Radarkit: How to Track AI Rankings + Automate Outreach!

Jay from RadarKit. How are you, sir? All good, Julian. How are you?

Yeah, I'm doing good, thanks. Tell us a bit about your tool and what you're working on. I know Jabez mentioned you. Yeah, so we are building RadarKit in a very competitive space, as you may be aware, that so many, you know, AI trackers coming right there.

So we've been into in SEO since last 10 years. We've been building a lot of expired domain projects. And recently we saw that, okay, you know, this AI source tracking is popping up. And we've been into automation.

So we've done a lot of automations inside our business. And we tried including our skills into this because none of the tools are actually giving you the real information. So what we did is we have our experience of browser automations. So we made our own browser.

So with RadarKit, what we stand is, you know, sitting in the UK, you can see what chargeability is responding to your prompt in Germany or in Japan or in India. So we are just solving that gap in terms of providing the real accurate data to the customer. And yeah, that is it. How do you track something like that?

How does that work? So what we have done is we've paired up with proxy companies. So you must be aware with MyPrivateProxies or WebShare. A lot of proxy companies are there.

So what we do is we buy proxies with them, from them. So if a customer wants to go into the market of Qatar or let's say Dubai, okay, so he can't just ask the same question, let's say, which is the best, you know, pizza company in Dubai, sitting in the UK, because, you know, if their ideal customer are actually in Dubai, so you want to optimize for that response. So you should be using a tool that is giving you the actual location based tracking. So that is what what we do at RedArchit, you know, buying proxies.

So we have partnerships with a lot of proxy companies, which gives us, you know, access to, you know, proxies from more than 50 countries as of now. So we use like mobile proxies, yeah. And do you track just across ChatGPT or you track like rankings across Cloud and everything? Across everything, across ChatGPT, Cloud, not Cloud, because Cloud is mostly B2B.

So ChatGPT, Persplexity, AI mode, AI overviews and Copilot. What do you see the differences between like someone ranking inside ChatGPT and someone ranking inside Cloud? Yeah, it is, it is very different because a lot of the times, you know, ChatGPT depends on the model to model because the models keep on changing. So in some models, the answers may be same, in some models, the answers may be very different.

So there is a concept called temperature. Are you aware of that? No, no, tell me about that. Yeah.

So the thing is, I'll just give you a basic example of how temperature works. So let's say you're standing at an intersection and you are point A and you want to go to point B. Okay. And there are six lanes that could lead you to point B.

Okay. But, you know, which way to decide is, you know, is the, let's say that's a ChatGPT. ChatGPT is at point intersection A, you ask the question to ChatGPT. Now, ChatGPT is gonna use something called temperature, which is if you're using API, you can actually set the temperature.

So temperature basically means the randomness, how many, you know, what, what randomness it is going to pick. So it is just, you know, it will, because it's an intelligence model, it is just going to do a rise, sorry, a roll of a dice. So if the dice says three, it will automatically choose an answer towards three, but it will lead to that answer. So if I ask, which is the best CRM software, be it the locations, be it, you know, be it Germany, be it UK, be it Dubai, it is gonna choose HubSpot as number one.

That's what we have seen, but the number two or the number three positions are always gonna, you know, change. Also the format of the answers is gonna change because again, it, it depends on the temperature. It depends on the randomness of these AI models. That is why, you know, it's, it's so different from SEO because the answers are always different.

Someday it's gonna answer me this thing. The next day the answer completely changes. Yeah. That makes sense.

And so like the temperature is changing all the time as well. Yeah. In chat GPT it is set by the system automatically. So it changes automatically, but if you're using the API and if you're building something of your own, you can actually set the temperature by yourself.

Yeah. So for, for people like going into cloud or chat GPT or whatever, it's like, okay, this is going to be random and it probably spikes up. Very random. Yeah.

So what you have to do is if you are optimizing for visibility, let's say if you want to optimize for more visibility in chat GPT, what you should do is you should look at an answer pool of, let's say, let's say I've seen your website about birds, right? So if I want to, if I, if someone wants to ask, which is the best website around birds, so why it will choose your website depends on, you'll, you'll, you'll, you'll have to look at a data point of, let's say at least a hundred responses over the course of seven days. So in seven days, what you should do is you should use a tracker to track. And then once you, once the tracking has been done for the last seven days, you come to your data, if you can build it by yourself as well.

So you come to the data, you see that, okay, these are the times when the answers get repeated. And these are the citations that get preferred more, you know, because let's say Reddit is coming and LinkedIn is coming or a normal blog is coming and your bird website is also coming. But what if, what if the occurrence of your website is only once there, there is the data, you know, that is the data that one should look at. You should look at a bigger sample size and then analyze based on that, because all the answers are always, you know, randomized.

Yeah. Yeah. That makes sense for sure. It's like harder to track than your typical sort of normal rankings.

Right. And then when it comes to actually ranking, do you see, like, how many data points is it actually taken from? Like, for example, is it looking, do you think at the first two or three pages or the first 10 pages, for example, for Google, how many sources is it pulling in? It depends from model to model.

What I have seen is maximum 10 to 12 things back 10 to 12 in AI mode. GPT sometimes goes overboard with, you know, with its research. So it's all token based. Right.

So what happens is when you saw something, if it's in their training memory. So what is the cutoff training memory if it's on, let's say, the last time it's November 2025 or something like that. So before that, whatever data they have, they'll give it from the data. But if you ask something very particular, let's say, if you ask, like, I need a CRM software because I'm a law firm.

So that is a very particular question to charge GPT. So here it will do a web search. So web search, again, it is using the SERP API or else, I guess, Google's API. So obviously, whatever's ranking Google, it will fetch that.

But again, even if it's even if it is visiting a website or a source, it doesn't confirm that you'll be there in the answer. The answer is very much dependent on, you know, those, you know, what say if it's coming from a discussion like, let's say, Reddit or something, they choose those answers way too faster. Or else if it's coming from a very high authority website, where an author in the similar niche has written that, they tend to choose that. So sometimes your source will be there, but you will not be chosen.

So that is one more, you know, caveat into this. I mean, so like, if you're asking it a question from when it's got the data point from already, like, let's say, for example, the cutoff knowledge for GPT 5.6 is like November, just as an example, like, does that mean that it's harder to rank unless you're ranking previously? For that answer? Yeah, if you're tracking for 5.6 all now, and you were ranking in 5.5, then you'll have to wait for the next training data.

So what you'll have to do is, let's say, you were ranking before, but now you're not ranking. So that means they have updated the, you know, the citations, their memory about your brand, about your niche. So what you need to do is take more action, based on the sources that it's taking, it's learning from, it's getting influenced from. So what you can do is, if it's getting influenced from Reddit, a lot more, or it's getting influenced from a third party website, let's say, for Zoho, for Zoho's case, Techradar is one of the most influential source.

So what I'll do is, I'll kind of, you know, build a partnership with them, or, you know, with one of their authors, ask them if they can, either, you know, post something like that, you know, how backlinks work, you can ask them if they can just publish a guest post for us or sponsored post for us. Because that is, because when the next, what to say, the cutoff happens, when the, you know, memory cutoff happens, you're already there in the memory for those, for your terms. This is how we take it. But what we suggest is go after a data point called web searches performed or not.

So for your terms, let's say you run your prompt 400 times, and then you take a percentage on how many times a web search was performed. So if it's more than 70%, so that means, you know, every time it's performing a web search, so you have to be on those web search results. So automatically, you'd be in their minds. But if it's not performing a, you know, web search, if it's like, less than 50%, that means it's getting the data from the last cutoff.

So that means, you know, it will be harder for you to influence because it is, you know, it is not getting, it won't be influenced because it has already made up the decision on your topic or your question. Yeah, that makes sense. So if it's doing a web search more often than not, then basically it's going to be changing the answers and you've still got a chance. But if it's using no web search or not using a web search often, then it's probably going back on the previous set of training data.

And so like, it's kind of like a core algorithm update where you just have to wait six months for the next one, right? This is so new for us because we are waiting for their updates for sure. They're going to be an update. So I mean, they'll be publishing updates from now and then.

And they should give us a web search console as well. Something like that, where we can see that our brand is performing as well. Because Bing gave us. Bing obviously no one uses Bing, so we need Chattopadhyay or else at least Google to give us the data so that we can look at data.

Yeah, that's it. I mean, have you looked at the new features for AI inside Google Search Console? Yeah, I have. I have, you know, seen the update, but it's not pushed yet to everyone.

I guess it's only been pushed to, you know, the UK users and it's very, very few people have gotten it. Yeah, it's weird. Like it just showed up on my Google Search Console today for like one of my websites. Yeah.

One more thing. So if we, can I, can I just, do you have another question? You can carry on. Okay.

So, so the thing that we were discussing, let's say if we are not getting into the minds of all the 5.6, so what we can do is we can go after the query fan outs. So those are, you know, this is a new thing again, I guess Jabez was also discussing that with you. So query fan outs as well, there's a lot of folks around, a lot of people saying a lot of things about query fan outs. So the heavy, there is a person called DJ and he has also published a website.

So a lot of people, what they're doing is they're just taking their keyword. They're just asking Chattipati, you know, that give me the fan out. So obviously that is not going to work. So what fan, what a real fan out is that when we search something, can I show my screen?

Of course. Yeah. Yeah. Go for it.

Perfect. I made a video showing, you know, how query fan out works and I have done that in real time. So this was my video. I went into Chattipati and I asked for this term called best CRM software, as you can see over here.

And then I went to the network tab and then I can just pull out the conversations module. And here you can see that Chattipati is actually, when I search best CRM software, Chattipati is actually searching for best CRM software 2026, small business enterprise comparisons. So, you know, it's giving me so many keywords right here, here only. So what we do as Redarkit is we just, you know, utilize each and everything automatically from those browsers because we are not using any APS or because we are not using data for SEO or any API out there to fetch those results.

We are actually, you know, built on real browsers as well as, you know, we are using proxies to fetch those results. So, you know, you can see what a real user is getting the answer in US or Germany. This is what we do. So what we do is we have just extracted all the information.

We give it to you over here in the query fanouts tab. So if someone's searching for best CRM software in the US, you'll get this, you know, these are the query fanouts. And within the query fanouts, there will be some fanouts which get seen more. So we have to optimize for those fanouts.

So this is what no one talks about, that even inside the query fanouts, there are fanouts which are, which have higher occurrence. So, you know, if let's say we have hundreds of query fanouts, so you want just, you know, you have to decide, you know, what to say, impact level things, which are the highest priority ones. So we give you those keywords because we fetch those things within the models, within the browser itself. So here for a prompt like best cloud-based CRM software for a small business, these are the prompts that the LLM is actually searching.

So it is searching a US, you know, phrase just after because we are searching in the US as well as you'll see a lot of, you know, HubSpot name, there's Zoho, PipeDive. So you get a lot of keywords automatically on which you can just automatically optimize. That's smart. Yeah.

I've never seen that before. Like I obviously heard about query fanouts, but actually that was, even that was like something I only heard about recently, but the fact that you can go into the network tab and then see the actual code that's being used and what it's searching for, you know, this reminds me of something I was going to ask you about. So like when I was checking my search console today, I was trying to find like keywords we're getting impressions, but not clicks for. And like some of the keywords were stuff I would never imagine someone searching for.

It almost looks like an AI is searching for it, but it's, but it's within search console. It's not like the AI part, you know, have you seen anything on that? No, it's like only impressions. It's almost like the AI is like, I feel like it is an AI search and like, they're just harvesting who's trying to rank for that keyword, you know, it could be you.

That's a new insight. Yeah. Because we don't have that. I am just waiting for someone to share it with me so that I can look at it.

I have seen the Bing one, Bing one looks really nice, but here, yeah, I mean, this will be very informational. Let me see if I can pull it off here. Let me show you what it looks like. Sure.

Are you visiting any conferences? Yeah, not so much this year. Yeah. Yeah.

Not got that many plans for it, but it's something I'll probably get back into next year. Are you doing so many? No. What we are doing is this time we are going to the Brighton SEO in San Diego.

Yeah. So we are exhibiting over there. Yeah. We are sponsoring as well as we are exhibiting in the Brighton SEO.

So there is one Brighton in Brighton SEO happens in Brighton as well, right? In the UK. Yes. Yeah.

Yeah. Yeah. Yeah. So this is a details here.

So it's like generative AI features and then you can see like your impressions here. Okay. Have you come across this before? No, I haven't yet.

Just waiting for any of my website to show them. Yeah, I literally saw it this morning and it's pages that I wouldn't expect to get many views for. Is it showing for which prompt have people search this or any term or any keyword or something like that? It's really weird.

Like, so it doesn't show you any of that data. So you can see generative AI features in beta, total impressions, and then you can add a filter, but the filters like country, device or days, like the least useful data that you'd want for something like this. What would you do with this data now? Exactly.

Exactly. The only thing that's maybe slightly useful is like looking at the URL and looking at like it's growth over time, but even that's kind of like, nah. That was just impressions. I guess.

So these are being cited or being shown inside the LLM hands. So I guess, I guess shown inside the LLM hands. Yeah. How much traffic did they send us from this?

They should give that data as well to us. Yeah. It should be like, okay, what's the prompt, how many clicks you get in. And then the, cause I, even the impressions, you can have a website that's getting like 10,000 impressions with no clicks, right.

If it's just ranking for the wrong stuff. Yeah. I'll just show you something more, I'll just share my screen. So I'll just show you like, even we got the data, right.

We saw the data in, you know, the generative engine insights. So what we did is we built something like page radar inside Redarket. So let's say for this, this is a company that, you know, is, is from a friend of mine and we are doing a case study as well inside Redarket on this one. So what we did is, what we do is the, we've added this project since two, three months now.

So once you add a project in Redarket, what you have to do is just, you know, let it add your prompts and then just wait for, you know, citation data. So a lot of people ask like, what should be my prompts? So here, what we have done is we built something like agent. So what you can do is you can just connect, just like cloud and all, you can just connect your GA, you know, Google search console, and you can just ask our agent, can you just suggest prompts for me?

So here, our agent, what it will do is it will just read all the impressions that you're getting on your thing, and it will automatically add, suggest prompts for you, and you can add it inside Redarket. Once you have added the prompts, you come to the citation data and here, you'll learn that which kind of content is actually influencing the, you know, the results. So here, you'll see that a product page or a tool page is actually influencing the result 34% of the times, followed by a social platform like Instagram, you know, YouTube or LinkedIn also influences the results 24% of the times, followed by a listicle form of article, which you may have heard, you know, a lot of the times because listicles are working so much. So listicles improve, you know, this thing, you know, around 20% of the time, then differs a lot from brand to brand.

So, you know, because Zois was an AI video generator, you saw their citation data, here you'll see Zoho, which is an enterprise software, which is a competitor to HubSpot, you'll see that for them, a listicle form of content influences the result 43% of the times. And you'll see how different it is, you know, for a B2C software like Zois. And then once you have this data that, okay, these are the data points that are influencing the results the most, then let's say, if I want to just create a listicle, I will just post that, you know, I'll just, I'll just come here, I'll just see that, okay, YouTube is being the number one source. And then agent, which will obviously is a competitor, say they're not going to give us a link.

And then you'll just, you know, browse and browse, and you'll find any website which can actually, you know, give me a link, you know, so you can just outreach them. And how do we do that? You could just, we have an agent for this as well. So you can just click, you know, in the connectors, you connect your own Gmail, you go to the outreach agent, you can ask, you know, can you do some outreach for me?

So what it will do is it will just set up outreach campaign for you, it will ask, do you want to buy backlinks, or you want to just get mentioned, we'll just do buy backlinks, because we want to just, you know, get out there as far as possible, you'll just select for which prompt do you want to buy the backlink for. And let's say we'll just create AI twin generator, we'll just click continue. And then you'll see that, let's say, there are 100, these domains get cited the most, you can just select that. And it will just ask, do you want to do follow ups as well.

So these things will be done via your own email, it's asking. So I'll just do, you know, let's say two follow ups, if the author is not replying, we'll just do more follow ups as well every two days. And then if the author sends me a price of $500, we want to negotiate as well. So it will do the negotiations as well on your behalf.

So let's say I said 10% over here. So let's say it says continue. And then it is asking me, do you want to sign off with something, I'll just do skip it. And then this is what we built before what used to happen is everyone used to use Redakit and then there was no sheets, I mean, there was the action was to taken by them.

But now what we built our agent, the agent, you know, takes the action. your behalf. So now the agent has found these domain, they get cited the most. So like captions.ai 70 times, HN65, there's AIjoin.

So for sure, this is a third party domain. It gets cited around 19 times. Now I've just run the workflow. It is automatically finding the emails of all these domains.

It will automatically find the email. It will email them. It will wait for the reply. If they don't reply, it will follow up.

It will ask if they reply with the pricing. It will do a negotiation as well. It will do everything and it will initiate the first point of contact between you and the author, which is getting cited more. If you want to do the third party citations, if you don't want to do that, if you want to publish content, you can utilize the agent.

It will automatically publish content as well on your website based on the query fan outs, based on everything. And then once you have that, once we have found the emails, it is sending the emails, what you can do is you can come to brand radar. So this is one specific feature, which was similar to what you showed. So what we can do is we'll just add, let's say this is actually a third party URL that was purchased by Zois.

So this is, we actually paid them to get this article, best AI talking photo generator. So once you have that published, so I'm just, I'm just showing you the workflow. Once you have that published, how do we know that this is bringing us sales or bringing us traffic? So you can just come to brand page radar.

You can add that page. Once you have added that page, you can just look in the last 30 days and see that how many times this specific page brought traffic to me. And so it's just loading. You'll see that AI John has got, being cited around 22 times and you can see for which specific prompts is getting cited as well.

So you'll see that it is getting cited for two different prompts. So that means it's working for us. Even we published a LinkedIn post. So this LinkedIn post gets cited around six times.

It's also getting cited into two different prompts. So you complete the whole loop of creating, finding out which form of content is being cited, finding the website, using the agent, you'd set up the outreach using the agent. And then once that is done, you'll just, you know, you acquire the links as well and you track which links are performing the best for you. That's great.

Can you create content as well? Yeah, you can. So what, what we do is we create content based on, have you used Surfer SEO or you've used Neuron? So we create content based on that.

So there is a lot of, you know, should I show, should I show it over here? Yeah, go ahead. Okay. So what we, what we suggest everyone to first of all, foremost, look at the query fan out.

So let's say if I want to rank for best AI twin generator, so in the query fan outs, let's say we'll pick up, pick up, let's say this one, we'll, we'll pick up this query fan out term. We'll click on create content. We'll just click here. So what we do is we extract because in Neuron or Surfer, they used to extract the terms from the, you know, SEO side, right?

I mean, so, but in, in these LLMs, what we do is we actually extract all the terms from the prompts itself. So your will, what we'll do is we'll just utilize this prompt and we'll just click on, right. And now it is just, it's going to ask me, do you want to do a listicle or a blog post or a comparison page? So let's do a listicle because those work the best.

So your, I'll just mention that mentioned, so I, as number one. So what we have developed is something called brand kit. So you might have heard about, you know, controlling the narrative inside these LLM models. So if you want to be, you know, because these LLM models, they learn about you through YouTube, through Reddit, through LinkedIn, through your blog posts, through third-party citations.

So they are, they want to learn about your brand. What is your, what does your brand stand for? So you have to speak that thing everywhere. It's like search everywhere optimization.

So that is why we built brand kit. So in brand kit, you have to, you have to add each and every information about your brand. What is your positioning? Who do you sell to?

What is your target customer? What is your product and services? What is your key differentiators? You know, do you have any awards or do you have any approved facts about yourself?

So you can just add everything inside this brand kit. So these things will be utilized while creating content so that we control the narrative about your brand. So here, once you add the prompt, it will just, you know, fetch the SERPs. These are the SERPs that are ranking for your, for this term.

And these are the citations that you get writing. So what you can do is you can select, so these are baseline articles. So what we, we ask customers to just select some articles, which are highly optimized from the SERP as well as from the citation. So you stand, you know, in two, in two of the boats, you're always optimizing for the LLMs as well as for the search results.

Once that is done, it will automatically just go out there and create an outline based on a lot of factors, a lot of data points that we have collected inside Redakit that can be query fan out, that can be, you know, the citation things as well as NLP, the site, the, the, the, the terms as well, the semantic terms. So each and everything is being extracted based on the baseline articles that we selected from the SERPs as well as from the citation data that you have on Redakit. And then it creates a very highly optimized content, you know, which is optimized for both. It's just, it's just gonna take maximum 30 seconds.

All good, man. All good. One thing I've seen recently is like the amount of investment going into AI ranking tools as well. It's pretty insane, right?

It is, it is. It's too much. I mean, you'll see profound, profound is like more than 1 billion now. Prompt watch, they recently raised 6 million.

Peak has already raised more than 30 millions. So it is going, I mean, it is going bonkers up. The thing is these people, they, they, they're not coming from the first way of SEO. So if, if, if someone who, who comes from the first way of SEO, then it, you know, the tool will be better.

It is what I believe. So here you'll see that, you know, these are the, do you want, were you saying something? No, no, you carry on. So this is the outline that it came with and every outline, it will tell you why we have picked this outline.

So we have picked this outline based on a metric called GEO. So we introduced this metric because we found that a lot of the LLM models, they're choosing articles based on the content that's present. And if, if an article is citing something which has facts, which has some sort of data, you know, something like, you know, market market gap or any percentage of growth or something like that. If you're adding some search form of data, the LLM are likely to choose your articles.

So that is why we try to add those things as well inside your article. So let's say this H2 was selected because it was find, found in a, in you, you must be aware about PA. So it is found in PA, it was also found in the query fan out as well as it is, it is also found in the entities. So we are creating all these H2s and H3 based on the query fan outs, based on the entities and all the details so that we are optimized, not even for the main keyword, we are optimizing for all the query fan outs, all the entities, all the NLP terms, PA terms, everything, you know, you name it, you know about the SEO writing thing.

So we are trying to, you know, club everything. And so then we can create a very highly optimized article, you know, for, for the brand. And here you'll see, it has generated this article, which is, you know, which comes up with an, with an image as well. You'll see that, you know, it has kept Zoys here as number one, then agent and synthesia.

It has mentioned each and everything. It has made mentioned frequently asked questions as well. And it does some, you know, I would say outgoing links as well, but these are not do follow. So they are all no follows.

And yeah, you can always just, just like surf for a neuron, you can see which all terms were used, which were, which all entities were used and you can automatically optimize it. You can see that, you know, which facts were utilized. So this is the geo term that I said about that, that I said to you, you can also generate schemas. Even though Google said that they are discontinuing with the schemas, you can add schemas as well.

You can add facts. It has included these three fan outs based on the six fan outs that we have for this prompt. We have included these PA questions as well, each and everything, you know, we have added that so that the article becomes more and more, you know, optimized. You can change this paragraph as well, if you want to, and they can automatically publish it to your WordPress, so it just completes the whole circle, you know, our tool as well as a lot of things we've added.

I mean, this is, I mean, they've been building it since one year. And since we are doing it ourselves, we are using the listicle module. We had everything. I mean, I just show you one example.

So there's this company use hall. Have you heard about use hall? No, no. So they just, you know, went bankrupt.

So, I mean, you were just talking about how much money these people are raising. So now there is a lot of tools who are losing money and going bankrupt. So I found two companies as of now, which were funded and they've gotten bankrupt. So use hall is one of that as well as use bear.

So use bear or something like that. Yeah. So they also went out of business. They were backed by YC.

And then obviously, because I guess because they are building on top of API and API costs a lot. I mean, if you're just, you know, going after thousands of prompts, it's going to cost you a lot. And that's why I guess they also ran out of business. So use all, I guess, I don't know why they went out of business.

And so we wrote an article, you know, best use all alternative on my LinkedIn. Okay. And we got a very big enterprise client just because, you know, he was just searching for that. And you'll see that Redocket is the number one alternative mentioned inside AI overviews.

And you'll see that how it has picked that information. So what I was just talking about you, you know, how you have to do search everywhere optimization. And this is the best example I'll give you. So this is the LinkedIn post is my LinkedIn post.

Okay. And this is one YouTube video that is our channel's YouTube video. You'll see this is this is a short video, which I made it with the AI channel. So I have a big channel with 51,000 subscribers, a lot of videos on LLMs, you know, everything we published like every day.

And this video also got cited. So obviously, you know, this data also helped. us get into the rankings, a lot of, as well as, you know, this Reddit post, this is also ours. So we, what we, this is my subreddit.

So we have over here as well, you know, I have done for this keyword, what we did is we took the 360 approach. So the whole search, the whole search for this keyword is ours. The videos, the LinkedIn posts, the Reddit posts, the blogs, each and everything is ours. So that is why we weren't able to, you know, get these things.

And we got the idea just because using, you know, tools like Redarkit, we're finding the keywords, finding which alternatives these LLM models are searching inside their heads, using query fan out technique. And then we are creating content on that. So this is one of the best examples I would, I could give to someone if they want to optimize, you know, for any keyword. That's crazy.

Smart. Actually, I love the fact that you got your own YouTube channel there as well, GEO with Adam, right? Yeah. Yeah.

So what do we do is we use Agent and Zoys as well to generate videos. But that's getting good views, man, like for SEO videos, particularly. Yeah. We don't get, sometimes we get, yeah.

That's awesome. That's awesome. Any, anyone you see like really winning with AISEO, like, wow, these guys are the gold standard. I guess, I mean, the older brands, I mean, I'll say Zoho, Zoho as well as HubSpot is doing very well.

What I've seen is HubSpot has, they're doing amazingly well. Even they reach out to us. So I'll just show you this thing. Yeah.

I mean, if you search AAU trackers, so HubSpot, this page ranks and they are not this one concept. They have this page on their blog, blog, blog here. This one, I guess. Yeah.

So what they did is they are buying a lot of backlinks. So even they reached out to us asking that, can you, you know, mention us in one of your blog posts? I was like, yeah, for sure. I mean, I'm getting a DR90 blog links for why not, I'll just take it.

So even they know that we are competitors, but they did mention us and they took two links from my blog, which I don't understand, but I'm okay with it. I mean, so they are doing it on a very nice level. They are just outreaching each and everyone, whoever is ranking for their prompts, whoever is ranking for that, you know, keywords in the LLM models and they are just outreaching them, asking them for mentions. So this was the blog.

I'll just show you that. Should I show you the blog as well? Yeah. Yeah.

Go ahead. Yeah. So this is the blog and then like 25 search visibility. So they, they asked us if they want to be mentioned over here.

So they got mentioned over here by us and they mentioned us in the, in their pages. So they have been doing it very nicely. This is one brand. I would, I would see that they are just flying inside these LLM models.

I mean, they're just doing it so nicely. They have so many pseudo channels, YouTube channels, LinkedIn posts. They also have this organic thing as well. So they have started this new project called HubSpots AEO or something like that.

And so that is where they're shining. They're doing very well. Their target is just B2B. So not in front of us, obviously, but they're doing it.

Nice. Awesome, man. Well, thanks so much for coming on today. That's great.

It was, it was awesome, Julian. Awesome. We could, we could do it again. I mean, I'll, I'll come up with one more study.

I'm working on this case study. So this one involves using PR articles. So what I have seen is if you just do a normal PR blast with AB Newswire or something, it actually changes the answer, but it doesn't last longer. So we are just testing out two, three PR results and we'll see that for which niche it lasts longer and for which niche is just, you know, just be great.

Let's do it. Let's, let's do another follow up soon based on the case study. Yeah. Once I'm done with this case study, I'll surely ping you and then we'll do it again.

Perfect. All right. And I'll link to RadarKit in the description as well. If anyone wants to check it out, I'm actually going to send this over to my team so they can watch this too.

Good training. Perfect. Thank you, Julian. Thank you so much.

It was great talking to you. Thank you for inviting me. Thank you. Yeah.

Bye bye. Bye bye.