How Generative AI Is Changing Online Behaviour and What It Means for Marketers

Hello everyone. We are now live. Welcome to today's webinar which is about AI search and how generative AI is changing online behavior and what it means for marketers. I'm Charlie, CEO at Exposure Ninja, a search marketing agency, and I'm joined by the fantastic Fernando from Semrush. Hey, Fernando. How are you doing? Hi. Hello everyone. Good to be here today. Thank you so much for joining us. We are incredibly excited to see your presentation today. And I know that everyone is going to have lots of questions. So before we get into it, if you do have questions, we're going to make sure there's a little bit of time at the end. You can pop your questions directly in the chat and I will make sure that Fernando gets all of the best ones. There will also be a recording of this webinar. So if anyone has a power cut, a thunderstorm or the postman arrives, it's absolutely fine. A recording will come to your inbox after via email. And without any further ado, Fernando, I will hand over to you to take us through the AI effect. Fantastic. Thank you very much and welcome everyone. I'm Fernando Gul again. I'm senior market research uh manager at Samrush. been working for the company for about uh 15 years now and over the next uh 40 minutes I'm going to be sharing with you I'm going to be guiding you to explore uh something that is quite quite interesting and quite impressive that is how the AI the LLM platforms are influencing search behavior and strategy uh for a decade as you must know uh the goal was crystal clear rank on Google right if you are an SEO or if you are doing digital marketing you need to rank on Google mostly for a couple of decades conversions drive traffic that's it but something has changed and that's something that I want to share with you not with my own opinions but with data data as you know uh means facts and we have that data something that we are uh sharing with you today it's going to be a lot of data so be prepared in terms terms of getting acquainted uh we might we might connect on on linkaging uh there you can see that I've been uh working a lot with data lately with companies like statista eurom monitor we were gathering in a lot of uh data to do overall research and that's something that I'm really uh passionate of so in the chat give me a thumbs up if you use sand brush before. Let me check that. I can check that. Yes, definitely. If you use Smrush before, nice. I can see some thumbs up. Perfect. Yes. Well, great. So, what is SMBR overall? I'm not going to beat you the product because more or less uh if you're a digital marketer you're you already know what is Samrush but basically we are an online visibility management platform and we have been in the market for a few years actually 18 years now and we know a little bit about SEO of course and right now we are really passionate about what is happening with with AI these companies are trusting us they trust our results and this is uh something really great for all the industry. What is SEO? Well, I'm not going to tell you what is SEO, right? So, just imagine what what is the first thing that comes to your mind when you are thinking about SEO? Keywords, back links, rankings, all of that. Yes. And you're right. Is SEO dead? Of course not. SEO is in a really good shape these days, but we have a new uh expansion to our realm which is AI search. SEO is still there, but AI search is a new comr. So what is going to be the agenda for today? We are going to be covering four main aspects today on our agenda. First, what is actually happening? Where are we in terms of AI search? What is the situation? The second point, our approach, SR brush approach to AI search. How we are using our own tools to work our AI visibility. Then we uh we will be analyzing one of my favorite companies. Uh this is going to be a case study. So you're going to see how they are crushing it on AI search. how they are performing, how they they are creating their uh context engineering and the last point is going to be AI search with Simrush, how we are uh using the tools and how you can use that as well. So, hope you are ready. Let's start what you're going to be learning today. This is also very important. How is changing how people search, evaluate and trust? We need a foundation before I start developing uh this topic. So if you want to share this with your teammates uh inside your company uh we need to have a foundation. Second, why zero click results and AI are generated answers are reshaping the SEO strategy. No more clicks. Yes, but maybe that's a good thing for all of us. And the last one is going to be how brands can build uh their AI visibility and uh credibility in this AI search system. So now that we are all established, let's start. So what is actually happening in AI search and let me be clear. And the name is not a problem anymore because uh we started 3 years ago with search generative experience. Uh Google was was uh promoting was um telling us that okay search gener generative experience is the new way that we're searching. Then we had the AI overviews massively adopted uh since the last couple of years live globally only last year and this year we are in fully AI mode. So if you are searching something you can be in AI mode and maybe you don't even notice that yet but this is where we are. This is the same revolution. Three names one direction. AI is becoming the interface between users and information which is huge. It's massive. This is the data behind all the uh graph stats and trends that I'm going to be showing you today. We have been tracking [clears throat] billions and billions of co of of keywords uh through all these countries more than 190 countries and we have the um the support of 10 million marketers that already tested tried our um professional tools. Everything in this session is going to be grounded in observ behavior not a speculation. So let's start from search to conversation. So this is the foundation of of this um overall conversation. For about 25 years we have the keyword-based search. So we needed to think about our needs on any product or service translate that into keywords. Basically using the uh search box uh to type our needs converting that into uh keywords. Then we could pick choose any source that we want. Usually everybody knows we have the first three positions that they were that they have the most uh impact on our choices and we can click on those results. 25 years two years ago that has changed completely. Now we have conversations. Now we have conversation of the AI natural language. We don't need to translate anything. We don't need to choose anything. We just have one single syn synthesized answer and basically zero clicks. Nobody's clicking on those results because they are no results. You have the full answer. You have the complete um definition of what you want. Brand visibility measured by AI mentions in that specific answer that you have after the prompt that you have. So 25 years gone, nobody's saying anymore. Uh Google it when you're going to do your research. Most of the people that are not saying anything uh at all. Some of them are are saying let's chat GPT it. I heard that phrase which is pretty new. But Google it. It's not trending anymore. That's a significant change in our minds. So that critical thinking that we had before in our brains the layer that we must choose something that we can uh relate information that we can translate this information that we will be comparing judging uh deciding is forever gone. Right now that cognitive shift is change in the side of conversational AI and the implication is well basically there is a new way to acquire knowledge that's the important thing here the new knowledge acquisition system is here with us and we need to trust it we need to work with it basically we were changing from critical thinking to algorithmic trust which is how we used to think and how we are thinking right now single recommendation from AI search and passive acceptance we just accept the results question for you are you using AI 100% of people yes I I know for sure you are using AI the question is are do you trust the results of AI there's a lot of hallucination over there so overall only 20% of the um users alm platform users they trust the result they need to check it they need to verify it, they need to validate it. So that's the uh paradox that we are living today. In terms of keywords, let's take a look on which keywords are actually triggering these AI overviews. So we are changing the way that we are looking uh the AI overview results by intent, right? We have the commercial, navigational,formational, transactional uh intent in terms of keywords. As you can see here from this graph for uh for um last year, right, the the last couple of years. Overall, by far the informationational intent behind keywords are the most popular one to trigger AI overviews. So yes, knowledge acquisition, informationational intent, they are all aligned. So people, users are going mostly to LLM to be informed about something, to discover something. And you can see the numbers. That's a huge huge chunk of those graphs month over month, year over year. completely different from the transactional and the commercial. This year by far we can we have seen that okay information is still there navigational this year navigational is also entering in this uh uh user behavior what is happening well you can see data from this year so navigational is here informational of course transactional is coming also mostly chip so this is uh these are the start from chajipro chat activity prompts that match traditional search terms and having navigational or transactional intent. So yes, people they are trying to compare prices. They are uh trying to uh check for better options. They're going for the transactional part of the search behavior which is really interesting as a development. Last year they were trying only to find information right discovery. This year more navigational and transa transactional intent is coming mostly in charge GPT. what is happening in terms of overall traffic, right? In terms of overall traffic, you can see that well, basically uh Chad GPG is growing a lot since December last year, but something happening in January this year. Something big was happening overall uh January this year. I don't know what what was happening. Maybe that was cold. Maybe it was Iran something but they uh stopped growing uh with their monthly traffic and they stabilized in in their uh traffic trend as you can see but the growth was impressive. Then they are going a little bit down in terms of total visits and total users. Overall the referral traffic from CHP to the rest of the internet is huge by far. You can see that line. I hope you see that line. That trend line is going up up up every single month and it seems unstoppable. The penetration of LLM usage in this case Chad GPT is by far one of the biggest ones that anybody see have seen in the couple of decades. What is happen with the referral destination? Top Chad GPT referral destination. So after going uh to other platforms let's focus on chat GPT you can see Google by far is the top one then you can see YouTube then GitHub then Amazon then Facebook those five are the top destinations referral you can see the rest of the list but you can see here that these are open sessions on JPT people are just getting more into the platform thank to these referrals what is happening And this is something interesting. You can see here a decline on the trend. What is the trend here? So this is the average chip prompt length. If you can see at the beginning at the mid of 2025, more or less the words uh the number of words that have been used on a prompt uh where if the search was on or the search was off were about 30 or 28 words for every single chatist. For example, uh rewrite me this document uh with a corporate tone for my uh CEO whatever you write uh 28 words a huge um long text uh that was for a prompt. But this year just have a look February 2026 the standard amount is normalizing if you have the search on activated or search off or all it's around 10 words. So basically people they already understand that they don't need such large text for prompting. they are realizing that they okay I know what I want I just gonna describe it right so the amount of numbers that people are using in the last couple of years is changing really really fast and that's is uh fascinating as you can see over here you can see here for example this is extracted also from the last from the last graph 24 words when the search is off when the search is on is only four uh words as as as a medium, right? Uh this year the amount of words overall has decreased in search has increased just a bit but overall the amount of words that people are that users are using for uh prompting is reducing a lot. So what about the search language versus the conversational language right? So in terms of what what was the uh keyword first direct navigational awards the most popular ones was uh were best coffee NYC for example uh Python sword list uh iPhone uh 15 price those are keywords right but still this is these are this is search language this is the keyword uh mindset that people were using those type of uh prompts which are for us are keywords for um the last 25 years are in based 15 or 35% of all chat GPT prompts keywords still we're used to it but in conversational language the int intent first the curious one the contextual ones the most popular are explain this to me like I'm a designer or I'm a doctor or I'm a CEO for example what should I do about this situation help me think about this decision these type of prompts are 65 till to uh 85% of all chat GPT prompts so right now you can realize why maybe the amount of words for each prompt has been reducing so the conversational language is getting more more established so yeah that's that's the base for for this uh for this trend Now let's move to zero clicks right why zero clicks is accelerating now that we know u some insights about what is happening with jbt right what is happening with users intent with their prompt behavior with their search behavior uh we also have noticed that okay AI is giving you the direct answer so you don't need to click anything at all in your results in your LLM platform, the answer is already there. So you don't feel the urge to choose something, to select something. We already established that AI overviews are growing in terms of uh keywords triggering. AI overviews are on more than 25% of all search queries. That's a lot. And that number is growing all all the time. The SER has changed fundamentally for the organic links because they are no longer the first uh the first thing that you're seeing actually if you are um asking a question in Google you will see that basically you have your AI overview you have some YouTube videos you will have some ads and you have will have all the recommendations but the first position the ranking number one is going to be at the bottom. So user behavior has shift as well. User they are increasingly trusting more on AI synthesized answers over clicking the 10 first results of uh the search page. This uh evolution process that drop traffic is bypassed by far today. Three marketing shifts that are happening. Well, from traditional keyword research, keyword strategies, we are coming to uh semantic approaches to the context engineering. From link building, we are going more to API or MCP service for from ranking metrics. We are going to citations to mention metrics. This is something that we need to establish. We need metrics. Without metrics, we cannot we cannot grow. we cannot um improve our our results and this trust paradox is overall the whole um answer to all of it. We have the operational trust. Yes, you say that you were using AI. Of course, if you are into uh digital marketing world, you need to be using AI because that's leverage. That's leverage that you have as a professional market. Marketers achieve this uh professional efficiency with I for example by using uh to produce content uh you can uh create more you can do it more at the scale you can reach larger audiences but the at the other at the other side we have the strategic trust however we need to check we need to verify so at least a bit of this critical thinking is implied on the trust question that we have we need to be sure that the content is 100% readable, is legit. Uh we need to be sure about the accuracy of the facts that the AI is giving us. Uh the brain safety, the content quality. So all that strategic trust is transmitted by professionals like you and me. We need to validate those results of course. So these distinctions reveal a couple of things that we need to measure operational confidence separately from strategic trust. This is helping us to grow but we need to verify we need to validate with experts which is the situation that we are uh right now in terms of AI tool users our falls from statista we're analyzing this uh situation um the reality is that this isn't going anywhere it's just going to be uh exploding for the next uh five years at least. 68% share of US adults are going to be using generative AI tools um by the end of 20 uh 20 2030. So that's already a lot that's already a huge change in terms of AI implementation AI usage. What about the industry that are receiving traffic? And this is from Chad Remember that right now if you if you just started to uh follow this webinar, we are position oursel in into what is happening in the AI world. That's why I'm I'm throwing you so many data, so many facts. Right? So these are the top industries receiving traffic from Chad GPT today. Our industry is the first one online services. By far around 10 million uh users per month are requesting information, prompting uh doing all the research related to online services. Second one, the largest one by far is mass media PR uh information knowledge still there. Third one publishing. Then you have computer, software and development. Then education all of them knowledge acquisition industries. Then you have in information technology, news, retail, government, finance and so on. So by far right now our industry and the one we are working is the more required by users overall to have more operational uh leverage or to have more informationational uh leverage. in this case. Okay, let's move to the what zero click changes for SEO and content and we are going to be moving to the uh use case part but first we need to think about the new metrics right we are used to have the search volume clickthrough rate conversation rate bounce rate in the AI erometrics we have the AI citation frequency we have the zero click satisfaction well we don't have it yet but we will go There we have the conversational data. We have the source attribution patterns uh the citations as well. Right? The challenge is that we don't have many metrics to work with. We need more metrics. That's why we elaborated also uh new sources of metrics to understand what is the visibility that you earn by working your um AI efforts. Right? So just as an example when you are searching for something this is the situation today for example how to create an SEO strategy you will have this type of results overall um I'm right now looking at my uh 13inch uh MacBook Pro and I see that okay where is the first result where is the first position there is no anymore first position relevant here the AI overviews are massive are here to stay and they are really interesting to create an SEO strategy first define your goals. So I have all the information I need. I don't have the need to click on anything. Right? But also you will have news and you will have new ad formats. You will have uh PR results. You will have Reddit and forum discussions and you will have YouTube videos. Those these four type of results are massive. And of course if we want to understand how we can be here we're going to be following some few steps. But first let's try to uh position ourself in the timeline. Right? So these are these are the um the trend of traditional search the u blue lines uh versus the AI tools visit. If you can see here January 2023 the amount of users for from AI uh tools was around 5% 4% you can see every single year for the last three years this number was just growing so from being 0% in 2022 to being 4% in 2023 January we can see that in June last year the amount of um AI tool visits in the US this is US-based data was around 40%. 40% and that's already a lot. Then you can see here this is the distribution of visits to AI tools from US desktop uh devices also last year. You can see here uh zero business per month, one to three times business per month till 10 or more times per month. So the usage overall was growing in the um timeline for about also uh 10% 20% 40%. People are using AI tools these days 40% more than 10 uh 20% 10 times or more per month and you can see also four or nine times per month around 30%. So the usage is growing really fast. The implementation is here. The usage is already really high. Question for you for you all. So if you will be asked as I'm going to do it right now uh what is the p perception that you have? What is the most popular LLM platform generative AI tool these days? You can ask you can answer in the chat. I will I will read it. Fantastic. Uh yes, as you can see my graph over here, right? Most of the people think that Chad GPT is the top one most popular uh use used LLM generative AI platform, right? Um the second one you can see Microsoft copilot then uh Gemini, Perplexity, Clo is also really relevant these days, Como and others. But when we seek for data to uh to um analyze if this is true or not, the data is showing another reality. Thank to Patrick Stocks that he analyzed all this data for us. So basically Google AI overviews are the most popular results by far. Maybe you don't want to use it, but you already have. If you are searching for something on Google, they are just going to be appearing all over the place. Chachi PT is in the second place and you can see there they are big in uh if you are comparing to the rest to to Gemini to compile you cannot see their compiler perplexity and and and the others by Google AI overviews are massive and well the implementation is already past the uh beta or the alpha. they are already here and they are really prominent. AI is not expanding discovery. It's not replacing it. AI is expanding discovery. It's not replacing it. So if you want to be found by any users, if you want your product or services to be found by any user, you need to have this type of results. You need to be on uh the AI overviews on Jad GPT on these platforms because nobody is clicking no more to find information to gather information about product or service on Google because there are no clicks right and welcome AI AI SEO has entered the chat. We have a new friend. We have a new expansion for our realm. AI SEO is here. is part of our lives. uh it seems that it's not going to be going anywhere because it's just getting mass adoption and well what we should do about it well let's give it a name right we ask our users on on Twitter with different list how what you what should we call this new era of SEO most of the people are saying okay AI SEO sounds good AO something something else entirely bloggers influencers were saying okay which one is going to be SEO O go a lm whatever keyword research keyword research was telling us generative engine optimization uh in Google search console impressions you can see it's 72,000 so it's been massively used right AI SEO second one go SEO well it's the same thing the only thing that we need to use is use it and implement it right we we are uh witnessing of this massive change. So if you can see here those uh links, beautiful blue links, they're not relevant anymore. False. They are still relevant, but they have a different uh usage. Right now you can see you have the answers, but also you have the full uh prompt answer to our search queries or to our prompings from blue links. Okay, they are still there. They're still relevant, but we have answers. And if you are an entity, that's another talk. Of course, if you are an entity, brand entities are mentioned in the in the AI answers, which is great. If you are defined as an entity the if the um knowledge graph is uh [clears throat] is giving you this entity status that's perfect because AI search visibility is going to be mention you and if you are lucky enough it's going to be giving you this that citation maybe from your own sources or from other related uh sources. Most company get one or the other. They are mentioned or they citated by these LLM platforms. Very few have both mention and citation from own sources. What is happening? Well, when it comes to um being on once AI minds they prefer by far review platforms for product comparisons uh which is in terms of marketing customer success right they are helping users to identify the best possibilities to improve their um tool experience service experience Reddit threads for pricing complaints that's social media right every anyone is complaining they're using social media and social media, the ORM team is there. Developer forums, implementation details, that's support, that's actually actually also community building. New sites for company credibility, that's public relations. Support documentation for feature explanation, that's also customer education. Then you have your company's pricing page. That's the product marketing team. That's a product team who establish the market. So these six pillars are the ones that AI is digging digging all the time. We have identified these ones and we have related that to the teams that already are working on the same thing. Well, we're not in Kansas anymore to be uh only thinking about, hey, I need to have that ranking position. I need to have that uh that link. We need to be thinking about the overall result. This is an a company effort. This is a team effort. If you want to be on the blue links, sorry, maybe that's not the situation right now because people are looking more information on AI platform. That's the reality. And to be on those LLM platforms to have that AI visibility, well, we need to do we need to be doing other stuff. So that's the second part of my uh of my uh talk today. I know the first part where we are is a new thing is something that we need to be uh working on to accept it because that's the reality and here I'm going to be uh talking about what we are doing inside of Samrush to be better to be more to have more uh visibility on the AI results. First of all, as I was telling you, growing AI visibility is a synchronized companywide effort. Why? Because a LLM platforms are digging are exploring all that information from different aspects of the web and those aspects are well basically for us are different teams inside of company. Um here is how we are measuring success. We are expanding our own KPIs. So we have the brand visibility which is basically how often your brand name appears in AI generated content. That's the brand visibility, right? It can be a mention or a citation. That's already brand that's already good. Then we have the share of voice. So what is share of voice? This is how often your brand is mentioned within an specific topic area compared to competitors. Fantastic. We have our new metric, new KPI, right? Then we are measuring also sentiment. How many of your brand mentions are made in a positive way? How many of them are really positive in comparison to your uh competitors or they are neutral or maybe they are negative. Right? So three pillars for brand uh for to measure AI uh success brand visibility share of voice and sentiment and this is our framework. Oh something has changed there with the with the text. We have this this framework gets in and be trusted. Easy enough right? gets in win fab favorable mentions uh in AI answers and bet trust earn citations gets in bet trust that's it that's bas basically I can close my computer this is the most important slide but how to get seen right this is the sentiment game so let's build presence on the right review sites really important let's participate in community discussions or quarter are the leaders of it engineer UGC and social proof. So UGC user generated content and social proof create a feedback loop that AI s systems amplify. Well, this is how we you can get seen on AI platforms. How this looks like? Well, basically we have the listicles from user generated content. We have the review platforms. They need to be in a good shape. Of course, if you are delivering a service or a product, you those product needs to be good. So that's going to be redundant to say. And the review websites are also a great source. Uh actually they are the most important source information about who who you are and about what type of experience the users are having with you. Then you will see that okay, you need to scale that on Reddit. Reddit is well it still is. It was last year the first source of um experience knowledge for LLN platforms last year. It still is important massively important. So you you need to go there and create conversations. Uh take a look about the sentiment of users exploring your products and services and you can apply here something that is great which is customer advocacy or employee advocacy. Customer advocacy you can have your champions or your ambassadors right affiliates are also good for this they are going to be uh talking about you from their own experience most of the users they have a back bad experience they go to social media nobody's pushing uh users to share their good experience you can't do that so you will not have those negative that are particular that are are uh damaging your your brand image right an employee advocacy. Well, basically all your employees are also recognized as members of your company. So, if they're talking about you, social media, of course, that's also great signals for a great signal for LLMs. Scaling Reddit plus customer advocacy, that's a massive uh weapon. It's a really good strategy. Reddit having um ask me anything sessions for example uh sharing conversations there. You can see here we have uh different Reddit topics we are uh being mentioned is similar as for example worth in 2025 we have the sources of this uh prompt in on AI mode as you can see basically the sources are G2 crowd Reddit or Lar Lwing uh some of our partners that they are sharing that and voila we have a massive uh AI overview result or a prompt answer there. That is basically the information that we were um working on to be uh seen for um to be seen for these LLM platforms. Additional to that we can activate by affiliates videos on YouTube are a really great choice. You can see building partnerships with other uh companies for example today as well. This is a an important signal. But I can stop here just for a bit and say, hey, that's SEO, right? That sounds that feels that smell like all good SEO because it is these are good practices, but right now they are again relevant more than ever because LLMs they are not going to be taking the time of SEO 3 months, six months to recognize you. They are recognize you immediately. That's the power of AI SEO. So building a content ecosystem news for search marketers SEO education case study trend uh data for for strategies the things that we were doing those uh actions were really good and how to be trust. So this is the authority game you optimize your site for AI platforms doing all the stuff that I was uh sharing with you also applying Wikipedia. Yes, Wikipedia is relevant again after uh three decades when it was launched for the first time. All human knowledge information and of course the knowledge graph accuracy are really important. Again you need to have all that information about your company about yourself. If you want to be an entity recognized as an entity you need to have all of that in a good shape. Publish transparent pricing and expand FIQ. Well, there are companies uh that are not sharing their prices online. Uh uh uh LLMs, they love to see how transparent you are because they don't want to hallucinate about the question, what is the price of this service? And if you don't have those prices anywhere, they're going to be just hallucinating and they're going to be entering into a loop. So just analyzing how much you cost, how much your product cost, that's something that is really important today. Freshness and consistency. So this is um internal guide that we have internally in summarize. Uh this is our meta prompt for uh talking to the external sources right lead with an answer word mirror the syntax of the heading uh make it quotable uh chunk tightly use question based headings. So this is the metaprompt that we use in order to create uh more content and UCG content. We have dedicated FIQ pages which are great. We have industry-leading data studies that are telling you um that these are the companies that are implementing all these things that we were also implementing. Number one, Google of course. Second, Samsung, then Apple, then Microsoft, then Fidelity, then Chase, then Amazon. And if you can see the pattern here, we have huge tech companies that are implementing all these changes also. and also fintech companies, financial a huge banks that are doing that. You can see American Express is there, Capital One is there, Chase is there. So they are sharing the uh the the share of voice in terms of LLM is really big. So if you want to check your um AI visibility, how good you are in terms of AI, use this free tool that we have, the AI search visibility checker on Seamrush. uh you just need to put your domain there and we you will see how good different LLM platforms J or place it called you name it are perceiving you are perceiving your website and you can compare it also in seamrush one with the overall SEO if your SEO is in a good shape that's already a great opportunity that you have to increase your LLM to your AI visibility but maybe it's not your case so you need to check it you need to track it right Wikipedia page again very important to have everything um on date, everything fresh, all the information in an accurate way and these are the results that we are having. AI visibility changes that we implemented last year June last year right now are exploding just in a couple of months you see that just in a couple of months all the changes that we were make making had increased our AI visibility from 10% to about 35% and this year it's going even higher two months the efforts are very clear traditional SEO is of course good AI SEO emerge. That's also really cool because you can just adapt the whole stack of information of optimization that you have and implemented that for optimize your online visibility. Let me give you this um this um this structure this five stages of how you need to work your AI visibility. First research. So listen your customers the language across all your touch points. Then the architecture part you need to design navigation that mirrors customer mental models where they are searching how they are behaving they have more commercial information this is SEO my friends build the content align all the sources that you have all the different teams you have the um customer success educational team align that and that's the the reward of all of that is your AI visibility AI rewards death structure and authority. So basically one of the brands I'm I'm about to finalize this this part brand that is winning in AI search. We really love this um this brand is something that what this looks like in practice. Everything that we were implementing they are doing it as well. So you may recognize this brand right? Yes of course Salesforce. Salesforce they are doing the same thing. they are uh building their own their AI visibility in the context of their clients. Every exclusive client that they have they are working the context around it. Let's uh see just for an instance that they have their partner half and in their partner half they are building content for certain industries. Let's say the automotive industries. Yes, Salesforce is not they're not talking about uh automotive industries. Well, it's going to be very difficult to identify that. Here you can see a couple of agents, right? These small robots because these pages are meant to be found by agents. You can see automotive automotive leaders. You have see there's uh Ford, Mercedes, Bentley and other friends that they have, clients that they have and they're building these pages exclusively for them. Oh, a unique page for Mercedes-Benz from Salesfor um the customer experience is great. This is not just in the car, it's also in the CRM, right? Well, maybe you will find this type of of of pages, but these are meant to be for to be find for LLM platforms to build the context around it. And imagine that they are building also YouTube videos with their uh affiliates with their uh with their partners. How industries AI and automative cloud keep drives drivers safe for explain. They are also using all the method methodology that we were implementing. Our teams are also working on a single effort to give the visibility the context to the LLM platforms. Salesforce are doing that and voila for about 10 or more years they weren't uh publishing their prices. Right now we have the prices published and even from Carter automotive cloth automotive uh cloth agent forces can you believe this? Can you imagine this? Now for every single source for every single industry they have been publishing their uh prices because that's the thing that LLM platforms loves to gather to don't have any seen issues. So in terms of their eye visibility, they were using also uh the AI visibility checker from Samrush. The AI visibility is great. You can see here the total AI visibility chat GPT working great. AI overview is working great. Oh, AI mode is working even better. So yes, they can analyze that. Last three points, you can have AI visibility without organizational clarity. You need to work as a team. You can have organizational clarity without audience obsession. You need to love your audience. Not only your product, but your audience. And your site architecture proves whether you have it or not. That's everything for me. If you have any questions, I'm open to the questions and answer. Thank you very much, Fernando. That was incredible. Thank you so much. And I love your energy for this topic. We have loads of questions, which is really exciting. So, I'm going to bring them up on the screen as I go and hopefully we're going to remember what the context of all of them is. So, this question was from Caroline. It was much earlier in your presentation asking about ChachiBT. You mentioned they're losing some of their traffic, some of their visitors. They had a slight dip in the graph. And she asks, "Is that decline because they're losing out to other AI platforms like Claude? What do you think?" Well, this is the part where I can finally share my my opinion because the data is telling that they're declining, right? Uh they are not declining in terms of affiliate traffic in in terms of um sending traffic to other platforms but in the usage. Something big has happened with cloud this year. I'm also cloud user these days. Right? So if you are into this world yes we you can say that anthropic was doing a great work uh since January this year when they launched the cloud code then cloud code work and that was well impressive is that enough to uh attempt to chat GPT uh empire I don't know I believe yes but I cannot that's only my opinion right maybe that's that's uh the claw situation or maybe there is something else but the fact is that they're declining in traffic. I also think I read from Semrush just last week that there was a published chatbt traffic analysis that you guys did saying that chatbt the referral traffic is growing quite a lot. It was over 200% was the figure that I read. So even if they're having a slight decline here because of Claude because some people are using new models of Gemini as well which I think people have their favorites right they have their favorite tool what's really interesting I think is that referral traffic still being so large all right true and and that was the second graphic that I show right traffic the usage is uh going down but the referral traffic is just going up yeah all right next question is actually A double question from Andy and the first part you definitely answered already. Where does chat GPT rank relative to other AI tools in terms of usage and traffic? So I think you shared that on your fantastic pie charts worth catching in the recording if anyone didn't get it. And do you need different content strategies for different AI tools? Oh, that's a beautiful question and it's an important one because if you are um if you're willing to create uh traffic to be shown in an element platform that's useless, right? You need to understand where your users are first and then build content upon it. uh for example for some interesting uh cases in terms of industries I remember we have also uh that study uh published on on simrush uh for the enterprise category the most used uh LLM platform is copilot that's Microsoft you have all the Microsoft ecosystem and copilot is right there to be used for other for the more creative uh part uh let's say you are on Etsy or you are in the beauty e-commerce category that's something really interesting but it was perplexity there that is giving you more advanced research I don't know I'm not using perplexity but we have seen that there are a lot of create creatives that are using perplexity but for the rest by far is Gemini so Gemini AI overviews are the top of the of of the pyramid there you will see is Chad GPD. So just understand where your users are, what is their use case and you need to be there. Yeah, absolutely. I love this question. It's actually maybe my favorite which is from Caroline asking, "Does all of this mean that we don't need to manage our Google business profiles? Are they less important or will AI actually be surfacing content from them as well and analyzing them?" Nice question. Yeah, by far no. Your Google business profile is important more than ever. That's a source of information for LLM uh LLM platforms. If you are there, you have all the chances to become an entity. If you are already and all the reviews that you have from there are just exquisite food for LLM platforms. So if you have good reviews there, fantastic. If you don't, well, you need to work on on those part. uh all the details about your business, about your profile, about what you are doing, that's pure 100% information for LLM uh platforms, it's not going to be less important. It's more important right now. Even as I was mentioning uh having a Wikipedia page is important right now. Having the knowledge graph in a good shape is important now. So having all those sources in the right place is important. [snorts] Absolutely. and I see them come up on Google's AI mode all the time. They bring up business profiles all of the time, maps listings as well if it's a local business. So, they're really really huge for that. We've got another question which is uh around how we're actually optimizing which is all the extra coding, all the schema, is it worth it? Is it worth our time to do it? Nice question. And actually this this is also a wood one a good one. Why? Because you need to think that uh lls are robots, right? They understand if you uh talk different languages. Sometimes I speak uh Russian, I speak Czech, I speak Portuguese, I write, sorry, I type Portuguese, I speak Czech, I mix the languages, they understand me anyway. I make grammar mistakes, they understand me anyway. in terms of coding in terms of JSON LD or schema that's their language if you already have implemented that that's the first thing that they're going to be noticing because those are robots that's that's coding as as well they are working they I believe they talk better between agents in their own own own languages so uh if you already have that that's your great opportunity to be an entity so having your schema uh your um micro formats uh in a good tone already there established. That's a good um a good strategy. Yeah. And I have to say we review hundreds thousands of websites across the years and I don't think I've ever seen one with a perfect schema setup across the site because often what people don't realize is there's like hundred different types of schema. There's so many different types and setting it up correctly takes time and you want to make sure you're doing it in a way that's really clear. you're not confusing AI systems with incorrect schema or weird schema. And quite often people just haven't reviewed what they put on their site 10 years ago that might not be fully accurate as well. I see that all the time. Um, yeah, [laughter] absolutely. Okay, let's talk about e-commerce. Fernando, I'm being asked by Thea for e-commerce brands that are also across larger websites like Amazon and those get cited in AI quite a lot. How do we try and get our domains a share of that traffic and also get cited in those AI answers? Nice, beautiful question. In terms of strategy, when we're talking about uh e-commerses, um there is not a huge difference because at the end of the day, you need to think about the funnel, right? You need to think about from where your users are going to be um from which channels they're going to be founding who you are. They're going to be finding uh your services or your your your products. If you are an e-commerce brand, well, basically you you already have an established uh visibility. But if you are for example Amazon, you don't need to have an established visibility. You already have your whole ecosystem to have that from AI and that's the the good part. You need to be visible doing all the reverse engineering or on on what the LLM platforms are showing you right now. the things that I was telling you, you need to have your uh reviews in a good way. Product reviews are massively important. Uh YouTube videos or having these uh experience videos on these uh product platforms, they are also massively important. So it doesn't matter if you are um good at what you're doing, it's important to be visible. right now uh for example for this year mainly the important thing is that users are going to be knowing about you in on LM platforms because that's the discovery um source the discovery path that we have right now there we have we don't have any more clicks right now we have AI visibility from those from that discovery uh path that is starting from your product if you are an e-commerce brilliant answer we are so tight on time and so high on questions. So, I'm going to take us to a final question which is coming in from Julie. She says that she finds Reddit users and answers can be very US biased. How can we know the use of AI platforms by country and optimize accordingly? I think what Julie's getting at is do these Reddit threads get cited as much in the UK for example or Germany, different countries, or is it really big in US? How do how do we know what platforms we want to actually focus on? Well, in terms of uh local visibility, in terms of uh user conversations, right? Uh so the conversations are going on different uh platforms depending on the location that you are on the the country that that that you are. Reddit of course is massive for from from for English-speaking countries, but for uh let's say if you are in Australia, Reddit is also it's also massive. There you have Quora. Quora is also a fantastic source. But for example, in my case, I'm in Barcelona. Uh for us, well, Reddit is not that uh relevant. Uh there are some parts of Reddit that are really good in Spanish, but it's not massively adopted. For that, you will go definitely to social media. So any other social media, let's say Instagram, uh let's say even Tik Tok are massively adopted just to identify what users are saying about your product. But by far right now is the um is the Google [clears throat] is the my my my business is Google my my business. All the reviews that you have from there are localized already. So having these conversations um already established on a local language it's basically Google my business or it can be a Reddit or different review platforms. Uh here for example we have the G2 crowd. uh you have other crunch base even uh how people are commenting on your product or services those platforms need needs to be taken uh very seriously. Absolutely. Fernando, thank you so so much for your time today. I know that everyone will have super appreciated this webinar and if you have more questions for Fernando because there are loads of them, you can find him on LinkedIn. I've just put his LinkedIn on the screen so you can drop him a message and I'm sure he will be more than happy to answer questions. Fernando, is there anything else you would like to plug or anywhere else that people can find you if they want to follow your work and what you're doing? Sure. Well, basically, uh, you you can find me as a contributor for this for the Sebra blog. You can find me also fernando.com. It is my portfolio page. Uh, or you can just follow me on LinkedIn with Fernando. Thank you very much, Charlie. My pleasure. Thank you so so much for joining us, Fernando. Thank you so so much, everyone, for your fantastic questions. Uh we have really enjoyed this whole session and look forward to seeing you in the next one.