Surviving AI mode: What it means for the future of search

Okay. So, I think we are ready to begin. Well, welcome to today's webinar on surviving AI mode. See the number of you guys are now joining us in the chat. Welcome. Welcome. Um as you enter in, you can uh type in hi and where you guys are logging in from. Glad to see um who will be joining us here today from around the world. It's always good to see who are the different people that are joining us in today's webinar which is great. Yes, wonderful. Thank you everyone for joining us today. So we are already at four o'clock. Uh so uh I will be starting the webinar in just a few minutes here. Just making sure everything is ready and set up. Yes, streaming is live. So again, thank you everyone for joining us today. It's a contentious topic about as we draw to the end of 2025, we ponder about what will be happening for the future of search as it continues to evolve as the big AI giants start to bump heads against each other and continuously fight for their users attention and the market share. And AI mode is Google's answer basically to many of the AI assistants out there in terms of how uh people can use it to better answer their questions than traditional search. We will talk a little bit about that today and what it means for us as marketers. So a bit of housekeeping as you guys are joining here in today's webinar. Um if you have any questions please use the Q&A function inside Zoom. If you type it in chat, I may lo I may miss it. So, please um check the navigation at the bottom of your Zoom here. Uh it should say it should be under like a question mark icon or it could be under more. It you should say Q&A there. From there you can ask your questions. And for those people who are looking at the questions being asked, if you see a question that you definitely want to see the answer for, um you can upvote it and that will get prioritized when answering. Uh this webinar is I'm trying to keep it within an hour and if we hit the hour mark so we start a bit late so like we can go over for a few minutes but if we hit the hour mark and there's still questions unanswered I will follow up in uh a me email to all of you guys uh with a link to today's webinar replay as well as answers uh to the questions that are being asked. So your questions will be answered but we may not answer all of them uh in the live webinar today. Okay. And so, uh, without further ado, let's go ahead and begin today's webinar. First question is, how many of you guys have actually used AI mode before? You may have heard of it, but have you used it? How many of you are actually like, let's do like a raise of hands here very quickly. Yeah. Yeah. Yeah. So, now more and more people are using A mode. I think maybe sometimes you're forced to use AI mode as well when uh Google by default may switch you to AI mode instead of Yeah. So um a number of you but like not everyone right because it's not uh Google's really not forcing you uh immediately to switch over to AI mode but you can imagine and proportionately that's how many people will be doing that too uh when they're searching about your brand or trying to find answers online. So for those that have not seen AI mode and only heard of it, this is what AI mode looks like. The user will enter a query using text or they can upload an image or even use their voice and they'll get an answer um generated by um Gemini 2.5. And how they create this answer is that they'll first create fan out queries which is that they'll take your main prompt and ask like related questions to it and use and and use that to search uh oh thank you for letting me know it's not happening. Okay, one second. There you go. Everyone can see the screen now. Thank you so much. It's not showing up for some reason. That was weird. Okay, so this is what AO mode looks like, right? You type in a query using text or image or you can use your voice and you can and AI mode in this case gen Gemini 2.5 the latest algorithm or I think now it's three sorry it's currently three um will be um creating fanout queries which is related questions to the prompt that was entered in and they'll use those related questions to search on the web. So, for example, um if you have a very complex prompt and you type it into AI mode, you can ask something like gift ideas from Osaka under $50 for people who already have everything. People that are hard to like give gifts for. Press enter. So, what will Google do or Gemini do? They will create fan art queries that look like this. Uh Osaka artisan gifts under $50. Uh food products that are budget friendly. experience gifts, Osaka that are affordable, hard to shop for people gift ideas. These are just example of like um fan out queries that get searched on the web and all of these will come up with like uh source answers and AI will synthesize all these different results on the web and generate a textbased answer with citations and sometimes as you can see in the case of this uh answer in this video here they will create a table or they can create um they will list a bunch of different uh shopping results or they will create like a board of different pictures. So the format of which AI mode generates the response also depends on the question that that's being asked. Is AI mode just a longer version of AI overviews? Right? Some people are saying like is this just like if I have to optimize for how AI is responding about my brand? Can I just assume that AI mode is just a longer version of AI overviews and I only have to worry about what AI overviews is saying? Well, here's what we see when we look at the data. We see that if by analyzing uh the the responses by AI mode versus AI overviews, they generate very different looking responses. On average, AI mode's content length is four times time four times that of uh AI overviews and only 0.51% of the time AI mode and AI overview output the same response. So by and large the responses in terms of just content length in terms of uh like the words actually being used they are completely different and only 2.51% of the first 50 letters of AI mode and AI overview are the same and not only are the text results not the same the URLs being cited are also not the same there is only an overlap of around 13% um by the two modes And 3% of the time there are no URLs cited at all in the AI mode generated answer. In AI overviews, this is slightly higher of 11%. However, despite having different text content uh different text length and text words and using different URLs, semantically the content is very similar. So there is this um uh data science metric that you can use which is the average conign similarity score. And what happens is that you can basically plot a vector uh using uh on the words that are being used between one chunk of text versus another chunk of text and they they vector that based on like what is associated meaning of each word that's being used. So then you can average that out when you do that for every word in like a paragraph. You can average that out and compare that to uh a comparison like uh block of text A and block of text B and you can see directionally if they're similar or not. If they are similar then there'll be a cosine similarity score of one. So basically they'll be pointing if you like you can envision this in a vectorzed space they'll be pointing in a similar direction and that's how you sort of uh like analyze how similar content is across like large data sets especially with text. So we see the score to be 86. So that's very very similar in terms of like so and when they answer questions the kind of um answers they give in meaning wise is very similar and the pages that AI mode and AI overviews tend to site are also quite similar. So by and large the majority of um sources that they site with is articles and then followed by core pages and a toss up between listing and videos and like the rest of this is just a very small percentage but by and large you can see that mostly articles are being used as citations for uh creating answers uh generated for when people ask them either in AI mode or AI overviews. And interestingly, many of the site pages in AI mode don't actually show up in in like traditional SERs. So they may not be ranking for any organic keywords at all. Our core pages are things like uh homepages, landing pages, uh like pages of a like the like the the sort of basic pages of a website. Yeah. So, as opposed to like a blog article, there would be more like a homepage or a page that's describing a business, things like that. I hope that answers your question, Sally. But yeah, so going back to the presentation, um, when we analyzed like 6,300 AI mode responses, it cited over 60,000 URLs and we see that many of these URLs really don't show up in any positions at all in in the organic keywords that we have in our billions of keywords database. even though they don't perform like in Google in traditional SEO, they may still be used as answers uh in uh AIO and AI mode. So, and if you want to see in uh brand radar, you can see at the bottom here that a lot of um different different sources are being used. And in this case, you can see that this is um uh AI mode response in uh Japan. And here it consisted actually a lot of different forum posts and not just uh landing pages, not just blogs. It was very interesting. Um other factors like um DR still seem to influence citations as well because we looked at um the chat GPD citation frequency. uh by domain rating and based on the top thousand sites we see that many of them still are in the higher domain rating um still come from higher domain rating websites. So even though the pages themselves may not rank organically, they still may come from uh websites that have traditional domain authority and even though this is chat GBT, we know that um chat GBT uh scrapes Google when they create uh generate responses and we are also looking at things like AI mode and AI over overviews that do reference Google's giant web index. So like we can see that in in general these traditional SEO metrics will still be important when we want to improve visibility about our brand or just visibility in general in AI responses. So as a summary, AI mode generates answers differently text-wise. They source pages differently, text uh source page differently compared to AIO's, but then semantically the answers are similar. And pages that get cited may not need to rank organically, but they may still need to come most likely need to come from websites of high domain authority. How does AI mode as a mode affect search behavior? So, um there is this one um in-depth AI mode usability study that was conducted by Kevin Indig. Kevin Indig is one of the speakers that recently presented uh at HF's Evolve. If you have not if you did not attend, but you can check out uh his newsletter. And in that um newsletter one of his in-depth AI mode usability studies they looked at uh had a handful about 30 people and they gave them specific tasks to achieve and they said you have to use AI mode to achieve those tasks and they they basically studied to see what they looked at what they clicked how much time did they spend on different uh parts of their task and that helps us understand that in depth how a person would be reacting uh or making decisions using the information presented to them by uh AI mode and they saw that users basically hardly left AI mode to achieve these tasks. These were tasks like um research a uh like best product or try to answer a question and things like that. Um and most of the time people are not leaving AI mode. They they do not click out to other websites. um and they spend more than double their time in AI mode as opposed to um spending time reading AI overviews. And the like one kind of damning fact is that the median number of external clicks per task was zero which means most of the time people were not it's not like on average people clicking one or or like um clipping clicking more and and then overall it was less. It was just people just did not click at all. and the very few clicks that were made were basically already going out to pages to make a purchase. So by and large it implies that the content that is being served by AI mode is very very useful and people are very satisfied with their um with their like prompt or their task by AI mode and they do not see the need to go into different websites. This is basically uh something that we have to keep in mind as we uh develop our websites and information about our brand moving forward. So what does the future of search look like as a result where people will stop clicking less? I mean they'll stop clicking and they're using all these AI assistants to answer their questions and get things done. Well, right now Google is still king because even though people are using AI and I'm sure almost all of you guys will be using AI in some shape or form in terms of the traffic distribution across all the categories of sources uh that come to different websites, we see that search is still at the top. Organic traffic is still the the biggest source followed by direct. And um if you're wondering how we we have this data, we actually have this website here which you can check out where we analyze what are the total traffic sources that are coming to uh websites as a whole and we monitor them on a month-to-month basis to see how things are trending over time because we're also seeing how want to see how the landscape is shaping up and but however even though Google is still king traffic from GBT is steadily growing. So, we're seeing month on month that Google's um traffic um share is going down ever so slightly and chat GBT's um traffic is going up every month in some percentage. Um but like if I go back and I show you that the since there's so much more search than the um organic search compared to uh AI search this increase in percentage does not translate to like lots of traffic every month. It's still a very small percentage but you can see that overall there is a change in um market share and it'll be interesting to see how this continues to evolve. And of course like as you know every month every one of these AI assistant providers are not uh laying down. They are very aggressively trying to do tie-ups, optimizing the uh search experience, optimizing the algorithms, doing uh tie-ups with local brands to basically try to capture more and more of the search market of the attention uh market of people so that any decisions or any knowledge or basically usage of their platform can be done and be more uh exciting on their platform versus other competitors. and and Google um themselves actually really want to see AI mode replace traditional search for good. Uh Sundar Pichai himself has said on uh the Lexen Freeman podcast that he wants like the current goal is for this to basically become um the main experience moving forward for basically the rest of the world. And we will maybe like most likely say goodbye soon to our familiar 10 uh blue links. And in fact, they are really serious about this um new change because they are starting to see how can we monetize this this platform? Will people continue using it if we start um pushing ads into the results? Um this image is taken from um Brody Clark who on on LinkedIn who recently posted to see hey they're actually starting to run ads now uh in AI mode. So you can be looking out to see whether or not that is showing up for you as well because they like eventually Google has to make sure that their main business model still works in a new mode. So we shall see uh if Google continues with this approach and overall we know that clicks will continue to go down because we saw earlier that people as people get more satisfied with the AI answers they have less reason to click. However, that doesn't mean that people are not still searching. Um like these responses may lead to um direct search into um the website of your brand itself. um they may then do further research on other associate websites um by typing it manually or seeing it through other social media posts and stuff like that like the mention or the visibility of your brand then sticks in your brain that does not get captured by the typical clicks that you get in uh traditional search. So here we're showing that and typically if you sign up for HFS I think for many of you guys when you sign up for the first time you would have seen us ask this question where we ask you where did you learn about hrefs and you can see that a lot of um especially when chat GPD became first became more popular people were still getting um people were actually responding to say yeah I learned about HFS and chat GPT or they learned from us from Gemini or they learned from us from uh learned about us from Perplexity. So, so this is like a layer of um attribution of your success that may be useful for you moving forward because sometimes you just have to ask these people because it's going to get a lot harder to track where people are learning, understanding about our brand through the traditional means. Um Joanna, if possible, could you share your question in the Q&A? Um I I'll try to remember that you've asked a question, but it might get lost in chat. Thank you. And so as a result, we can't be relying entirely just on organic traffic. I think many people have um are sharing about how um because of clicks are going down, that metric is not going to be as useful anymore. And we have to look at a number of other metrics um to better understand what is working and whether or not performance of your marketing efforts is is working or not. So things like share of voice and share of search are going to be important. How much um like share a search is basically how much of uh the search has your brand name in it. Uh what percentage voices things like what percentage of the results mention your brand as they get um as people like search into these different platforms versus your competitors or versus no brands being mentioned at all. um thing traditional sales metrics become important things like customer lifetime value and like u number of new sales. Uh this is just something that you as a marketer now we have to pay more attention to just because um we have to understand in the end these are the core metrics of success that everyone has to be aligned to. Uh and then things like social mentions will be important and direct traffic. Why? because uh if I learn about a new brand uh or I learn about I think a new feature that's being done by a certain platform or something I might just type the website straight into um the search bar you know I will bypass all these like clicking through links now and that is a common um common behavior that we see moving forward with a lot of people when as as a respond of use as a response of using all these AI platforms So the next is what should you do to be more discoverable in AI and search. So I'll give just an overview of like things that you can do today but there are so many things that you can do if we go deeper into like how you should build off page man off page mansions. There are a lot of ways to do that uh and different angles that you can take but I will go through just like the the key points that you can do here. So the first is to audit your brand's share of voice. You have to understand basically where your brand currently sits now uh compared to your competitors. And the first and of course I have to talk about our cool uh core tool of doing this which is brand radar which allows you to see um mentions of your brand across all the different AI platforms and now most recently YouTube as well. Um and from there you can see over time whether or not uh it is growing and you can filter for things like uh um different industries or different niches. You can also filter to see what are the top ranking uh tops cited um domains is and pages to see in your niche which are the places that people are or which are the places that AI is more likely to take to generate their answers from. And of course if your if your brand is not there or it's being mentioned in a context that is not favorable you have to look into it and see how that can be improved. Uh so for example, if I want to find AI responses that mention competitors but not me, um I can click onto AI responses, I can filter for my brand versus my competitors here and I can optionally put in a market or niche and I can click on AI responses and there I can stack on different filters to specifically filter for u mentions that of of my like of my competitors but not my brand. then I can see what they look like. Um, optimizing your content structure. So, there's a lot of talk about chunking your answers, whether you should be taking uh your articles and being like basically changing its format um so that it becomes more uh address like more easier for AI to digest and use as a as a source. Um, typically LLMs love structure and clarity. So, like as opposed to having just a page of 2,000 words, um having it structured into your your your headings and your sub uh your sections, the main key takeaways at the top and making sure that uh things like the author and and um is is stated information about the author, uh publish dates, all these things are are important, but it's also nice for human readers. So it's not just optimizing for AI like turns out that if you optimize and structure content in a you know a clear manner that like your human readers will also benefit as well. So you can see example on HF's blog that's also how we go about structuring our content where we start with like the the key takeaways first and then we um break it down into different sections trying to answer the the main topic of each article, right? and a guide of how you can go about that is in this QR code. So if you'd like to learn more, you can definitely scan it and and learn more. Uh refresh your old/outdated content. So by and large um freshness or content we we see is a like sure indicator of whether or not um your these pages are being going to be cited by AI or not. So like things like reviving your top pages and making sure that the content in your pages that you have not looked at for like many years for example still contain accurate information or if they can be improved you can go about uh updating those. You can basically find the older pages that were previously not answering questions just as well and add sections to that. So that's like filling topic gaps. Um and then as you go about this um process of refreshing content, they then become opportunities to reshare them on social media uh with the new changes that are made. A full guide can be uh checked out by uh here if you scan a QR code. This is Louis's my colleagues Louise's uh blog about how you can refresh content. It goes through from analyzing how when to decide which pages should be prioritized to how you should go about optimizing uh refreshing your content. Some needs more effort, some needs less less effort and she'll shared a guide about how you can do that as well. Build off- page mentions. So, it's not just about optimizing your website alone. It it really really matters when other websites are talking about your brand and how they're talking about your brand. And we find that in our study of over 75,000 brands that off off page mentions like off- website mentions is the key is the strongest factor to determining whether your brand gets mentioned or not in a AI generated response. So this is of course um like not just Reddit or or AOR but also uh opinion or based websites that like in your industry that are important. They could be PE influencers that are running your blogs. Those are still important. PR that's why PR is becoming so prominent these days. uh where all these um mentions in in articles coming based on high authority websites will matter more and more. And you can test this yourself even if you don't have brand radar by entering a prompt that is really like a typical question people might ask in your industry or even just a question of like what is this brand about or what is your brand about? you will see that many of the sources that are being um that are being used in the AI answer is not your website. So it's very important to that you start building off page mentions uh to get more visibility of your brand and that's because brand discovery in general aside from AI but also for like normal human beings that are learning about brands they are now multimodal they are people that are listening to podcasts they are people who are um like on Instagram scrolling through looking at pictures they are people who are like following down long threads in the forums and there are people who are watching videos, short videos, long videos. These are the like these are the platforms that um people are learning about different brands or best practices or ways to solve their problems that um that will be important for you to consider. So, it's not necessary that you have to be on every platform, but you have to look into um which are the platforms that your core audience hangs out and see how that you can at least set up um your your brand's presence like properly there like like anything from the just doing the basics of setting up a profile and making sure it's authentic that links back to your website to to um answering questions and engaging your your audience in authentic discussions. And of course, it's not just uh like people talking about your brand with a branded channel. It's also um industry influencers and experts that are that people follow that will be talking about different brands. And like in fact this ties back all the way to the very beginning of how HF started where when HFS was still a beginner brand. Um Tim basically emailed out to all the industry uh influences about SEO tools and he asked them hey you know like we have this tool uh at HRS can you please try it out and let us know how it compares to the tools that you guys were already doing which at the time were like the biggest players were majestic in mods and that's like that's the the beginning of how relationships can be started and how these people who kind of um have a play in controlling the narrative in which what brands are good or how they're good. Um this is this is these are the ways in which you have to start u making sure that you are visible to them so that people outside of you are talking about your brand. Uh and if you want to look into how which pages are being cited uh when generating AI responses, you can do that using Brand Radar, of course. again um in this section about cited domains and cited pages where you type in your brand and your competitors and then you can filter for either a brand or your competitors or I can or for example I like to put in like YouTube or Reddit and and many of these like big websites to see like at what percentage of the time are these domains being cited and like what kind of responses are being generated that are taking uh responses from these big uh from these big websites. So these are things that you can also do with brand radar um and do outstanding things. So the reason why I mentioned this is because it's not just about the like the same routines that everyone is doing. I think like today now in marketing if you're doing the same thing as everyone else you're kind of not going to stand out. An example um here that we have um shared before in our blog about Ryan Traan's 50 states challenge is um where this particular company electric ebikes saw that Ryan Trraan a like big influencer was doing a um traveling across all 50 states to raise money for the St. Jude's Chris uh children's research hospital. He started with 1 million goal and then he eventually raised like $11 million and every one of his videos were pulling in millions of viewers. So this brand electric ebikes came in donated $100,000 and then showed up in person in one of like as he was traveling across the US uh and said here we'll give you an electric ebike and every day that you ride with this bike they will donate another $10,000. So, like they could have just paid for an ad on YouTube or something while um in in like Ryan Shrehan's niche, but they decided to do this like very um like classy approach to to giving like donating to his cause and saying like they they they share in it, they believe in it. So, it's less like an ad and more of a like generous genuine like like act of kindness that they're doing. And of course and now like because of this then Ryan started to ride his bike in every one of the videos. So now every video video has electric ebike like showing up um which is getting millions of views and of course and and because of that then people have a reason to talk about how you know electric ebike in a way that's not just about oh should I try this bike or that bike. they go, "Oh, this is amazing. Um, I bought electric because electric bike because of Ryan, you know, like I I have so much respect for the brand." These are the gold standard of off-brand mentions that you you want to get and just because of um their approach, they managed to get that and not paying people to do it, but that people were so inspired by this gesture that they will now associate lecture bikes with this gesture and the overall impression of the brand. And this information of course gets fed into um all the search um all the AI assistants when they do their um like searches and in their web indexes to find answers they will like parse this data and they will train it in their algorithms like that's how that like it is like a win-winwin on different levels and of course like after after that happened then mentions did indeed increase over time um for this brand uh across AR overview uh AI overviews as well as uh other AI platforms. So like you know it works and we can definitely see that it has worked. Um, sorry, this is the Japanese version, but I will share a link later on about an 82point checklist for improving AI visibility where if you really are stuck and you're not sure where to start, um, you can check out um, my colleagues blog on this. Just one second, I'll get it to you guys. Um, 82 point. So I'll share it in the chat here. So if you are stuck and you don't know what to do to improve your AI visibility, you can check out this checklist and start there. I'm sure you'll find um an idea that is worth pursuing to improve your brand visibility uh across these platforms. But is there really nothing extra than just SEO to improve AI visibility? I you know that that is a common question that people um typically ask to say like is this just SEO again you're saying that SEO is just enough um to optimize um visibility for uh all these new platforms that are doing different things. Uh, one second, guys. I just shared the link. One. Okay. Can you see it now? Yes, I think people can see it now. Awesome. I will also share this link uh in the follow-up email that that you get after this webinar so everyone can access to it. It's just a great place to see all the different things that you can do to analyze your AI visibility as well as do things to improve your brand visibility like off- page, on page, things like that. Okay, definitely do check it out. There was a lot of work put into this blog. It's really good. But coming back to the presentation, is there really nothing extra than just SEO? Google says so. they they plan to say that hey just don't need to do AEO or go uh to rank in AI overviews and um in as well as AI mode. Um they say that traditional SEO is all you need. Um and of course it's more than just to take um Google what Google says at at face value. you we can also see uh at least at hrefs when uh chat GBT first came out and all the mentions about our brand were showing up in AI responses we ourselves didn't do anything different uh to optimize for AI responses we just stuck to um like traditional um good SEO for for many many years and to show up in these answers so it may seem that that is the case but there's a lot of talk out there to about how you can do little tricks or AI hacks to basically find a way to be mentioned more in in AI responses. Uh and actually in this time it is kind of easy to do it. So again I will share a blog um from the same colleague of mine Despina that talks about SEO being in it um a AEO being in his black hat era. This is an example of uh a security company which is focused on transparency and privacy and AI. They actually ran a test to prove how easily a model could be corrupted by just poisoning the data with um I mean fake mentions of a particular brand or or like facts that were not true and and when that is pushed into um um that was trained on you can see that they constant that the AI and LLM was confidently repeating fake news and you know like inaccurate responses because of how the training set was affected. The truth is like LLMs are not particularly good in understanding what is you know more reliably true or what is actually wrong. Just the more um like more of something will affect its response they'll think that it is uh they it would affect the way it generates answers. That's the reason why all these AI companies are so focused in um in the more traditional metrics such as like backlinks and and uh like domain authority and trying to make their own search engine and trying to build their own browser because like they're still relying on these factors to which is you know inadvertently human factors to understand whether something is valuable or not. The LLMs are not very good. So if you want to manipulate it, you kind of um and then I will not talk about it here but you there there are people doing it and they're kind of getting success from it. What does that like? I like I can give some examples here. Um like people posting many many many best off lists like you know best way to do things, cheap ways of doing things and then just putting their brand on top. They don't have to do any sort of uh research or or actually use the brands or use the products. um they just keep creating these articles and LLMs keep um searching those and like regurgitating them as answers. Um you can hide frequent mentions of certain entities or co-mentions like your brand versus like great product or or like it's fast and easy and just and hide that in the background of text or even images. I've seen that as well where like they will hide these prompts inside images or or website code and that gets you know picked up by LLMs and goes into the way that they generate answers. Uh chunking or creating Q&As with entity rich content. So like a lot of um people who are trying to manipulate uh responses will just create long long long lists of of um Q&As's where basically they try to match as closely to what they think of fan art queries of anything that's related to a brand to try as much as possible to get their brand mentioned more. And then like when some people are trying on the reverse to just spam LLMs with prompts that are like questions that people will not normally ask about your brand or like whether it's good or what do you think about Xbrand? How do I compare Xbrand against other things and then using that to basically um go back into the training of these algorithms. So like these are ways that you can artificially influence the way things are being mentioned. But um here's the problems with these gimmicks which are similar to the problems of black hat gimmicks which is that um a the algorithms like both in search as well as LLMs they can change like anytime overnight even right and and the way in which um algorithms change the way that they respond is massive in LLMs you can like recently with the way that uh open AAI recently changed the way chat GPD was was responding from like really detailed long prompt answers to very short answers they can just do that to an algorithm and then all your efforts to optimize for a specific way which the um AI is answering can basically be gone. Um the second thing is that all these AI platforms still want to provide valuable and verifiable uh answers to their users because that's what users want. So if they find that the quality of responses are going down because of all these tactics, they will penalize and they will um put out rules to to fight against this. They will not just take this because it ultimately affects the experience of like normal people. So um just like black hack black hat practices, right? And then uh finally even if your brand gets more mentioned or they get um you know more visibility in all these places because of your efforts um if these efforts are uh of like trying to manipulate um answers artificially get known to your customers or they get basically get reviewed by people that doesn't like correlate with a viable trustworthy brand. This is something that is going to work in the short run but will not do any do you any good to build a long-term good quality brand experience later. So this is what you're comparing against like are you like chasing very quick results now versus how it's going to impact the way people see you like five years from now. These are the decisions that you have to take when you look when you see the next guru or the next uh like exciting post that talks about quick hacks you can get to improve your AI visibility and ultimately you're selling to people. Bots are not going to buy your product. Bots are not going to engage or subscribe. They're not going to, you know, follow you. Like I mean they could follow you but they're not gonna talk about you in an excited way to other people or and and and get like grow an audience in a in a where a meaningful way. You're selling to people and then that you have to remember all the time and we may forget that sometimes as marketers. So at the very end of it it's someone that's making a purchase decision or someone that is making a decision um like making that uh after all their research analysis to pick up that phone and call you. So, is your representation on these AI responses and your website and these different, you know, social media posts or whatever, are they consistently genuinely trying to help people? Are they trustworthy that people should trust um that people should, you know, decide to go with you? Because that would reflect humans as well as LLMs both. And um I I would just I mentioned this already earlier, but like we were cited fre we're still cited frequently by AI and we're not doing anything different than just good fundamental SEO and provide creating content that is meaningful and useful for for you guys as well as um the rest of the digital marketers out there. like this all takes like a lot of effort to create a lot of research and like we try to optimize for good quality content and writing as well as creating um good experiences in our videos and stuff like that. So these are the things that are going to stick more um than just the quick hacks and I mean it does translate to these other uh platforms visibility on like AI as well as normal search as well. So strong brand fundamentals will win in the long run. This is creating a good product, building topical authority by um you know doing research or doing the hard work and then sharing about it, addressing the pain point of your customers, chasing excellence and making sure your messaging is consistent across platforms. Make sure they reflect your values. If it if we go back to the previous slide about what electric bikes was doing, they did this uh donation thing because they actually have a philanthropy uh part of their business. So it's on brand. So are you doing things that are on brand to what you are um that that people should associate you with, right? So um that so that's what I want to say about how we can prepare ourselves for the future of search. we can't do like it's going to be less of the shortcuts and it's more about the the hard work that matters to help us stand out. Thank you. And now I will help to answer the questions that you guys have shared. So I will go ahead and uh look at the Q&A. Uh and I know some people asked in the chat. So but first I'll have a look at the Q&A and see there are nine questions. Okay. Chris asks um sorry if this has been said again said already but will you share a copy of this presentation? Yeah uh this time I will I will share the slides some of these parts like if there uh maybe the links that I can turn them into Q like QR codes I can also add the link as well. So I'll do a few changes to make sure they're easy to reference uh when when you guys get a copy of these slidesh and then if there were things that were covered in this Q&A that will be useful I'll share that as well. Uh, yep. And Ry asks, "How can uh information/advertising revenue website survive this?" Uh, it's going to be hard. So if your if your business relies specifically on clicks like you have display ads and you need that click to generate revenue, you don't sell products, you don't sell services, you don't sell a uh like a paid subscription or something like that. Uh yes, your business these businesses get affected the most. Uh I I've seen I think many websites in the educational category. So like they they answer questions like you know how do you do this in Python and basically the page only like the the value is that people click on it and they'll like get their answers in Python and then they'll go away and they're running ads in the background. These are the businesses that get hit the most because they have sort of no other way to um attract revenue stream and also a lot of times it may not translate to brand. So they are not that people will not remember them as a brand. they'll like remember, oh, I read this cool thing or like I watched this cool thing. Um, so a lot of these websites actually have started to pivot. So we know some that basically uh took the website entirely off of like AI visibility and like started to do like a paid subscription instead. So different different um business models may need to be considered um depending on it really depends on your business but there are different ways. Uh also uh you may have to look into how your website is more useful than reading uh articles at and on our website we have a lot of free tools for example and no matter how much AI mode and AI overviews can take content from us if you want to access um and do something useful with these free tools you have to click on our website. So that's like a way that we still keep some of the traffic right and we sell a subscription of our platform. So eventually even as people understand our brand, they know about our brand, they know what we do or like the the great content that we reach out, if they're interested, they still come to our website to buy a subscription. So these are the ways in which um businesses that are like traditionally not caring about these things will then start have to caring to care to care about them. Uh Nadir asks, "Would established domain authority be the most weighted metric coming into AI search? Is there a change in content that could prove effective as well?" Okay, so I'm going to go back to the previous slide that I had where we did analysis. This one is specific to chat GPT only but um we see that happening in general maybe not the same percentage but also on other uh AI platforms which is these are like the factors that we see affecting uh brand appearance most in a overview. So it's from branded web mentions to branded anchors to search volume then domain rating then number of referring domains branded traffic. So um we we are constantly looking at the data uh to try and figure out what is going to be meaningful or useful uh as we collect different AI responses from different platforms to to sort of correlate how how well brands perform and outperform over time. Um but we see that it's not going to be the most important thing. It's also like domain authority. It's also not the main thing that you need to focus on even in like normal SEO. [clears throat] Uh you know different factors always come into play. The quality of your content still matters. What people are saying about your brand on those pages also matter. So, like you can get mentioned in a PR, you know, on a on a high authority website, but if they're just mentioning things like, oh, your company has gone through uh a new change recently, or they're opening up an office in a new country, does that translate into a context of, oh, you should use this brand to solve this problem? No, it doesn't. So, like, you know, focusing on domain authority um alone does not help necessarily. And uh I will again go back to like I shared the link earlier about um this SEO AI search checklist. So definitely check into that to see um different tips and tricks. I'll share it again just for for easy accessibility where you can see tips and tricks on how whether or not changes to your content will make an impact. Sometimes it's not worth the effort, but there could be old pages out there that are saying outdated things about your your brand. They could be updated to get more um helpful information that that is something that you may be important to you and therefore require you to change your content. So, uh okay, Chris also shared another question. Um, is I'm going to answer live. Is an off- page mention a backlink or do you mean literally that they're talking about you? I mean literally the page the page's text like the main body of text mentions your brand. It doesn't necessarily have to be a link actually. Now these days it's not necessary that you have that you need a backlink that they necessarily have to link back to your website. I mean, it's useful if that uh article or that blog or that page is um both mentioning you and linking back to you, but it's not necessary. And for example, a lot of uh like Reddit and Quora and many of these other forums, they would usually ban links in their posts. So, like they can't link back to you anyways for many of these platforms. or like YouTube like like in a video they could be mentioning a a brand and there's sort of uh it's not necessary that they will link back to you but these are all mentions of your brand in different places. So yeah, I mean um on these days just think about like if we take a step back and like just try not to look just in the data. Just think about how people discover a brand. Anyways, just think about how you've recently learned about the last five brands of like uh laptops or like you had to buy a phone or you had to buy a car or something like some you know decision purchase decision that you have to make. Did you only read an article to make that decision or did you watch Tik Tok? Did you, you know, see an ad? Did you read up some like guy's review uh on on Reddit or, you know, other platforms that you may use? You know, these are these are now ways and then if you did, then these are places you need to start thinking about. Okay, how do I um sort of try to find ways to be spoken about on these platforms or be mentioned about? It may mean that you make videos. It may mean you work with an influencer. That's that's, you know, things that you have to work out. Chris says, "I just close my eyes and choose the first brand." That's uh not necessarily the best way to choose a good product, but it could work. If it works for you, man, that's that's pretty good. Uh Darkos asks, "I have a few good blog posts for my business that also serve as landing pages with forms. I'm afraid if they change them or cluster them for AI, I could lose rankings. Do you have any experience with this at HFS?" Uh, so this is similar to uh the question of like should I refresh my old outdated content? I know it's not like old and outdated, but like maybe it's a page you probably have not touched in a while. Like it's been working uh and it's doing a job well and like maybe it's not performing as well. That still kind of counts. um in in my colleagueu's blog here I think she has a section that basically you need to validate like whether you need to change it because you may change the page and it became it becomes worse. So um first things I typically look like if I want to update a page is a is it still useful for the reader like reading it that's the first thing and then the second thing is that um like if this page is no longer ranking well uh or like it's dropping in rankings in traditional search just traditional search first I go and look and see which pages are being cited instead like which pages are ranking higher or which pages are cited more to answer the same question that your page is answering and I see what like how those pages look like how are they better than mine and and try to see what improvements I can make on my page as a result and you go with that approach and then it's possible even after that when you make the change that that it may not have like a a positive impact but you use these benchmarks to see how you can improve the page rather than just analyze just optimizing it for so the AI picks it up more it's usually not that use useful to spend that time doing it. Okay, so I hope that answers the question. Darko, uh Andreas asked, I think you've come in our uh webinars before, so welcome back. Uh backlinks are becoming increasingly important again. Yeah. Uh I don't think backlinks was ever unimportant. I think the quality of your back links probably matter more than quantity. Uh and then mentions is also like quantity and quality matters. So mentions number of mentions on relevant websites will matter. Um but like you know it is not it was never like backlinks was never unimportant. I think it was always something that is worth sort of uh building as long as people are blogging, as long as uh people in the industry like fellow like websites in industry are talking about uh your brand or something like that it there should always be um they would always be reasons to link about something. So um you know if you if your strategy as a company is just a okay I got to like stop SEO and just focus entirely on AIO or or geo first thing is you know like organic search is still the biggest source of traffic today still so like you should not like stop it entirely and it's it's not a like like a like AI or search discussion most of the things that you can do to if you if you create great content for example, people will like organically link to you, you know, like if if you make stuff that is useful for people, people will talk about it. So yeah. Yeah. Hope that answers and thanks Andreas. I guess that you found that to be useful. Uh yeah, don't please don't guys, please don't like drop your SEO efforts and and stuff or like I know it can be for example a hard discussion with maybe your manager or something if it if it comes to say our traffic is down what do we do? uh as now in the SEO space I think it is the responsibility of every marketer to to really um have this discussion to educate uh people in your company or in your team about how why things are changing and how they're changing rather than sort of like um fall back and say oh you know SEO is not important anymore. It's important to have that discussion so that people can all sort of move forward in in in a better way and create like more useful Uh, okay. We are getting more questions than we have time. So, here's what I'll do. Um, I I saw that there were still some questions in the chat. So, please make sure that your questions are are in the Q&A because like that's how I'm going to take it back to um answer the questions after this webinar ends. I will answer uh three more. How about that? And I sorted by most voted. So I answer three more questions and then uh after that we will take the rest of the questions offline. I hope that is okay with everyone. So I will start with Josh's question. Um actually I think this was related to what we had earlier which is about backlinks. Uh so so again I will go back to this slide where it's specifically what the factor that we see is mentions. So I mean backlink is a mention. It's a it's your it's your brand and then hyperl to your website but it's also when your brand is being spoken about but it's not linked. Okay. So and is Google trying to filter the truth from the lies? They are trying. They've been trying to do it. I think all these AI platforms are trying to do this because uh generally authentic genuine content results in better responses that everybody wants. So they are trying their best to make sure that they can always um make this happen in in the responses that they make. It's not easy. Uh Yogesh asked the question uh we're using a site map sitemap file and XML. Do we have do we have anything for geo? Um I think this sounds like this llm.txt txt file that is um being mentioned that you need to optimize for we don't see any evidence that it matters whatever document is used um that to to like you know are these AI assistants using that um document we currently don't see it like that is impacting a uh having that big of an impact we see more of uh like the traditional sitemap or robots.txt txt, you know, whether or not they're allowed to crawl your website or not. So, that is a bit more prevalent. So, we don't see llm.txt like optimizing it or adding information about it being necessarily worth the effort at all because like the evidence shows that they're not using it right now. all of this that we talk about today like you got to watch the trends of how all these like OpenAI and Entropic and and Microsoft they want to make changes that can be really big in the next like few months or so and so my answers may not be valid if they may choose to use lms.txt txt currently they don't not that we see so but in the future they might so like you you can watch out but for now I think it is so much more worth of your time and effort to focus on the parts that we know that are hard but they matter which is to create good content and structure your website well and all the like the boring traditional SEO stuff that makes your brand and your website great that that is more worth of your effort right now because we know those are important Uh okay, Rajat Mahultra, do we need to focus on traditional SEO? Should we put more effort into AI visibility? So go back to the original thing which is that currently it um traditional SEO U factors help a lot. Um there is more focus on off-brand off- page mentions. Um there is um like so basically more more of those are important into SEO rather than focusing entirely on just backlinks or more monitoring traffic and things like that. Uh like you know structuring your pages those are also kind of uh similar to traditional SEO like this one that we talk about optimizing content structure. This was useful to SEO anyways. So a lot of times there's there's quite a bit of overlap. Uh but like the the building of like different mentions um off-site in social media in in forums and stuff like that that kind of blends into other parts of marketing as well and not necessarily SEO that would benefit AI visibility um uh in these ways. So like yes still traditional SEO but um AI visibility is slightly more than that and focusing on both will help your brand no matter what. Um, okay. I'm going to answer this specific question by Idle Fishman. That's the last one here where where they're asking about this website that we that we're supporting right now. This one, which is our Google chat GPT versus google.com uh like uh website where we're showing Google traffic share decreasing. And then there are other reports that show Google usage almost doubling. Um we are using the data sources that we have. I think um if if maybe you want to share that study with us or that that website with us because I'm not sure which one you're referring to. Um likely they're they're sharing getting data or they're using data about Google in a different way than us. And we can see the reasons why they're different. I can't really tell you that right now. Lots of um websites and lots of um platform providers like ours are sharing their own studies and analyzing what is true and what is what is impactful and what the data is showing. Uh these are all everyone are using different data sets. Uh like rather than sort of say oh you know like because of that we can't decide on anything. Maybe we can see how the data is being analyzed like the size of the data sets, how they go about doing the analysis to see what um insights are more reliable than others and that would be a good way of seeing whether like you know why um this the outcomes of two studies can be different but it can still be useful in that way. So I hope that answers the question. Um there are many many many other questions today and we are at time. So what I will do now that we have reached the time is that I will take the questions that asked here and I will create a document which I will share with everyone post the webinar in the next couple days and uh you can get your you can check out the answers there later on. Um thank you everyone for joining today's webinar. Uh, I want to go all the way back to the last page here. And uh, remind you that if you share a review about HFS on G2, you get a u, you know, free gift card as a thank you for sharing your review. Share like be be honest like it's okay. We want to hear feedback from you. And when you close this webinar, you will also get a feed a quick feedback form about this webinar today. And uh please share your your honest thoughts about how whether you liked it, whether you thought some things can be improved and how you'd like future webinars uh to be done or what content you'd like us to uh present in the future. Okay. Thank you so much uh everyone. Uh have a great um start of your day, end of your day, wherever you guys are logging in from. And for everyone who's watching this webinar uh later on uh thank you for coming to the end of the video and I hope to see you guys in the next webinar or um see you guys in you know future events that HS holds. Thank you so much. Take care and bye-bye.