How AI and AEO (Answer-Engine Optimization) are Reshaping SEO | Metehan Yeşilyurt | Glasp Talk #55
This is the fifty-fifth session of Glasp Talk.
Glasp Talk delves into intimate interviews with luminaries from various fields, unraveling their genuine emotions, experiences, and the stories behind them.
Today’s guest is Metehan Yeşilyurt, a top 100 growth hacking influencer and Growth Marketing Manager at App Samurai. With over a decade of experience in SEO, product-led growth, and AI-driven strategies, Metehan brings sharp insights into how digital marketing is evolving in real time.
In this conversation, Metehan breaks down the shift from traditional SEO to AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization), new paradigms for optimizing content in the era of LLMs and AI search. He explains why outdated SEO tricks fall short, why crafting AI-optimized content fuels growth, why the “Share-on-AI” memory hack unlocks LLM reach, and why structured knowledge matters more than ever.
From AI search behaviors to curiosity-driven growth tactics, Metehan shares practical advice and fresh mental models for anyone navigating a post-SEO world. Whether you're a growth professional, content strategist, or lifelong learner, this episode will reshape how you think about discoverability and digital growth.
Read the summary
Transcripts
Glasp: Hi, everyone. Welcome back to another episode of Glasp Talk. Today, we are excited to have Metehan Yeşilyurt. Metehan is a dynamic marketing manager at AppSamurai, where he leverages a decade of experience to help mobile apps and SaaS companies scale through SEO, data-driven strategies, and product-led growth. A recognized thought leader, Metehan was ranked number 22nd among the top 100 growth-hacking influencers, and is a BrightonSEO speaker, sharing insights from working across global markets, including the US, UK, UAE, and more. And he's passionate about branding, marketing with AI and machine learning, and is currently focused on AI engineering, ML pipelines, and algorithmic optimization to boost growth efficiently. And today, we will explore Metehan's journey from SEO specialist to marketing strategist, and hear his thoughts on how AI is transforming growth marketing. And thank you for joining us today, Metehan.
Metehan: Thank you all, and thanks for the invite. I'm really excited to talk with you. We will cover many important topics in the market, and we are currently chasing AI traffic for our websites, our products, maybe for our blogs, etc. So we are here, and let's start.
Glasp: Yeah, thank you so much. So, first of all, we are huge fans of your newsletters and blogs, and also, now you are a growth marketing manager at AppSamurai. You do so many things, and could you tell us, because I introduced you briefly, but could you tell us, you know, what do you do at AppSamurai, and what's your role and responsibility, and so on, with our audience?
Metehan: Yes, sure. I started my SEO career over a decade ago, and I was just building many free websites at high school, and then my SEO love started in high school, and then I just continued to build more websites, and I was thinking, like, if I open a website, then all people will be notified, and then they will just visit my website, and then I realized that won't happen, because you will need some traffic channels to drive more traffic to your website. Then I met with SEO, and my whole career began at that moment, and right now, I'm working as a growth marketing manager at AppSamurai, and I help product teams, content team, and we have a lovely marketing team at the moment, and my main mission is bringing, actually, more customers end of each month, so I'm working hard on mobile marketing. I'm trying to understand what, especially, enterprise gaming or non-gaming company needs, and we are using many traffic channels, but organic search channel is our biggest lead channel at the moment, and I'm working with, our Sierra, and trying to drive more traffic from every organic channel, and now including AI and LLMs at the moment.
Glasp: Wow, exciting, and as you mentioned, since 2022, since ChatGPT came out, I think it changed our game a lot, SEO and AEO, GEO, and that changed a lot, so do you see the change? I mean, I know in the SEO world, once Google changed the ranking algorithm, people like SEO consultants and specialists work harder to understand what the next algorithm will be, so do you see that kind of level of impact in the current AI, LLMs, since 2022 or so?
Metehan: Yes, I believe, yeah, mostly, but the current AI hype is just everywhere, so we are living in a fast-pacing world, we need to evolve our businesses, we need to run many things at the same time, synchronously, in a fast mode, and just, I believe, I guess I started my paid subscription to OpenAI in 2022, I was curious about all AI things, AI is the new reality for everyone, and we are just trying to align with all publishing more blog posts, more news, more articles, research papers with using AI, and all of us trying to use AI effectively, so we are searching for many important points, how can we improve our jobs, our daily routines, etc. So, I believe AI changed a lot for our lives, our daily stuff, even, you know, even for our private life. So, the current hype is real, it also changed a lot for me, for my job, and Google, I believe Google has the best technology for AI at the moment, even OpenAI plays a big role. So, last year, we read many deep dive articles and blog posts about Google's ranking algorithm leak in the last May, but none of us doubled our traffic since the leak. So, because it's a very dynamic environment, and you need to consider many things at the same time, like UX, on-page SEO, user development time, and bounce rate, we need to check our content for EAT, and we need to ensure that we deliver this post to the right channel. So, I'm using AI for all this; it has changed a lot for me at the moment. So, yeah, it's real and happening.
Glasp: Yeah, totally. Yeah, it's changing a lot, and yeah, I remember Google's search engine algorithm leak, so they hadn't been saying, yeah, we don't have domain authority, but they actually have, right? And yeah, and now, so AEO and GEO are very, like, not hype, so real, you know, topic, so people are thinking, but for our audience, so could you tell us what AEO and GEO are, and what is the most important factor for AEO and GEO?
Metehan: Yes, this is a great question. I believe many SEOs are also looking for the real answer to your question at the moment. I think I still believe in SEO, and there is a new acronym in the search industry, like AEO, GEO. AEO stands for answer engine optimization, and GEO, generative engine optimization. Because we think, let's think like this: if we need to search for a question on Google in a traditional way, like five years ago, over a decade, we need to click every time blue links in the first page, we need to read them all and then just trying to summarize and we are we were basically looking for the right answer that with our thinking pattern. So with the LLMs, you just ask a question and you can read all the stuff together, and you can modify your question if you didn't get the right answer, you can modify many things on the website. So it's just becoming a new thing for the search industry because, especially with the AI overviews and AI mode full rollout in the US at the moment, we just ask something and we get all the answers just in seconds. So this is a new thing, and I'm calling it a traditional SEO because many issues are also using this, and I also believe that everything is just coming faster, you know, we're just swiping the videos, scrolling. So we are living in a way, many fast changing world. So it's basically happening in the search industry, and right now, most of the web also contains information queries like what, how, where, which, etc. So if we are basically looking for answers right now, we just ask our question and get the whole maybe with a deep research answer from LLMs in a way, basically. So, it's our search behaviors as users also changing very fast. So I believe people need to find a new term to define this behavior, and is it different than SEO? I believe many things are also working for AEO and GEO, but it's obvious there are some different patterns to driving more traffic or being more visible in LLMs at the moment. Yes.
Glasp: And I, yeah, we realized that the content work for SEO doesn't always work for like AEO, GEO, and vice versa. Why is that? Could you provide some insights around here?
Metehan: Yes, my personal research and many also research papers and many famous SEOs also share all of this information, and also Google announced this, I guess from 2015, they're using semantics, vector space, and embeddings. So Google and all LLMs can understand natural conversation, your question, for almost any language. And so I will continue from most SEOs, famous SEOs, and based on semantics and vector space, vector databases, they're all actually working in a very fast environment. So why the traditional SEO content not work for LLMs? I can give an answer to this question like, because LLMs it's basically depend on their training data sets, but it's obvious to all LLMs that structured content. I mean, if you use bullet points, tables, they will be able to understand more efficiently. So for the LLM site, you will need more structured content, and you need to connect all entities altogether in a meaningful way. And of course, you need to be cited in authoritative other websites like, let's say, Reddit, Forbes, New York Times, New York Post, or San Francisco Chronicle, etc. So it seems it's now working like this. You need to be more consistent in your content, you need to show your expertise, and these are all also working in traditional SEO, but you will need more structured content at the moment with powerful semantic connections between all entities. I mean, San Francisco is a location in the United States, and you are running a startup named Glasp. That's your organization's name. And you need to use all entities altogether in a meaningful way, and most importantly, in a natural conversation like between humans. It's also valid for traditional SEO, but it seems there are some minor nuances between LLMs and the classic Google rankings at the moment for the content part.
Glasp: I see, yeah. And could you tell us what structured content is like, you know, what kind of content AI prefers to cite or learn?
Metehan: Yes, first you need to get indexed by search engines because we already know ChatGPT is working with Bing, and sometimes there are some rumors that ChatGPT is also using Google as a search engine, I mean the web search part. So basically, you need to get indexed first, and then for the content structure, you need to use a good hierarchy between your heading text H2, H3, H4, etc. Then you need to explain your main topic, main idea in a meaningful way. You need to understand, you need to explain your expertise, how it works, maybe with a final conclusion part. You need to mention all entities that you need to mention because if you provide richer content, I mean for the entities, let's say iPhone 16 announcement. If you only mention iPhone, but if you don't mention Apple in your content, there will be a missing part for semantics and entities because iPhone is a product of Apple. So basically, it's working like this, and you need to show, you need to create some maybe tables with some comparison because it's also easy to tell and show your readers, your visitors, maybe your leads in a better way. And probably if you use a highly engaging and clearly structured post or article, or a news article, people stay longer, and probably they will want to visit more pages on your website. So you will get more engaged users, and Google will love it. And for the LLM parts, probably they will cite you more, and you will also start to build your brand legacy, and it will also help to reach more people because they will trust you. And if people trust you, probably LLMs will do that in the same way. Let's just think about the Forbes article or Reddit, or Hacker News. We all know these platforms are great, and if they say something wrong, probably there will be news for that author. So basically, we can explain your question in this way, I believe.
Glasp: Yeah, totally. Yeah, thank you. And it's related to content gap analysis as well, right?
Metehan: Yes, yes. And as I see, many solo founders and developers are also looking for their next big thing or their next entrepreneurship adventure on Reddit. They just search for something. Is there a tool that threads or questions, right? So Reddit is a great platform, and since the AI hype, we all know that Google also has a partnership agreement with Reddit, and they're using Reddit posts and all interactions to Gemini or the internal AI memory or AI systems in general.
Glasp: Thank you. So yeah, related to memory hacking, so you are writing an interesting article about sharing on AI. So yeah, could you tell us about it? Also, how did you come up with that idea?
Metehan: Yeah, sure. This is the funny part. I believe this can be the funniest part in our video call, video recording at the moment, because I called this method a sitemap. It's basically an alliance with any people's any user's personal AI memory, not the global LLM model training. But it started with a basic question. I just read many tweets on X, like asking this question to ChatGPT, and you can learn what ChatGPT knows about you. And I just asked only one single question: What do you know about me? And then ChatGPT started to explain to me my personality very well. And then I realized that, okay, this is a very powerful weapon for any company, especially for OpenAI. And then I started testing some dumb things on my personal websites, on maybe AppSamurai. So then I realized that if I have also user behavior like this, I love reading summarized documents, especially for any new research paper or new website. I love using some AI automation tools like M2M, like browser use, Monos, the most popular AI tools in the market. And of course, Glasp. So I realized that when I started to ask about these companies, these products, and in future conversations, ChatGPT started to cite my past conversations. "Oh, you were looking for this, and this can be helpful for you." Then I came up with the maybe dumbest idea in the AI search industry. And then I just placed some basic buttons on my personal website, AppSamurai. And what happened? It worked. So I was the happiest man, the happiest growth marketer in the world right now. And then I just started to share this on my social channels with a blog posting. So it went a little bit viral because it contains very conversion-oriented topics, because it's not good if you interrupt the user session and send them just to another website. Let's say you are running a blog website. You are driving 20,000, 30,000 monthly traffic. Is it good to send your visitors to ChatGPT? And do you need to wait for that user? Can you bet he or she will just come back to your website again? So there are many questions, but I just decided to publish it all. And people have already talked about this, and they shared very logical questions. And even some communities publish other WordPress plugins, which is very impressive, and they just started testing it, but it's still working for me. So it's all stuck with the basic question to ChatGPT: What do you know about me? And just, I applied this as a method on my personal website. This is it.
Glasp: Yeah. Thank you. Interesting. Yeah. We saw many posts on X, "guess my IQ based on conversation history," or "tell my personality based on conversation with ChatGPT." Yeah. It's interesting. And so you mentioned you implemented share on AI. So on your website, did you see an increase in traffic from ChatGPT, or are you not tracking?
Metehan: Yes. The numbers are not driving millions or 10,000s of traffic, of course, at the moment, I can be honest, but I know it works. And it's for the last four weeks, it's in a good trend right now, still increasing. But hopefully, when it reaches over 10s of thousands of traffic, I will also publish all analytics screenshots with the Y axis. I hide it. And I have many questions about it. I sent some private recorded videos to some entrepreneurs. But I know it works. And many methods work for LLMs. But of course, like the web search and citation, if they are working with a partner like Google or Bing, or let's say Copilot, etc., DuckDuckGo, any search engine, it also lies with the current traditional SEO rankings. And there's a reality that it's happening, actually bad for the small publishers and independent publishers, which is helpful content updates and all new AI features on Google. When you ask a question, AI always just answers that. You can ask for a recipe or travel plan in AI mode. So it all kills traffic. And basically, people realize that they haven't invested in informational queries, like where and how. They're very basic questions. And you don't need to visit the website to read all the articles, summarize them on your own, in your mind. So LLMs are covered. And Google, this area, it's going bad for many issues because impressions are going higher, but traffic is not the same and is decreasing month by month. I'm also experiencing this issue. And right now, it seems that if your brand is more visible on LLMs, on Google, it's just good for your conversions, good for your ARR, and good for your investors. So it's really weird times that we are living, especially with issues.
Glasp: Totally. Yeah. I would like to see your analytics. So, if you can share how it's working.
Metehan: Yes. I will send you a screenshot after this. Oh, yes. And you can just put all of these screenshots in a bigger picture, full screen, and then people can see what the number is.
Glasp: Oh, yeah, thank you. Yeah, we are looking forward to seeing it. Did you see any interesting growth hacks with ChatGPT or other LLMs? You mentioned there is a WordPress plug-in for memory hacking, but did you see any interesting growth hack related to memory hacking, ChatGPT?
Metehan: Yes, I'm still working on these useful and impactful growth hacks at the moment. I love reading archives from archive.org, I guess I'm pronouncing it correctly, from Cornell University. I just searched "LLM ranking stations" in the title field, and some good friends from the search industry are sharing some research papers with us. And one of them is talking about the methods named StatRank. It's basically a manipulative attack on LLMs because you leave some digital fingerprints on your website, if you are running an e-commerce website, on your product page, news article, or blog. It's basically telling that if you use some effective and semantically rich directions in your content, it can drive more citations for the future from LLMs, because you are using a language like "this is the best product in the market." I'm not saying that if you just use this sentence on your content, that will work, but there are some unique methods without using the top, the best keywords in your page that can work, and LLMs can cite you more. And they are using different algorithms right now, different algorithms like Perplexity, which we never mentioned in our chat until now, but Kagi is also using another web search model. Claude is just new to a search, web search feature. OpenAI also has an API for the web search feature. So they are all using different but similar methodologies. So it's easy to get more citations from LLMs, but it also requires a huge work, because you need to, if you need to show your expertise, your authoritativeness, you almost need to cover all related topics in your industry. It can be a glossary part, a blog part, service pages part. You need to tell everything. For the real-world scenario, I publish many blogs, technical stuff on my blog. And then we met on social media, because basically you read my blog, visit my webpage, and I visited your YouTube channel, your product Glasp, and all the landing pages. Then there is a trust between us. So, Google and LLMs are working in a very similar and technical way.
Glasp: I had a question, like a random question, but do you think in 10 years, 20 years, people will still be searching through Google, or Google will be gone at some point?
Metehan: That's a great question. I still think about this sometimes. We are like going into an agentic era, because yes, we love using LLMs at the moment, and it's reached billions right now, because Google has integrated Gemini with Gmail. And Meta is also using their own AI models in their social platforms. You can interact with Perplexity in WhatsApp. And of course, we ask questions like, "Is it true growth on X to verify some facts at the moment?" And I believe Google won't be dead, but it's obvious it will affect their total revenue from search and ads. I believe they're also still looking for the most efficient way to get more clicks from ads in AI mode or AI overviews, because it's obvious if people already engage and find the useful data from the old summarized AI overview or from LLMs, they need to create or adopt their revenue features, revenue streaming products a little bit more in a different way. And also, I believe Google has its entire web. You can ask how to create and ship a rocket to Google. Basically, they have this information from research papers, scholarly academics. And I believe Google has the best infrastructure because they are using real-time machine learning systems, especially their search and ads products. So probably they won't be dead, but it's obvious everyone in San Francisco, in Japan, for every industry, they're looking for new agent solutions for anything, for medical, for SaaS startups, for any even small businesses. So Google won't be dead, but it's obvious. They also announced Project Gemini, their new agent project, and Perplexity came up with Comet, their new browser. And we know that OpenAI is also working on a new agent browser. So the agentic era can be the new hype. And I really want to see who will get the lead because it's obvious Google is not getting the leadership from OpenAI over the last year. Agreeing with the popular coding startup, they were just about to join OpenAI, but then all agreements crashed, and then they all just decided to join DeepMind, another Google company. So basically, it looks like this at the moment.
Glasp: Yeah, totally, yes. But in that sense, people will Google less in the future, since you mentioned through like, because of like an agentic workflow. And if so, do you think people will be reading someone's blog in the future? I mean, currently SEO's algorithm is based on, as you mentioned, how many people click through the query and also how much time they engage with the content, or something like that. But if no one is Googling and no one, if few people are Googling, few people come to visit the website, and how can Google understand, oh, this content is good or not? So, if so, I think currently LLMs partially rely on something like those Google-like search engines' algorithms. So, if they can't use it and how does it work in the future? And I don't know. Yeah, what should we think about in the future?
Metehan: Yes, I don't know too, but I believe. In 2016, I needed to go to London, and I just wanted to learn where to visit in London. So basically, I started with a Google search and just started taking some notes. I guess it took around three or four days because I needed to ensure everything was planned perfectly. But right now I ask the same question to Google, ChatGPT, Perplexity, Google Cloud, or even Mistral, or any other agent LLM startups or companies, they just return with a very highly detailed response with a London trip plan, just under a minute. So for the future, and it's very funny, and we didn't mention Google's latest video generation model, which works very perfect.
Glasp: You mean Gemini Pro 2.5?
Metehan: No, VEO, VEO.
Glasp: Ah, VEO. Yes, VEO, yeah. Ah, yeah, yeah, yeah, it's okay.
Metehan: VEO, you can even clone your voice, and you can just click something, and you can even generate a movie in VEO. In a short movie, of course, it's very expensive at the moment, but the current situation is very funny at the moment, because we were watching Will Smith eating pasta, and all the first outputs were surreal, and right now we can export with a few clicks, very hyper-realistic movie sets from any new AI video generation tools. For the future content and blogging parts, I believe it won't be the same as today, because all AI and LLM search engines basically are reading your blog, so I also agree with the Cloudflare CEO, who is the latest speaker for this year. Yes, most of the users are bots at the moment, so I believe we will enter an agent-friendly blog posting or social media, so we will have many different agents for ourselves, I believe. Okay, this is my news agent, this is my social posting agent, and we will see many early examples at the moment. Okay, this is my YouTube summarizer agent, and the best, you know, the Glasp, highly recommend to everyone to use, and they will just scrape our content. There are also many ethical considerations. Of course, if you're running the New York Times, we agree to do that or work together with the agents, and they will all just bring summarized versions, even with a maybe shorter version, and we will get all the information, and then we move on. I don't know, it looks very chaotic at the moment.
Glasp: Yeah, there will be many AI agents for social media, for blog posts, for technical things, so it's chaotic. And so for new grads, like junior level, so marketers, growth marketers, SEO content writers, so what are their vital skills in the AI era, and do you have any advice for them?
Metehan: Yes, I'm also still thinking this for myself, because there are many great memes, like AI will replace our jobs, at the moment. Of course, I don't believe SEO is dead, or any marketing, proficient at the moment. We need to adapt to the fast-paced environment because everything is changing very fast. So we need to adapt using AI, especially automations, and they will need expertise. They won't need expertise at the beginning, but I believe if they have enough curiosity and an experimental mindset, they will probably win in their career. So you can build many MVPs, or your internal use mini-tools, using AI in just minutes. Maybe with many bugs, but you can create a real working MVP in just under 10 minutes, or under an hour. Basically, five years ago, during the pandemic, there was also another hype in the digital ecosystem. Many companies have just started to hire thousands of people. They tried to scale their business, and now we are living in another world, because inflation rates, all the world has changed, and then AI came. So they will need curiosity, and they will need growth-hacking mindsets at the moment, I believe. And if they share their ideas, experiments more, and if they become more visible for their expertise on social channels, on YouTube, on LinkedIn, on X, on Blue Sky, et cetera. Maybe if they're a designer, even Pinterest can work for them too. So they need to bring unique ideas and experiments, even if they fail, because we need to fail harder and harder nowadays. We need to test everything using AI, I believe. And just share with people, even if it's a dumb idea, or even if it's for nerds. So, I highly recommend that they need to learn storytelling, how to tell their idea to people in a more efficient way. Because if you, I'm not promoting my social postings, but I'm basically using three post templates. If you just visit my socials, you can see some similar patterns for every posting, almost every posting. And it's trying to align with the X algorithm, LinkedIn algorithm. So we need to learn the basics. We need to use AI efficiently. We need to save hours. We don't need to work hard, but efficiently. And they need to have curiosity with a growth mindset, and everything will come after that.
Glasp: Yeah, you mentioned like post templates, right? And for each social platform, is that based on storytelling? Or could you tell us, tell our audience, what it looks like? Like, could you give us some examples?
Metehan: Yes, and I don't have tens of thousands of followers on LinkedIn. I have around 2,600 followers on X, but I realized that there are many social posting patterns and templates that are working very well. It's not working in your first tweet, but I always start with like, "I found X using Y for AI search. It works, it's dumb." And then start explaining with maybe bullet points, emojis. And then people just start, Okay, I guess this can work." And the hardest part is being simple. You know, you may have many complex ideas or you may have complex solutions for your SEO operations, SEO projects, let's say, or an AI search project, but you need to tell all people in a very simple way. It works even for investors in San Francisco, you know, you may have a very complex idea, but it's valid if you can simply tell the whole story. You told me to explain your product, Glasp, with very simple terms. It's working like Pinterest for documents. It's very simple and people can remember this. So they will, you need to be simple, and you need to solve one problem at the beginning, I believe. And for these postings on my social, you know, three or four years ago, people were always telling the same things about SaaS. "You need to reduce your churn rates. You need to have a good monthly recurring revenue. What's your ARR? Here are the five effective ways." Or "you need to optimize your website's loading speed." Okay, I already know this, and probably you are the 97th website I read the same topics with the same ideas. So what's the difference? Then I just started to implement all my complex ideas into simple things. And then I'm not saying I'm the best, but I'm also a lazy person. So I believe anyone can do this for their product, for their teams, etc., maybe personal branding. But I believe if you simply learn storytelling, your influence and your reach rates can be higher. That's why I created these posting templates and a trying to be more friendly with storytelling. I hope I explain myself better.
Glasp: Yeah, yeah, totally, yes, thank you, yeah. And by the way, Kay and I are brainstorming, and we have some ideas like growth hacking ideas and so on. Then like you and ideally like your like a share on AI idea, memory hack ideas. But Kay already asked you, but you know, do you have some ideas you are thinking now or you think, oh, it's going to come in the future or you haven't mentioned today, kind of tips or hacks or in the area of AEO or GEO or anything like you're thinking or you haven't talked today?
Metehan: Yes, yes, I can, I can share with big. I guess I'm working on semantic chunking for any web page. Let's say, let's explain it in a basic way. Basically, LLMs have a context window. They can't remember longer chats, or if you can just upload thousands of pages in a single PDF, they won't process it because of their context window. So if you use longer paragraphs on your web pages, it's harder, it's hard to parse for LLMs. They are basically trying to guess the remainder of the paragraph because you're using very short context for LLM crawlers. Maybe it's also valid for Google or other AI search engines. My idea is basically also many SEOs are talking about this, I'm not the first one, but if you use maybe shorter paragraphs on your web pages, maybe it's better for semantics and LLMs to understand. Also, I'm trying to apply the stat rank. You can check from archive.org. It's a great research paper, and you need to use this sentence in your prompt. If you just upload to any LLMs, it probably won't work because it's a prompt attack. So, probably you will be stuck with the LLMs' safety filter because it contains some manipulative attacks. So I can also say this. And what's for the future? I have many dumb ideas, of course. So yes, let's see. I can, I can mention three of these topics at the moment. And I love using Screaming Frog, by the way. It's also good for your looking, reviewing your AI search performance. Of course, there are many tools like Profound, Semrush, Ahrefs, and Vacay. There are many great tools on the market. But for the technical part, probably Screaming Frog is the only software at the moment that's working on your desktop only right now. But you can modify and customize anything for LLMs for crawling your website, like LLMs at the moment. And it's working like plug and play. So you can insert your API keys from Gemini, OpenAI, and Claude, and you can also integrate with Google Analytics, Google Search Console, and all other third-party SEO tools. So it just shows you the big picture. You can see everything around your website on the technical side, like is my website is LLM-friendly or bots can crawl in an easy way? Do I have broken links? Because we also know there are hallucinations in LLMs. So Screaming Frog is highly recommended at the moment. Just a note, this screaming.
Glasp: Screaming Frog, right?
Metehan: Yes.
Glasp: Okay, thank you. And yeah, you mentioned the semantic chunk. So it's about the research paper by Cornell University published in 2016. It's the same thing.
Metehan: There is new research, but different, yeah. But on the same websites, a new research paper, yeah.
Glasp: Got it. Yeah, thanks.
Glasp: Yeah, okay, one more question. So do you see any interesting AI, how to say, AEO, GEO, so analytics or content generation tools? They're not only like, you know, current ones like HubSpot, you know, Ahrefs, but there are many like AEO, AEO tools like. So do you see any interesting tool also like, you know, that startup company or product like Air Opsthat works well in the future? If we cannot answer, it's okay, yeah.
Metehan: Yes, yes, yes. I can, I can say Profound because I believe they launched, I guess, a PR campaign. It just doesn't matter because they have good partners and clients at the moment, and I love them. Still, I couldn't have a chance to sign up for Profoun, but I heard their name a lot. And of course, there's a great AI tool named Glasp. You need to use everyone, yeah. And for the overall AI performance, I can say Vacay. We can leave the URL in the description. Yes, it creates a very comprehensive report. It also shows your entities, your competitors. You can compare your topics for Gemini, Claude, or Gemini Grounded Search with web search, Google, OpenAI, and Perplexity. Of course, I love their reporting dashboard. Very detailed, and it gives you the big picture. If you just export the PDF and send your client, they just love it. So I can recommend this. Of course, you can create maybe some mini tools using Cursor or other AI wrapper tools. Ahrefs and SEMrush have new dashboards for AI search, but the traditional SEO, basically, they are crawling every web page just like Google, and it's already a very expensive business. So and for the AI site they have I find their AI search products. It's a little bit expensive, but many great tools at the moment. And of course, you can also ask your brand name directly to Virtual XT and ChatGPT. It's also a great insight. And I highly recommend asking your brand or personal name in a temporary mode on ChatGPT to not use your personal AI memory, because if you ask a lot, it will tell you more details about the brand. So you can not fulfill personal memory. So you can just use temporary chat or a new account on Virtual XT.
Glasp: Yeah, that makes a lot of sense. So this is another random question. Not a question, but a source. But I wonder, because AHF and SEMrush are those search analytics tools for billions of web pages, and they have their own domain authority scores, and they analyze content. I wonder why they don't create their own LLM? Because they can understand which content is good. And if so, I think they could create the LLM.
Metehan: Yes, I believe. But it's another business. And of course, you are now in San Francisco. And I guess many companies are trying to build their own LLMs, also related to Japan. Maybe some big Japanese companies are trying to build the most powerful LLM. But I don't think we need to rediscover everything from scratch. There are many great open-source LLMs at the moment. But yes, I believe they found LLMs in a different business type, I believe. Basically, you are crawling websites for traditional SEO to calculate some metrics: visibility, domain authority, et cetera. But for the LLMs, yes, you need to crawl for citations. And basically, you need to track all prompts. And maybe you can just create more synthesized prompts to get visibility for any brand. Maybe that can be the answer. And I believe this is the most important part. In the 1930s, I guess, in the USA, there was a gold rush. But who got the most money? The sellers, the jeans sellers. So it basically looks like this. They're basically selling jeans for SEOs, Ahrefs, and SEMrush. So they might get more money from their businesses. Maybe that's why. Just a guess, yeah.
Glasp: Yeah, that makes sense. Yeah. Yeah, thank you. So since our time is, yeah, it's been a while. But so our last question. Since Glasp is a platform where people share what they're reading, learning as a digital legacy, we want to ask this question to you. So what kind of legacy impact do you want to leave behind for future generations?
Metehan: Yes, this can be funny. But you need to remember that every time someone needs to pay the bills. I have a daughter at the moment. When I was younger, everything just looked perfect to me, for anything. But when you grow up, you will get more responsibilities. So you need to create your legacy starting from right now. Even if you are 16, 17, or 18, it's all OK. You need to start building it right now. Even using Glasp, you can take some notes, summarize them, and maybe all your notes from your different sources. But you need to act now. Because if you use AI efficiently, you can create thousands of new blog posts in just an hour, twice. You can run a new website. It took weeks and months, five years ago. So you need to have a growth mindset. You don't need to be afraid of your failures. Just fail harder, try to learn something fast, and share your failures, ideas, and experiment results. And for your legacy, you need to be unique. I mean, you have a very unique product at the moment with a simple idea, Glasp, because you can basically summarize any PDFs with ChatGPT. Or basically, you can share videos on Facebook. But YouTube is the biggest player. You could create events on Facebook. But there is now a Ticketmaster event, right? So you don't need to be afraid of anything. And just, of course, it's a cliche, I know. But believe in yourself and start solving only one problem at a time, I can say. And if you are trying to solve a common and not unique problem, let's say, for SEO website loading speed, probably that would work, because everyone already knows about it. You need to solve problems like how to get indexed faster in Google, because people are experienced. So you can test your idea by using deep research on almost every element. Find that idea. Work harder, because you will notice more hyper-focused areas for your expertise. And then it will just all work. You can create your own legacy, I believe. It's working for me. I hope it's so valid for literally anyone.
Glasp: Thank you. Yeah, that's so beautiful. And yeah, thank you so much for joining today. We learned a lot from you. Thank you.
Metehan: Yes, thank you so much. I'm so excited and a little bit stressed. But Kazuki and Kay helped me a lot since the beginning. Thank you so much. And I kindly invite everyone, if you didn't try Glasp, our lovely tool, please sign up just right now and start building your own legacy. Thank you so much.
Glasp: Perfect. Yeah, thank you.
