By the LLMs? ⁓ and is my share of voice growing? And then based on that, you need to decide what to iterate next for the next loop around the process. Are you should you be changing what prompts you're targeting?

⁓ and should you be changing what type of assets you're trying to? Fab: It's almost like making space for asking the questions. And I think like a lot of the time we track results because we feel we have to address a specific ⁓ need or commitment as well. You know, you're working obviously with clients.

And I feel that almost that makes it feel rushed or either to really ask the right questions or just ask the questions so that we can give what people are looking for more than actually ask the questions to be able to prioritize what matters the most, which is not right or wrong, but I think it's really important what you said there about just that iteration and like because then there's gonna be another loop. And I think my only suggestion, I kind of there's a quick follow up there before we open open for QA's, is that A lot of the time we don't tackle what is most important, we tackle what's most or urgent, which is fine.

But I don't want us to miss out on some of the things that might not be quick fixer or might be something that we need to do straight away. Hello, I'm here. Lost me. ⁓ but actually stuff that we know longer term is gonna be beneficial.

And so I was wondering how do you find the balance? If you have to find the balance, how do you find the balance between looking at what is urgent but also looking at some of the things that maybe are results or data that is actually not as impactful today or tomorrow, but you know longer term is something that you want to keep an eye out. Like it's kind of like the results that give you what you need to do next, but also what you need to keep in mind. How do you find that balance both in acting on them but also communicating that with clients.

Ben Jacobson: Yeah, it's tough. ⁓ I think I think when it comes to AI search, ⁓ It it is a long-term game. ⁓ so you know, balancing short-term goals with long-term goals, it's less relevant ⁓ in this channel because basically everything is gonna be a long-term goal. It's gonna be ⁓ it's gonna be a marathon rather than a sprint.

On the other hand, it is easier to kind of validate that you're going in the right direction, I think, with GEO than with SEO. It it is a long-term game. ⁓ so you know, balancing short-term goals with long-term goals, it's less relevant ⁓ in this channel because basically everything is gonna be a long-term goal. It's gonna be ⁓ it's gonna be a marathon rather than a sprint.

On the other hand, it is easier to kind of validate that you're going in the right direction, I think, with GEO than with SEO. ⁓ with SEO. ⁓ It can take months to even see rankings change at all. With GEO, a lot of the LLMs have very strong recency bias.

A lot of them are going to ⁓ favor assets that were published more recently. And we've even seen situations where top citations, ⁓ one month, when you get to month three, month four, they're not in the citations at all. So Even though you might not see that asset as kind of an evergreen gift that keeps on giving, at least it shows you, okay, this is the type of thing I need to be doing, and just to keep going at it. ⁓ so it's all about kind of validating the the direction you're in.

Fab: thank you again for not only answering all our questions, but outlining a bit of a four-step journey that we can take too. Now, we have a opinion, a big question, but we'll see. is from Bruno. So is ⁓ you think the current geo still relies heavily on Ragdone on the SERP?

Do you already see many cases where the answers on AI overview are not just a summary of top SERP results? If possible, as you're answering this question, could you also help anybody who might feel this is a question that I don't even understand? to get a bit of context in what Bruno is asking too. Ben Jacobson: Yeah, it it's pretty technical what he's asking, but I think behind the question is kind of a broader question of to what extent is SEO with and ⁓ and ⁓ the answer a we're seeing, you know, especially ⁓ on the Google AI services, Also in Chat GPT and in others, there's a lot of studies that are showing that if you rank highly for a certain keyword, then you're likely to be mentioned and for AI prompts that are to that keyword.

Now it's not a total correlation and it's also diverging more and as time goes on. But for now, I am a big believer that good leads good GEO. and you can't really separate the two channels each other. Now going on with RAG basically it's it's these retrieval systems.

behind the scenes, if if you an LLM a question, it might decide to pull from its training data. Or based on the question, if especially if it sees it as a timely question, it's going to go to the web to look for answers. And so a lot of those considerations should what prompts you're going after and and how to go about influencing the answers to those prompts. Fab: Thank you so much for answering that and for giving us a bit of context as well on the question.

I hope it helps Bruno too. I have one final question myself, which is about obviously the people that are just starting out wanting to actually do some of this. They have an understanding of the technical concepts, they have an understanding of Buggio and SEO, but they're still a team of one, whether they are founders, they're entrepreneur, whether they are a marketer who has to do all themselves for themselves of others. What would be the best way for them to start?

It doesn't have to be even doing anything, it could be just thinking about something or just stepping back. But what is the best way for them to start before once again running before they can walk? And actually they feel like they are one step closer to building that system for themselves. Ben Jacobson: it's very hard to scale up on these things for people who are doing it on their own.

I think the best place to start is choose the best for you that will that will you track AI answers for certain prompts. Again, for the reason that I mentioned at the beginning that you know if you're only searching yourself, then ⁓ you're gonna see personalized answers. ⁓ you're not going to get a sense of what anonymous person is going to see in the answers. so using these tools, it'll also allow you to to track answers across different LLMs.

and and ⁓ and then on that, you can see what are what is what the gaps? What where does AI get brand wrong? and then think about what can publish fully in your control if it's on your own website. If it's on your social channels, if it's on some of these other of open platforms like Medium, what you publish that will help explain to the LLM what's more accurate about?

Fab: thank you so much. I feel it can be a bit of a beast just like SEO itself. I love that you reminded us that actually they are so interconnected. So the more we Understand of the one the more we understand of the other at the end of the day, it doesn't matter if it's geo-informed or SEO-informed anyway.

I am gonna do a quick follow-up from NES because NES just threw something in there, and I think actually it's really good. We talked about tools, right? And there was the mention of of platforms like Answer the Public, which actually we recommend to a lot of our students, is they're just starting out to understand what it actually means to look for queries. So is asking, what do you think about platforms like Ask the Public for Informing Geo Strategies?

That's a really good final follow-up. Ben Jacobson: Yeah, those platforms can be very good. I I look at it more for ideation they can set you in the right direction for looking, you know, formulating that might be relevant to your target audience, that might be relevant as prompts that you should be tracking and trying to optimize for. but again, it has to be an iterative ⁓ ⁓ it's not necessarily, you know, we don't have any visibility into actual prompt volume.

and and and the the logic behind these things is ⁓ fuzzy. There's all these fan outs. but yeah, if you have your your keywords that you're going after for SEO and you don't know how to turn them into at scale, that can be a very effective way to do it is to put it into one of these engines that that turns keywords into questions. questions.

Fab: Almost going back to their prompt search kind of line kind of way to kind of get their like qualifying or validate their ⁓ research a bit more. Ben, thank you so so much. Now, now now if people have more questions this one asking you more questions and or obviously they wanna connect with you, where should they go? Ben Jacobson: Yeah, I'm on LinkedIn.

⁓ you can search for me at Ben Jacobson or you've got my handle right there on the screen. ⁓ and my agency's website, inboundjunction.com, are good good places to go. ⁓ you know, I mentioned that the different LLMs will give different answers and citations are changing all the time and and the way it ⁓ the way the LLMs trust different resources are Changing all the time.

We have a data study coming out soon in a few weeks. We we partnered with some big organizations to kind of drill down into what the different ⁓ citations are that are most influential for different niches. ⁓ and I think that's an area where there isn't so much data available yet. ⁓ so if you connect with me now, ⁓ then hopefully that stuff will come up in your news feed soon.

⁓ you can learn a lot more of that. Fab: God, yes, please. And you know, make sure they actually, you know, get get the tea. I mean, we'll make sure that we do the same.

We'll also just ping you so that you can remind us when that is out. I'm always a lover of fresh data. I think it's something that we can never get enough of, especially with a lot of the stuff that we find on AI being all data, right? We just want the fresh stuff that comes from people and that knowledge.

So thank you, thank you, thank you. Thank you so much again.