How to make Content that ChatGPT and AI Overviews prefers - Backed by 6 months of Data Studies

Okay, can everyone hear me? How's everyone doing? Sorry for the late start to the webinar. Can I get some hello in the chat? I think there were about 80 of you guys waiting to come in. I apologize for the wait. Thank you for joining today's webinar. Yep. see how things are going on. How's everything going for you guys? Thank you for joining wherever you guys are joining from. I think uh today's presentation will be quite interesting for a lot of you who may or may not have been following our blogs and our content with the number of different studies that we've been pushing out. We're going to focus just on a few key points today of uh different studies that we've pushed out in the last 6 months or so. Uh some of them have been focused on uh observing what kind of uh sources get cited, some of it about the responses themselves, others to do with traffic going to different uh AI assistants and so on. So, let's just quickly share my screen here before I get started. I will put myself in the corner while I share. Okay. So without further ado uh let's begin today's webinar and hopefully if things work out I can also help to share you guys with you guys a way in which you can uh use the HFS MCP server to uh do some of the data analysis yourselves as you guys um many of you guys would be subscribers to hrefs. If you guys have not uh known, there is actually a way to uh use our API without having an API subscription and that is through the API server. Oh, sorry, the MCP server. Uh and if you have a subscription to a popular uh AI assistant, a supported AI assistant like Claude, you can connect different MCP servers to create um do automation, create the analytics that you need to uh make decisions. Okay. So, but let's go ahead and do some uh housekeeping at the beginning here. Uh so what we're going to do and welcome to uh today's uh webinar on how to make content that chat GBT and AI overviews prefers backed by six months of data studies. So I'm going to go through a number of different uh uh key facts that we found from analyzing from like a few thousand uh data points to hundreds of thousands data points depending on the studies that we've done where we try to understand uh what sort of content is getting picked by AI and from there we try to infer what are the things that we should be doing to our content to make sure that our brand or uh our our brand gets mentioned but also in a context that we we find that will be beneficial to us as uh brand owners or website owners. Okay. So what I'll do is I'll share the main presentation in the first half and after that we will open the floor for Q&A and if you have a question please share your question in the Zoom's built-in Q&A function. So there is a Q&A function um if you click on more next to your Zoom if this is your first time using Zoom there is a Q&A part and if you ask questions in chat I may not see it or may not get to it. Uh but if you ask them in Q&A function then uh we will go through those accordingly and if you see a question that you uh would definitely like to see answered first please uh please upvote that question so that gets u bumped to the top of the list. Okay. Um, this webinar will also get livereamed on YouTube. So, if you uh want to watch back uh part of the webinar that you've missed or you'd like to see again, uh you can do it later on um in on YouTube and the link to that will be sent to you after this webinar is over. Okay. So, without any further ado, let us begin. So today we're going to um cover sort of five data um findings. So the question of first is what kind of content length does AI prefer? Do they have a preference for a specific content length? This is a important question to ask because of mainly because of what people are claiming are important like kinds of lengths of how many thousands of words that their blogs now need to be now that AI is crawling as well as uh people are visiting the websites uh visiting the websites. The second is how much does content freshness affect AI visibility? uh in general we know that or you may have heard that it's important to keep content fresh but um we can also look deeper into the data about like what does this what does that mean and how often we should be looking to refresh our content uh the third is how much does YouTube mentions affect influence AI visibility and the fourth is how easily does best blogs uh best x blog list get cited in AI answers and finally is how much of cited content is within marketers control. Okay, so let's uh let's go ahead. First is I just want to quickly answer question why do I want to focus on chat GPT and AI overviews today when there's so many other AI uh assistants or AI uh platforms out there and the first is that uh AI overviews has a very direct impact to how much traffic you get on your website and that means two things. The first is that people are not going to your website anymore. They're getting an answer. They're getting the answers from what AI overviews generates as a response for like from as simple queries to even very complicated informationational queries. And um that means that if they are getting their answers um from AI overviews, they will stop they will go ahead with what is being re recommended there. Um, and when we looked at this study, by the way, you see that there's analysis 300,000 keywords. We looked at keywords that have 150 keywords that had AI overview show up and another 150 keywords that had no AI overview present to do a proper comparison. We did this uh last year and then no like the year before and then we made an updated study in December 2025. So like we're seeing that the trend is that people are um whenever AI overviews is present, people are clicking even less than before. And so it is important to recognize that what it it is basically a sign that people are really satisfied with the answer that AI overview is giving and they're just going to walk away um with that uh answer generated answer in mind. Uh next uh when it comes to um traffic so like chat GBT in of itself recently there has been some wavering as to how much um it's being used and there's like reports about whether or not it's being out competed by other platforms but there's um basically no contest in terms of how much traffic is driving because of all the AI assistance that we're able to look at chat GPD drives u more than 80% % of all AI AI related traffic. So it's basic it is the AI system that is inspiring the most people to click um like to go to a particular website after looking at an answer that is generated. So that's important to to recognize since the answers there is basically um able to in that way convert into uh potentially a deeper step in the salesunnel or however um your your website is motivated to drive clicks to your website. Um I will note however that overall the the traffic coming from AI is still like almost 300 times less than what Google um currently provides and drives this traffic uh to websites today. Okay. Uh and this data comes from a website called uh chatgpt versus google.com and this is a vibecoded website that ATFs have put out where we um update this data on a monthly basis. And you can basically see the trend of how um AI is stacking up against Google in the fight um for dominance is currently going on. Okay. So now let's go to the data and findings. So the first question content length essentially has zero correlation with AI citations. So the uh the correlation metric that we use here is Spearman correlation. Uh the average word count if we just take the average number is actually,200 but like if you just take that one number it's quite misleading. So, we've like distributed them into buckets here. And you can see that more than half of all citations uh in AI go to pages under a thousand words. And a thousand words is considered like short for the average blog post or website content. And it's well below what most SEO professionals would uh request in like a content brief. Uh a lot of our blogs are definitely longer than this content length. And it's definitely a lot shorter than what um a number of these uh mega guides and guruids are suggesting that oh they should be anywhere between like 3 to 5,000 or even in the 10,000 word mark. Um so like looking at the data that's not what we see and there also does not so like um maybe if the overall length is not long does a longer length therefore result in higher ranks? We don't see that happening either. In the top 10 citation uh citation positions um there is no meaningful difference. It's um that the length is just as similar to the same middle buckets of within 350 to um 2,000 words. And the so the Spearman correlation we looked at in terms of the word count and citation position is a score of 0.04 which is basically zero. One is the number that we want if we say is extremely highly correlated. Okay. So like the correlation we see here is 0.04. So what does that mean in terms of how we should um do our content or or like create our content? We should trim the fluff. We should not focus specifically on the like word count number like based on a particular industry that the priority is to be able to ask answer questions well and straight to the point. And in fact, um, structuring your content clearly and and focusing on clarity rather than length helps to not only for the reader to understand your content and and digest it better, but it also helps structuring content well as well as keeping it uh like within a reasonable short amount of length helps the AI also seem in this case to be able to improve your chances for getting cited. So it should it like the content length should not be like a goal target when you create your content. Uh next uh URLs cited by AI assistants are 25% fresher compared to URLs in SERs. So this is across 17 million citations across like multiple AI systems. But the point here is still true especially for in this case we're looking for a chat GBT. we compare it to like normal SERs. Um AI assistants just love um to site um fresher content more and in fact chat GPD shows the strongest preference for newer content uh citing URLs that are like um 300 to 400 days newer than organic uh Google results. Okay. And uh we will see that in in more detail when we talk about um the best sum best x uh blog lists that were being discussed uh earlier. Okay. And next uh but first we'll talk about YouTube. So YouTube mentions are the single strongest predictor in AI visibility. We looked at a number of different uh uh AI visibility factors here like things like branded web mentions or DR or number of referring domains like these are the uh traditional uh SEO factors that we said um that you that used to be quite important but when it comes to AI visibility um YouTube impressions uh and um mentioned impressions and YouTube mentions are at the top and and this is important to to recognize because it uh Most people may not or most um businesses or website owners may not uh necessarily even consider uh YouTube as a strategy. But that doesn't mean you have to create the YouTube content yourself. You can look into collaborating with uh influencers or people that have a reason to mention your brand. You can look into doing sponsorships. You can also look into um the ways of which uh your brand or competitor's brand is even mentioned in the first place. And that gives you an idea of what uh content can uh how you can get involved actively to getting your brand mentioned more uh more organically or even just like more blatantly. Okay. Um and that's just that's not just in citations but also it's important to note that uh that YouTube's uh video content does get um part of the training. It gets ingested as part of the training data. uh for both Google and OpenAI they have trained their models on YouTube trans uh transcripts in fact the New York Times reports that OpenAI's GPT4 model which is some time ago was trained on over a million hours of YouTube transcriptions so it's it's like a part of that is the reason why a lot of these big brands that get mentioned a lot in YouTube videos are so baked into uh the answers that are generated by AI. this is the part of that giant data set that gets um helps to uh generate the answers uh to people's questions. So, so even if you are not um in necessarily uh into creating videos yourself, please aggressively look into or not aggressively but like really really consider what uh being in involved or being uh present in YouTube videos or even like Tik Tok videos as well uh can do to help improve your uh brand's AI visibility uh in in both chat GBT in AI overviews and eventually also the other AI AI platforms. Uh and a good example would be to talk about our own brand because like even though um we ourselves did not uh originally start our YouTube channel to rank in AI search because we've been consistently publishing on YouTube since uh 2018 uh like 2018 our keyword research tutorial for example is one of the tops sources for how to do keyword research when we uh when people um search for that and using AI mode. So, um that's just proof that like consistency matters and like really exploring a space uh that maybe either like you've tried before, you can consider um consider creating video content again or if not work into partners that are already established in the space and see if there's ways to uh work together there. Okay. Um this is also true for new brands. So um like a lot of the big brands get like mentioned a lot. So um the impression is that AI makes it very difficult for uh newer brands to get started and established recently since they're fighting against all this like existing content that the AI has been trained on. But YouTube seems to be um very useful in getting a small brand get mentioned more and get uh uh yeah mentioned more in AI answers. Okay, now we are going into best x blogs uh represented in uh so so like how often do they get represented in all of the different page types? So, one second here. They make up like a significant uh part of uh chat GPT citations up to 43.8%. And 35% of these lists come from a low authority domain. So, it's not necessary that you have to um go to the the the biggest DR companies that maybe charge a lot of money or um they make it very difficult for you to um get mentioned or get a new page posted. So, placing higher on these higher on these lists also increase your chance of being recommended. So it's not just um like being like showing up on these lists but also um seeing where you're positioned in these lists to get recommended more. You can see that in proportion to the other uh page types like um the next following is things like blog posts and landing pages and product pages like blog lists that um consists of like oh best uh fridge for a college student or like um best um places to to go for vacationing in in summer 2026. These are the kind of like blog lists that we're referring to. Um but you can see here in my takeaway it says to avoid spamming and uh we we I'm just going to quickly mention the reason why uh this is a problem. Sorry, I have to skip ahead um a two slides because um very very recently just like less than two weeks ago um like there were studies that have been coming out to show that Google's really cracking down on these self-promotional best of listicles and that is when people are are seeing that they they've used to use this as like a hack uh to be able to um influence AI uh answer generation. So they just programmatically create a lot of best list blogs on um domains that they control and then just like spam uh the web with it. And we see that um not only is this something that Google typically like they try to punish any sort of um like blatant manipulation like this but uh also AI um owners platforms are are are looking to avoid like they're looking to penalize this sort of manipulation because overall dilutes the quality of uh responses and it also like um hurts the trust of the people that use Google and the people that use AI platforms. So that's why I wanted to mention that it is important that when you look at the lists, you should um look to get mentioned in in ones that have more authority but also reputation. Um don't just blatantly create best x lists out there and and spam them across the web because Google's already cracking down on them. Uh and another point that you can look into is making sure that they're updated. So if your um if your brand or your website is already getting mentioned on uh different articles that are um updated um like a long time ago and they were published a long time ago. Um note that Chad GBT um overwhelmingly prefers to site blog lists that are from like the last year or um and so much less than 2024. So this this study was done in 2025. So you can consider that like within the year if you can update any of the blog lists that your brand is being mentioned of as best x for something make sure that the content is accurate it's up to date and um see if there's any ways in which you can get yourself into uh lists that are very relevant to your business that communicates the strength of your brand. Uh and in fact uh 76% of these tops pages were refreshed within the last 30 days. So like the update uh uh frequency can be really really uh much sooner than you think. Um, and when there's a lot of talk about creating new content out there, do consider if there is any um, if there is any existing pages that you've created that has been some time that you have updated them and look to see if they can be uh, updated to get more uh, relevant information, more accurate information that properly represents your brand today. Um, but yeah, it's important to update your content. So um not just uh your own pages but also that on third party on third party websites that recommend your brand. Um those that are able to uh update their websites and mentions um pages that mention their on their brands more consistently um basically uh reap the best results. Okay. And then we saw this uh like slide already and content outside of your control. So we have to talk about this because um like chat GBT air overviews and other AI assistants are definitely um looking to uh sites from authoritative sources. They they tend to look for um different kinds of pages that they that the team be working behind are trying to get as um accurate as possible. And a number of these pages simply are not pages that like marketers can manipulate. Um, Wikipedia stands out as one of the most cited content types far ahead of the others that we've looked at. And you can see that only a third of Chat JPT's most cited pages are like something that you can pitch uh to the authors or to the to the PR companies to to say hey can you mention my brand or can I uh can we work on a piece that where I can include a mention of my brand in it. So there in these cases a lot of these um formats for example like uh educational pages is like a really small percentage. The rest are things like uh reviews, language and grammar sites, uh news and media actually news and media you can influence but like um basically like homepage and landing pages is a large portion of it and you just can't control that of other people generally. So and that is important to to take a note. So what does that mean for you when you want to improve uh your content or your mention uh or your brand mentions outside of these uh um because of this problem where a lot of it is not something that you can just pitch to get your brand mentioned in. Um the the short answer is just focus make sure to focus on the pages that like has your best uh chances of getting influenced. um find collaboration and PR opportunities uh by going through uh a tool like Brand Radar that can help you look into what are the cited domains and cited um cited pages that um are cited the most in your industry. Um like you can you can see that for example a portion of this uh comes from the QA community forums and like like for example Reddit and um these places are are known to a lot of these um the best gated uh subreddit communities basically um kill any sort of uh initiative to for of self-promotion. So many of these subreddits don't allow for self-promotion. So the best way is to just get involved in these communities like organically to help be helpful and like answer questions. So um you it's a lot of this is trying to go about go around these like restrictions um and that's just the case of the industry right now. Uh oh yeah and like um do note that just because uh a page has visibility in uh chat GBD doesn't mean that they have organic visibility in Google search and vice versa. In fact, you would see that in chat GBT oftent times there are pages that um for example may come from a reasonably um authoritative domain but uh the page itself bas does not show up in Google search at all. So um these are all opportunities for you to consider look at just because um something has uh like a a well-renowned uh website that is getting cited a lot uh may may be very difficult to get into. uh doesn't mean you have to discount all the other uh websites out there because they could still be getting used to uh get cited in chat GBT. Okay. Uh and now what I want to do is I want to switch gears a little bit and I want to go into brand radar because a lot of the cases that we mentioned here um is is can be answered via looking at uh a tool that for like Brand Radar that can give you answers about what exactly is getting cited and how to make decisions about uh what your content strategy should look like. So, just give me one second here. Um 96. I'm going to switch gears a little bit. Are there any questions before we we switch gears? There are questions. Okay, let me just quickly look through them to wait, let me just switch to sharing. Let's switch to brand radar here. Uh yeah, one second. Let's switch to let me just quickly look at the Q&As's. okay. Um what I will do is in the uh following up email. So a lot of these findings come from a number of different blogs and uh what I'll do is I will share with you guys a copy of the deck but I will add information about which uh studies that these uh data points come from. Okay. Uh and from there you can look into detail with like how they were conducted and um what other additional findings that we get from these data studies. And the alternative also is to uh if you don't want to wait is to go to this page uh actually I think AI search is better here this one where we have done a lot of blogs and different studies on AI search as well as some like more recently that I'm not mentioning here in today's webinar. So I'm going to post it over here to everyone. By the way, I shared the link of chatbt versus Google. So this is the page. Remember that we looked into uh where we mentioned about our vcoded page where we compare AI versus search analysis. Um this is the page. It's chatgbd versus google.com. I'm going to post it here. Uh and so you can see that like I think the graph that we took here uh so you can actually deselect. So I deselected all these links, all these sources and I just looked at the ones that we tracked for AI. You can see that um at least for our traffic share by source, chat GBT still re supreme across the other platforms. Okay. Anyway, let's go back to uh brand radar. So in the first case, we want to look at um like content length. So for example, do do people want to suggest uh maybe their well okay I'll show you the example that I've uh tried to make an example with which is I looked at the brand Oatley sorry just to give an example which is a brand of oat milk. so let's see if we can get some good examples here. Uh after this I can do this with uh any one of your um uh people in the chat if you want to suggest your brands afterwards. But here I think we can give something that has known u brand visibility stuff. So let's explore So how we looked at things is that we looked into the top cited pages. For example, you can enter in your brand or your competitors here. And then we um the first um in the first case for example where we looked at content length and we see whether or not there's correlation of how they get cited. We basically looked at what um pages get cited in AI overviews. So we're going to focus on AI overviews here. Uh and then we saw that uh like amazingly for example I think here I'm supposed to put silk over milk huh sorry that should improve things up. There you go. So, okay, you can see that um a lot of these uh pages get cited. And then if we want to look into like more detail about this page like for example this is the most uh cited page for this particular brand. Uh what I can do is I can uh go into oops site explorer. Yeah. So you can like analyze these things in site explorer to see um what kind of content that this page has. So uh with the page inspect tool. All right. So um this is a tool I think like um not everyone knows that we have but if you if you do know that it's here we basically at HFS have a equivalent of a uh way back machine where we are storing different versions of a website um that and whenever an update is pushed we can actually see um like what what updates or changes were made we compare them side by side. So let's compare for example with can I compare one second this is April and then I want to compare it with Feb. All right. So, like sometimes you get to see uh things like um how they've changed or let's compare to a year ago. Okay. But basically, can't really see the changes here, but basically um if you compare two pages, you can see if um there were changes in brands, there were changes in descriptions that were being made. Or for example, if you requested for uh one of the uh PR um posts to come out and they have not added a mention of your brand, you can you can basically use a tool like this to check what it was mentioning before and what it mentioned after. Sorry, this is not the best um example, but like this is basically a way for you to check um how content has changed over time. It's also mentions about like uh anyway. the other thing that you can do is look at um how for example how often YouTube is being cited uh as a source. here. It is not particularly cited, but we can look into uh you video visibility for YouTube here. Um if you don't know already, any of these beta indexes are available for you to try out even though you don't have a Brand Radar standalone purchase made. So, if you have HFS open right now, you can go into um brand radar right now and check to see um what are the um videos that are mentioning uh your brand or mentioning like uh product names or or like product terms in your in your industry. Okay. Okay. So in here like for example I can see like how many views it gets cited and like uh you can see that they are basically getting recommended different kinds of uh products within the OI brand different languages uh like the different content types as well as the influences that would uh create content that would mention your brand. And from there that gives you ideas of like what uh kind of creators that you can reach out to as well as what kind of uh content you can create to inspire uh the video content that gets created uh from these u providers I mean from these uh YouTubers. Okay. The same can be done for Tik Tok and the same right. So, so this is so this is a way for you to look at the a if you want to create your own content, what kind of content gets views and what kind of content gets cited. So, for example, this one got cited by uh like different AI uh perplexity answers and they're in German. So, I won't be able to give go into detail because I don't speak German, but you can see the point of how um data like this is useful to you. and then what questions like it helps to answer stuff like that. But this is basically useful in that regard. Okay. So you can see that this link uh this page got cited in AI overviews as well as AI mode. Okay. Um so the other thing also so let's do an example with uh suggestions from the audience. You guys want to suggest your brand as well as maybe your industry. Just give your brand and your industry and from there we can use AI to suggest your competitors. only one suggestion so far. Okay. Uh could you give uh your brand name as well as your industry please so that uh we can help to generate some competitors and see how um the data stacks up. Uh Dean, I have you have questions about the AI data insights and we will share like a few more. We can share a few more that I have backed up if you like. Um but for now I'll give you this example because a lot of the data that we talked about today it's also important to uh see how you can do the same study yourself and see what um is specific to your industry that you need to take into account when creating content. So let's look at okay we'll take the example of uh feedback post then uh I think the industry is employee engagement software. So, let's see what um brands get mentioned here in AI overviews and in AI answers. Um oops. Let's see if AI is able to find some suggestions here. Okay. So quite a number of these. So I'm not sure about glint but okay. Let's see. So like in this case we want to look at the cited pages. Oh. So we don't really see much here. Let's remove the niche and then see if there's any help. Okay. Let's refresh. I think there's some something [clears throat] strange going on with the formatting. So, I'll check with the team. So, you can see like these are the top pages that get um cited in AI uh in this case the AI overviews. And as as we saw earlier that um landing pages basically get um cited as well as a source. So um if we want to look into more detail, we basically have to see what sort of answers uh what sort of prompts um generate the sighting of this particular website. And then we can go into more information as to how it's being cited. Like here it is citing competitors but not uh your your brand feedback pulse. And then you can see all the citations here. And then what we would do, I mean the first thing you can do is do a cursory track and open up to see what sort of page this is. Basically it is a best x list that we saw earlier. So it's it's definitely still getting cited by by AI. Um but you see that like a lot of these um of the ones that generated before can are getting penalized over time. So let's go back. So you see best x list still um seem to take prominence here. See this is another best x list. Um and then this is a blog. So you can see read check out the blog and more information. Check it out in more detail. Right. So so basically this is now wait this has become the best sex list as well. So yeah, like uh I think a lot of these um that's AI is citing really is about like the different uh best um software for so use or best in 2026 and 2024. In this case, you can see that this is updated for 2026 already and it's only February. So yeah, updating your content definitely is important in this case as well. Um, if I want to go into detail, I can go uh into site explorer and put in the website and I can see and I can see like uh for what keywords where AI overview shows up as well. In this way, if let's say I don't have brand radar, I can do it this other way where I can add a filter for certain features. This is basically in a like API method. This is how we did a lot of the studies that we showed today. So we just looked at uh not include I want to look for includes target in and apply it. So I will not be able to see the generated uh response in AI overviews. However, I'll be able to see the keywords where the response was generated as well as what was cited in the AI overview. So, you can open up SER here. So, for example, this page was cited in uh for when in the AI overview that was generated for tool performance management. So, and it has been going on for a while like even from last year. So, you can see that uh AI overview has cited these websites. Um, another thing that you may notice when you look at the the before and after is that um AI can change the pages that it sites. However, um uh in terms of a semantic meaning, they do not necessarily change the the the recommendation or the answers the meaning in the answers that they give. So um so it's important to improve your um visibility across all kinds of websites and not just a select few. So yeah this is how you can see how uh whether or not your uh pages or maybe your third party pages that mention you uh how they get um cited in AI overviews like over time without using brand radar. So this is like something useful that you can take away. okay. And uh one other thing that we noticed when it comes to AI overviews. So I just want to pull up see if I have the slide here. Um oh we are already almost at time. Wait, let me let me help to answer the questions and then we can uh look at some backup slides that I have created where people who are looking for more for more data insights. Uh so some people have asked the questions. Where's the Q&A? Okay. Um, most of votes. Does that mean so Andre Andreas asks, "Does that mean you should completely renew your content every year if AI ranks fresher content higher than older content?" Uh, I think your most well your most important one should be refreshed. uh any pages that uh update incorrect information that they have incorrect information about your brand like wrong prices, wrong offerings, uh wrong positioning of your website uh would be would be pri should should be prioritized to be updated especially those that are anyways performing well uh in search or as well as um pages that are getting cited um regularly by um by AI and this is again something you can check in radar. I know brand radar. So I know you don't you don't want to see too much of the product promotion, but like if if I look at it, so for example, if I if I'm PlayStation and my competitors competitors are Xbox and Nintendo and I look at like the top-sided pages for example and all the top pages here are basically mentioning like something that is incorrect about my brand, I would definitely try to reach out and see if it if they can be fixed. Right? So um in this case you can see yeah the top is a YouTube video which is going to be hard to fix. So then so then you have two options. The first is to basically focus on the pages that you can influence and the second is to create content that is better than the pages that you're that you cannot influence uh can answer the questions. So uh for example like there were a lot of uh so for example there's this page right? Let's just give an example. So this uh page is basically trying to give a guide about how to do uh and it's from Microsoft how to record your screen on Windows 11. So do basically how to do screen recording on Windows 11. So why don't we just do this like look at it in search and see how many people are um creating guides trying to compete with Microsoft here. You can see that a like get involved in a community hub comm community hub will help to influence your ability to get um cited in search. But if you look at AI overviews, you see what gets cited, you can see that like okay, Microsoft is at the top, but it's also citing from all these other websites that are not Microsoft themselves. So they're they're trying to get a slice of the pie and try to get their their brands being mentioned or their recommendations being mentioned here. So that's a similar way you can look into uh if if there is a page out there that is um mentioning a brand or like um not not mentioning a brand and you want to um you see that there's opportunity for you to get yourself mentioned there. either create a better version of what it currently gets cited or try to see if you can influence that page or reach out to the author in some way. But those are your two options. but yeah, focus on ones that are like already getting cited and then focus on the other pages. Uh how often would you recommend to update content in a blog page to get help uh to help get picked by AI citations? We have one I think this is by our colleague Louise. She helped to create a blog as well as she did a webinar last year on how to republish your content. So this is she helped to provide a guide on how to decide which ones are worth updating, how you should update them and she had this eight steps here. Yeah, I remember. So, so like that talk about how you can how to prioritize which pages to to update. So, I'm going to link it here in the chat. Okay. Um, it it helps you because like when you update pages, you're doing you're hitting many like uh birds with one stone. you are updating information so that actual readers can access and and get things update um like understand the the new information that is out there. It's also for search engines. It's also for AI crawlers. So like you're you can update information basically to hit all of them with a guide like this. Okay. So it would help basically help you um figure out like how often you should in this guide you can help you figure out how often it is. But to in a general case, if you see your competitors getting um like if you if you do a search like this just in Google and you see that the the publish dates on all these pages, like for example, this one, this one, this page in particular, if I look into the publish dates, or update. So, I'll look into like any of these top pages and if they have a publish date or update date uh and it's, you know, last year or two years ago, maybe I will not worry. But if the top sited pages that you're trying to compete with are all in 2025 and 2026, that's your cue that you should probably update your content. Yeah. Uh, hi. You mentioned this is Sergey. How you mentioned the pages with SL blog, article, post, magazine, articles in your uh UR URLs right below. How about landing pages for services and products with URLs? Um I think in our slide uh service we had product with I think the service and product pages. If I change one second, if I share my slides again, product pages, reviews, news, maybe pages, I think it's not even it's basically it's way lower on this list. I remember that there was some discussion about it, but yeah. um product service pages are even less cited unless they can also double as education pages. So if there's information about how to go about a procuring a product or or purchasing a product that is less than just like a link uh to to a product and stuff like that um then then it can get your chances higher for being considered as educational page uh for AI to to consider insight. Now products and service pages will likely get a different like visibility by AI through the partnerships that a lot like for example OpenAI is pushing out where they're partnering with uh Shopify or they're partnering with uh other e-commerce brands like Walmart and stuff like that to uh basically recommend products to you and and offer a way that users can purchase um products or services u through AI platform without going directly to the website. So if that um like as that gets developed because it's not currently like developed fully as of now then yeah you will see you will see more of these uh like product/service um pages get cited more or recommended more by AI because it's literally a way for them to say would you like to buy something so that mechanism is not like you you know fully ready yet but uh a lot of the AI platforms are looking to ways to make it ready and then these service product uh pages where you can buy something or you can sign up for something will become a lot more relevant. So definitely look out on news for that. And uh I believe that chat open air has already started to roll out ads uh in the US. So I'm not based in the US so I don't see the ads but if you but um you can take a look at that and that would usually be people running ads for uh product or service pages. So that's also cases where uh you will have to consider that as a use for AI uh mentioning your brand somewhere and we don't have data about that yet but we're definitely looking into doing a case study on that very soon. thanks Nicholas. Um Nicholas asked where can I find these insight insights after the webinar? I will be sending them to you via email. um I will send a copy of the deck as well as uh some additional information uh related to what people have asked today. So like that would help to link to the blogs that were mentioned as well as u give more information about um what is basically getting um preferenced by AI's uh answer generation. Yeah. Uh anonymous person asked is there a significant difference between best and top based on our data? No. I think semantically um they are considered very similar and and as I shared in the earlier case that they can look at like they can appear as um seven most recommended seven or like 30 of the um most suited and stuff like that. So um in the AI era, it is better to look at what pages semantically mean rather than focus on whether I should use the words word best or top or latest or newest or uh most recommended stuff like that. Like it's more important to look at what the page is the content is semantically trying to communicate. Uh Bavo asks, "Will there be an addition of industry specific filters for YouTube visibility?" Uh for example, the renewable renewable energy sector. There's a Okay. Um this is seems to be a uh feature request. So uh we have we have because it's a beta tool and we're so happy that you're checking out this tool. So I will switch very quickly to our main website. I mean the the page I was showing earlier. Where is it? This one. Not this one. no it's HS. Can you Which page am I sharing? One second. Huh? Yeah. Okay. Can you see my screen now? So, we have [clears throat] where is it? It should be brand radar. Here we go. Um since it's beta, you can use uh the YouTube index to to basically try to see what uh videos YouTube videos are mentioning a brand both in the titles as well as the transcript and stuff like that. We are looking to improve the tool. So um if possible Babo and if anyone else has like re like suggestions they would like to see in the tool please go to this uh our feedback request board and and share it there. I can help to share it for you Babu but like uh for anyone else definitely look into uh sharing any feature requests they see like to see come to our tools and that will be the best way of getting uh seeing whether or not a a feature like this can be prioritized for you. Okay. Um, someone asked, "If AI systems mostly site content outside of companies control, what specific types of content can brands what's what's sorry if outside what specific types of content can brands realistically create and own to increase their chances of being referenced?" So, how would I go about this? I would basically so let's say if I think a number of people like always suggest like have been suggesting about their industry so let's say I want to use the example that I shared earlier because it was quite but let's let's do something like running uh sorry let's look at I want to look at running use. So I don't need to put in any brands here. I can just put in like a market or niche. What I'm trying to do is help to answer this question of like uh what kind of content can brands realistically created on. The first thing I'll do is check at what content is currently getting cited the most. Right? So in this case I' I've not put in any brand. and just put in like running shoes is a market or niche and we can influence Wikipedia but we can influence a lot of these. So a lot of these like in this niche they get cited by a lot of different an uh AI answers here. Uh and I think in this case is air overview. So let's look at chat GBT just for a change of pace. Huh. Give it a second here. Okay. So, um we're seeing the same trends happen again that a lot of the best x blogs are are basically showing up at the top. Uh and again, this is already updated for 2026. So, a lot of these like top websites that are getting cited, they basically know the drill. they they'll create these like best guides but they also come from like um websites of uh reputation. So I'm pretty sure if I check the so if I check the domain rating for example of this website right yeah 85 and this is chat GPT so like this is so and for for like a wellestablished domain as this one it is may be difficult for us to to penetrate and just write to the author and say, "Yeah, can you can you recommend my brand?" It would be like a hard sale. So then like we just go back to the fundamentals, look at like what sort of page this is. Um is there uh like is is it just because they are a website of authority like runners world that it's a useful um page and therefore AI chooses to site it or are there things that can be done to improve on this version of the page that page that they have? Right. Um I would I would that's like the first way I would go about it. It's like first is you can see what is very useful here is that they have um pros and cons like tables and allow you to quickly like buy the products and stuff like that. So these are these are points of like okay what is what is structured on this page that is good or not good that I can improve upon. That's the first area is like the on page like doing uh onpage analysis of its structuring and see if I can create better content that is first structured in a better way and then the second is I will look into what is let's maybe filter for one location so we can see the language yeah United Kingdom's fine I'll look into um what uh questions it gets cited by uh it gets cited as an answer in and then see whether or not there are questions that are not being answered very well and then create content around that instead. So let's pick this one. It's still so for UK it's still a small number but right they're they're able to answer questions for for different genders. They're able to answer uh but it's a very specific query. So like it's comparing one shoe versus another shoe. So maybe if you want to create a content that can compete for more nuanced answers, you basically get more specific. you like focus down on more uh specific like comparisons of different products or or specific um feature or characteristic of of your brand that you would be superior in and create content around that. So then when uh AI tried to look for answers, they tried to look for uh like different aspects that people ask questions about that you could serve a better need for that one specific thing rather than answering a a general question like what is the best A6 trainer. Okay, I hope that is helpful. Uh, okay. So, wow. This one is vote. I voted a lot. Clients always ask about KPI. Can brand mentions be a KPI for customer? Even brand mentions not generate clicks. Uh, clients always doubt how what is the value is if no one clicks on the website to buy. Um, that is a very good question. But, um, you should not. Okay. So like on on brand reader here, let's let's take uh one of the demo examples that any of you guys can check out as well. Any of these demos that you that I was clicking on right now, you can see all the data for. You can see that like the metrics that that we showcase here is AI share voice uh like search demand and like web visibility. Uh these two are not necessarily like AI related but at least in AI visibility we looked at um share voice. We show like how many times it's mentioned uh like how many citations there are and and then like impressions where we basically scale um we get like a rough overview of how much people are searching for uh for the question that gets prompted to the generated answer. So okay. So if what like in terms of how this helps to answer the question of like how much is it worth it? I think um it is not reasonable or meaningful to just look at oh because you get mentioned by uh 69,000 times in AI overviews. Therefore your uh therefore your brand visibility is good. this number is all all of these is entirely reliant on for example in our tool the keyword database that we have. So in reality, not only are a lot of these keywords, even though they are generated based on like what we know that people search, like real search data, um, but it does not definitely correlate with what people are actually entering into AI assistance or AI mode or or or things like that because until today, there's no like reliable way of getting that information about how often people search for specific prompts or queries and and to to get to see how eyeball. So, you can't really have the same mindset of how SEO was before with traffic and uh searches, but you can look at something like share voice because um if there is a large proportion of uh your of answers getting mentioned and that is growing compared to before that generally is helpful in in improving the impressions that um your brand has uh over time. So like absolute numbers are not that useful. But if like if you're able to see an increased slice of pie where your brand gets mentioned more and hopefully for you know positive and relevant things then then that basically translates into improved brand awareness that you can then communicate to your clients about that okay like in this space um you are increasing your like people are more aware of you as a brand compared to uh compared to before. And then the second thing is that if it does not generate clicks but it does generate revenue or sales then uh then in bottom line then that's the most important for these clients. So for example at HFS let's look at HFS right to to give a comparison our brand. Did we get hit in terms of organic clicks? We did. Uh let's look at let's so I'm going to untick this but you can see that like we got hit uh if we compare the last year or even like the last two years right like we had like like um traffic that was going up and then and then basically we lost a lot of it again like our traffic has not been doing so well in the organic side. However, revenue-wise, we are not like suffering necessarily. We're it's not like directly tied to the uh number of uh sales that we get or subscriptions that we get directly. We we don't have like a proportional drop in sales. Yeah. So, like I just want to see that there's a trend that we can show. It has gone back down to basically 2023 levels here, right? But just because the clicks have changed um doesn't mean that our sales have changed. So um if they are worried about KPIs um another thing that you can do and this is like outside kind of outside the scope of this webinar but like you can see if you can ask for uh your clients to ask their customers that they come on to and for more information about how they managed to learn about uh this brand and why what helped them decide to to sign up or they or to make a purchase. So this is a question that I think we have to start um asking a lot more now just because like the reasons for attribution of of a particular sale or a particular um subscription or getting a new client is a lot more fuzzy. People are using all sorts of ways to get to know or get recommended a brand. Oh, and another thing in that well I mean so but the questions you're asking is about KPI. So I still think I still stand by what I what I suggested today which is that you still have to look at bottom line and something like uh share a voice is going to matter more. Yeah. Uh, and I think what Ben mentioned could be useful to uh, and and if you have even if you don't have the brand radar um, add-on, you can still look at um, search demand. So like what Okay, so Ben uh answered the anonymous person's question about KPI, which is to look to see like whether or not there's an increase in branded searches. So uh this is something that actually you can check in brand radar. Now even if you don't have the brand radar add-on purchase you can look into how often a particular brand is being mentioned or being searched and how that is changing over time. I think you can wait I think it's best seen in the overview. There you go. It's this one. There's a search demand tab. You should be able to see data for if you search your brand against your competitors, you should be able to see search for them and how that has changed, right? If their awareness has increased based on like the efforts that that you've put into, people start searching about it more, even if they're not clicking, even if they're like getting to know things in uh air overviews. And a lot of times now when people search about something and they click on a website, they already have like like um purchase intent or they have like a buyer's intent. Right. So that's that's what I would most likely suggest. I hope this was helpful and thanks Ben for giving the the the tip. Okay. We are currently at 620. So and I think we have answered most of people's questions. So, what I want to do is basically um do another shout out to uh on our blog where we go into the different AI um search related contents that we've been pushing out. And if you want to see like uh new uh studies like quickly, you can also follow our um authors on LinkedIn because they do a lot of posts related to the stuff that they have been pushing out as well as we share a lot of interesting work that is done by people in the um AI visibility and SEO space. So these are the names. So Glenn Elop, which I think many of you guys will know, Ryan Louise. So definitely go follow them if you can. So, yeah, I'll go ahead and link this in the the blog for you guys. uh in the in the chat for you guys. Okay. Um and I want to do a quick plug. I don't have a slide for it but I'll do a quick plug where next month we will be looking into uh NCP uh servers as well as using AI for automation where where you can um uh like the HF's MCP server uh connector you can connect it to your AI tool and basically generate different tools as well as different reports um and I'll be going to a few use cases how you can do that how you can uh use that to improve or speed up things for you as well as like um help you create content better and things like that. Okay. Um, are there any more questions that we have today? If not, we can go ahead and end today's webinar. Anyone else has more questions. Uh, if not then, uh, we'll go ahead and end today's webinar. Thank you everyone for joining us today. Um when you close the webinar, you'll get a feedback form to um share what you liked or did not like in today's webinar. Uh and u we will use that feedback to improve future webinar content as well as and you can also give a feedback of what else you'd like us to talk about in future webinars as well. Uh thank you so much um and sorry again for ending it a bit uh starting a bit early uh bit late today but thank you otherwise for joining us and I hope today's content was useful uh and we'll see you guys in the next fun.