What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
Beyond Keywords: The Paradigm Shift from SEO to AI-Driven Search
Interested in being a guest? Email us at admin@evankirstel.com
The digital landscape is undergoing a revolutionary transformation as artificial intelligence reshapes how search engines connect users with information. In this riveting conversation, mathematician-coder Andreas from artios.io shares his unique perspective on the future of SEO in an AI-driven world, drawing from his experience building his first large language model five years ago.
Andreas cuts through industry confusion with a striking analogy: many SEO professionals are like pilots flying planes while the flight controls have switched from English to hieroglyphics – yet they insist nothing has fundamentally changed. This disconnect stems from a lack of understanding about how AI actually processes and evaluates web content.
"The content bar has risen dramatically," Andreas explains, challenging the notion that traditional keyword-targeted articles can succeed in the AI era. Where optimization once meant producing industrial-scale content targeting search phrases, AI demands genuinely valuable insights that actual professionals would want to share with peers. "A technology leader doesn't need to read 'what is ransomware recovery' – they want to learn from others who've navigated that challenge."
Perhaps most revealing is the paradigm shift in audience targeting. Rather than relying on keyword metrics (which Andreas calls "fictitious" and designed to maximize ad revenue), forward-thinking companies are data-mining conversations of their target buyers to understand what they truly care about. This approach creates content with proprietary value that AI recognizes as worth recommending.
For businesses watching their search traffic dwindle, Andreas offers both caution and hope. Those continuing with traditional tactics face increasing irrelevance, while those who understand AI's evaluation mechanisms are finding tremendous new opportunities. As search evolves from keyword snippets to conversational paragraphs, the companies that adapt their content strategy accordingly will thrive in this new paradigm.
Ready to transform your approach to visibility in the AI search era? Start by focusing on original insights your specific audience truly values, not just information they can find anywhere else.
Discover how technology is reshaping our lives and livelihoods.
Listen on: Apple Podcasts Spotify
More at https://linktr.ee/EvanKirstel
Hey everybody, Really fascinating discussion as we dive into the future of SEO in an AI-driven world and what brands need to do to navigate what's next with a true expert and innovator in this field. Andreas, how are you?
Speaker 2:I'm great and thank you so much, Evan, for having me. I look forward to sharing any useful knowledge.
Speaker 1:Well, you have much useful knowledge. In fact, I believe you wrote the book on SEO, not that I've read it. But you certainly have a deep and storied background in this space and you've thought a lot about the role of AI. But I'm jumping ahead. Maybe you could introduce yourself your background, your non-traditional biography, I think, in accounting, data science and other topics.
Speaker 2:How would you describe yourself these days? I would describe myself as a mathematician coder who's passionate about organic growth. That's probably in a line it you know I'm interested in solving problems. One of my hobbies is to uh optimize my stock portfolio using uh data science models, but it just so happens my day job is to help companies grow, with a particular emphasis on organic search, which is AI search or is now becoming known as generative engine optimization.
Speaker 1:Wow, we're going to have a lot of buzzwords today, which is good. I'm a fan of buzzwords. But even before that, what led you to start Artios and focus on a data-driven approach to SEO, versus the black magic that many people take to this topic?
Speaker 2:Well, this is it. So I started doing SEO around 2003, and I found that there was a lot of thought leaders going online, whether it's on webinars or published articles, claiming how Google worked. And you know, as someone who's quite data driven, I'll be thinking, ok, well, what's the veracity or the rigor behind your analysis? And I found, you know, it was quite rare, and I found it was quite rare. And so this, coming from an accounting, a certified accounting background, I decided to retrain in data science because search and digital marketing in general is incredibly data rich, and when I graduated, digital marketing wasn't really a thing.
Speaker 2:So I actually wanted a career in marketing, not in accounting. I only did accounting because my father said I should. So I actually went to the career in marketing, but you know, your ad agencies weren't really looking for economics graduates like me. They were really looking for for, for um for for economics graduates like me. They were really looking for psychologists. So I decided to to become an accountant, and so when digital marketing came along, uh, I was just like, wow, this is amazing. Finance role to digital marketing with specialism in SEO, and I've been very lucky to work with really incredibly talented and knowledgeable people who have backgrounds in really competitive sectors like online gaming, poker, you know from across the world.
Speaker 1:Yeah, it's such a fun area to be in, and let's dive into the hot topic of the day. Everyone's talking about AI and SEO. Is AI making SEO easier, harder, or something altogether different?
Speaker 2:Is AI making SEO easier, harder, or something altogether different? Depends who you ask, because I would say, for those who do not know how to meet the challenge of AI search, then it's making life extremely hard because most companies are seeing dwindling traffic from traditional search. If you ask the other type of person who does know how to meet the challenge, then they are having lots of fun, as I am, and from my own personal history, I built my own first LLM five years ago when GPT-2 was first announced. So when I did that, it gave me an incredible insight as to where the future was heading. And although I lived under no delusion or illusion, I wasn't deluding myself into thinking that, you know, I'd be able to compete with the budgets of open AI. It did make me, it did give me some idea as to how to respond to the challenge.
Speaker 2:And here we are, five years later, fast forward. Ai search is here and you know, if SEOs think the future or the now is scary, the future is going to be even scarier because, uh, especially for businesses, if they're not investing in ai search now, uh, they're going to be nowhere tomorrow, because tomorrow is is going to make use of agentic search where you know, at the moment, today we have people, we have a situation where people search and AI searches reactive, like a search engine is, so it's reacting to the prompt, but agentic search will be where your search agent already knows what you're looking for and what you want. So it's finding these things in the background which will cut down the amount of time you spend searching for something.
Speaker 1:And it seems also obvious to us as users. I'm spending all my time searching on chat GPT, for example. But let's take the big picture. What are some of the biggest myths you hear or read about in the industry? About SEO, maybe in general, and the impact of AI? What? Would you love to bust as far as common beliefs bust as far as uh.
Speaker 2:Well, well, let's start with uh. You know the, the seo belief. I'm seeing a lot of uh talk on linkedin, uh, and, disturbingly, among well-known thought leaders that you know ai search is not a thing. There's no such thing as geo, ie generative engine optimization um, it's just an extension of seo and therefore, you know um uh to you, know you to succeed in ai search and get, get your business recommended by ai, you just need to do seo. You know it's not search engine optimization anymore, it's search everywhere optimization. But that, to me, can only be a view founded up by, founded upon by people who have no idea how to build an ai or how ai works. Um, you know it's almost like the, the, the google or the traditional uh search channel.
Speaker 2:Uh, plane is is bombing. You know the planes go and the pilots are the SEOs and they're going la, la, la. You know no such thing as GEO. You know you can succeed in AI search using SEO, but the problem is is that the flight controls have completely changed language. It's almost like you know, before it was written in English, for example, and now it's in Egyptian hieroglyphs and they're going la, la, la, no, no, no, it's still an extension of SEO, is there's no such thing as geo? But the plane is still going. You know, it's still dive bombing um, uh, for, for their, for their clients, and so this is, this is what I think is the issue, and and you know to take that, if I may extend that further, I'm I'm also seeing, um, some advice on how to respond or meet, meet the challenge with.
Speaker 2:Oh, you know, what you need to do is you need to create content, think mode or deep search mode, and that's incredibly flawed because it's like you know, you wouldn't rehydrate by drinking your own sweat. Sorry to use that analogy. Why would you ask AI to get smarter on its own outputs, which are extremely diluted, summarized and derivative, if you know what I mean? So that's not going to work because that kind of content is not making AI smarter and therefore AI recognizes that and is definitely not going to recommend your business.
Speaker 1:Got it. So it's evolving into something new and different. Seo, I don't think you'd say, is dying, it just will be something else. So what will it be in two, three, five years, a year, six months? What's the evolution? What does it look like?
Speaker 2:well, this is what I think is what's happened now is seo has been relegated. Uh, I don't want to use the word back office operation, but uh, you know, there's a layer on top and the reason for that is the seo basics still apply. You know, you still need to have a website that has content discoverable and when it is discoverable, it needs to be passable so that both traditional and AI search technologies can evaluate the value of it and make a decision as to whether it's going to include it in its results, as to whether it's going to include it in its results. Now, what AI has done?
Speaker 2:Because we've seen at least two years of SEO abuse to such an extent that the Verge wrote an article saying the SEOs ruin the internet. Now, obviously, it's a little bit clickbaity and perhaps a bit extreme in its interpretation, but there is a bit of truth in that. You know, we've we've seen the internet filled with what is target keyword articles? Okay, um so, but what ai has done is it's raised the content bar in terms of quality. So now we can't get away with loads of these low effort ultimate guides to whatever your target keywords are, because it's simply not enough. You know, in order to get into AI, you've really got to provide high value and therefore that requires a lot of effort. Substantial resources requires a lot of effort, substantial resources. So this is the difference between SEO and GEO. There's a real gap between the tactics and the content quality that has to be met in order to get into AI.
Speaker 1:Got it. So let's take this little chat, for example. You are I think we all agree an expert guru in this field. This discussion will be deployed as a long-form podcast, a video, short-form social media content that will go out on Twitter and YouTube and Facebook and Instagram and every platform iTunes, spotify, youtube, podcasts. It'll go out as a written blog from the transcript that captures all the discussions. Is that all I need to do is have more high level, expert level conversations, or there's? There's more black magic that has to be done to, to, to get get ai seo to recognize, uh, what I'm doing here well, as my daughter often corrects uh, adults around the dinner table.
Speaker 2:It's not magic, it's science, and so I I'm gonna have to side with my daughter on this one. Um, it's not so much the black magic, it's more about let's contrast SEO content. It was produced at an industrial scale so that it could maximize the business's exposure across multiple topic areas that their customers cared about. But that content, whilst it was good for search and, in particular, google and bing, it would not be good enough to share on social media. Why? Because you know, let's say, your target icp is, um, a chief, uh technology officer. You know, technology leader now, uh, a technology leader does not need to read an article on what is ransomware recovery, right? So, uh, this is where we've got to.
Speaker 2:But this, you know that that was going on with seo because it worked. But now, with ai, a technology leader already knows what ransomware recovery is. What they really want to know is what can I learn from other technology leaders that have gone through that experience? Okay, so that is the difference. And you can't produce that kind of content simply by copy pasting, you know, analyzing the top 10 ranked articles, taking all the best sections and all the common sections and then plus one-ing it with you know, a longer word count or richer detail, because really ai can see right through it. Okay, because there's no real information gain or proprietary value. So there's no reason for the internet, which was built to inform, to prefer you and therefore there's no reason for AI to prefer you and therefore recommend your business.
Speaker 1:If you're following those tactics, Wow, that was a mic drop moment for sure. I think I'm starting to get it. It's only taking me 15 minutes and I'm getting to catch on. So is Google losing its grip on search to chat, TPT and perplexity and Claude or will their own AI sort of come to the rescue and save the day Some evidence that their search business is fine with their current AI initiatives and their push to be an AI search engine as well?
Speaker 2:The mathematical mind in me hesitates to make such bold predictions about the future, but I will say this if Google was personified, I believe it would be thinking if you can't beat them, join them.
Speaker 2:And to some extent it's doing that now with AI mode, where AI mode is a similar user experience to these LLMs like ChatGPT, where there's no 10 blue links to be found. It's pure information, answers, enhanced with citations. And it's funny how it's come full circle, right, because when we look at when Yahoo was around in 1998 and Google was around, google had the simplicity of just the simple search bar. And even when you, when you got the search results, it was just simple 10 blue links and and you had maybe a couple of ads. But now now, now the you know search results page of google looks like you are in 1998. So it's really funny how it's all come full circle. And then chat, gpt and perplexity, they all come along and they start with the simple search bar. You know how can I help you today? And? And um, yeah it, it just reminds me how you know AI today reminds me of what SEO was like 25 years ago.
Speaker 1:Wow, speaking of what SEO was like. So fast forward. You have giant SEO agencies who have perfected the art and science of SEO. So are they going to go obsolete or maybe reinvent themselves? Why and how?
Speaker 2:Well, those who don't reinvent themselves will go obsolete or they will be relegated to the very, you know, the lower value adding components. And the reason why I say and the ones that do adapt will obviously thrive. And you know, the SEO basics still matter. I think one of the other myths, if I could speak to that, is there seems to be another common myth that you know, ai SEO is just sprinkling some AI SEO fairy dust and, and you know, if you just restructure your content to FAQs and put the immediate, the answer immediately below the header and and put, you know, jump points at the top, you know that that's enough to get you there.
Speaker 2:Um, the trouble is is that if everybody, if you could do that, then everybody else can. So you have no real competitive advantage. And that goes the same for deep seek, you know, or or a deep mode on um, on chat, gpt or perplexity. If you could do it, well, so can they. And even if you spend two hundred dollars a month, well, so can they. So you have no real competitive advantage. So if, if, if, um, if these agencies um keep thinking that way, they will go obsolete, whereas those who understand and use that understanding to adapt to the challenge of AI will obviously thrive.
Speaker 1:Brilliant. Well, there's hope for us some of us in any case. The other challenge, it seems that, I guess, is the favorite term of art, which is why I love these kind of conversations, because this entire chat will get indexed and I turn it into blogs. And it's the opposite of AI slop real, unique human content, generated content with maybe some AI fixing up a little bit. But what's the role of creativity and original content versus AI-generated content for the new paradigm of search?
Speaker 2:So one of the things I realized when building my first LLM is that it's immediately clear when content is AI-written. Now, a couple of years ago, I was hearing arguments that it doesn't matter whether content is AI written or not. What matters is the value. But I would argue against that because I mean even pre-AI. I believe that any copy or content you write, every single word and sentence, needs to fight for its life in order to be included. So AI, I'm afraid, is not at that stage where it can overtake a human to make a sentence fight for its life in order to be great copy or great content.
Speaker 2:Um. So look, it's just way too obvious statistically to an llm or an ai that um, uh, where content has been reproduced. So if you're taking shortcuts there, then it's not a very good way to introduce yourself as a brand. Trying to get your content recommend, trying to get your business recommend, recommended to your target buyers. When you're um, you're presenting it with content that may not be, uh, top notch, as we say in the uk yeah, a great, great point.
Speaker 1:And so let's talk about artios. Um, I've heard a couple already in our short time knowing each other amazing recommendations on how you changed a couple of your clients' businesses. I mean, you don't hear that sort of thing too often. What was the secret behind your engagement and how do you help in general clients transform from legacy SEO to this new world of AI bots?
Speaker 2:Yeah, it's, thank you, evan. It's a paradigm shift in order to achieve the kind of rapid, significant results that we're achieving. And the paradigm shift is this Traditional SEO is relying on keywords, but even that was flawed during the traditional search era, and the reason why it's flawed is because keywords don't really tell you who's searching and therefore, and also the numbers are are fictitious as well. The numbers are based on what google ads things, so those numbers are purely designed to maximize the revenue per click for google. They are not there to serve uh, they're not there to serve um marketers, and so, um, the.
Speaker 2:The paradigm shift into ai search is well. First of all, let's just recognize that ai, as a search technology, is far more advanced. You know, people are now changing their behavior, from typing a few keywords and iterating on that to typing a paragraph on. This is what I'm trying to do, this is who I am, or why, my context, this is how I want it done, and this is why I want it done, and this is why I want it done, and, um, and, and, and. Therefore you get a very customized answer, which you couldn't before through traditional search.
Speaker 2:Now, if you bear that in mind, that means that your input parameters have to be far richer than just keywords. Okay, so what we're doing to meet the challenge is we're actually data mining, um, we're data mining, uh, the conversations online had by, uh, your target buyers. Why? Because if you know what they care about, then your content strategy is going to be far more on point. Okay, and you're not going to be producing, um, articles that appeal to everybody. You're going to be producing articles that appeal to your target buyer, and that is the difference.
Speaker 1:You make it sound so simple, but congratulations on peeling back the layers of data and analytics to make this work Really intriguing conversation and can't wait to see what's next.
Speaker 2:Yeah, totally. I mean, when you think about how we're in the second wave of AI, okay, so I'm a bit past the excitement of LLMs. Having built my first one five years ago, you know you soon realize that, yes, the results are impressive and definitely a leap forward from traditional search. And the first wave of AI was machine learning, okay, and then the second wave of AI is where we're at now. It's LLMs, and really LLM is just synthesizing from a large memory and instead of giving you a list of documents, it's actually using the language synthesis to give you a summary. And that is quite impressive.
Speaker 2:But it's still just based on memory and that's why, most of the time, if it's faced with a situation that it's got a memory of ie its model is trained on something that you're asking it then its hallucination is like 0 to 5 or 10 percent. You know, hallucinations may vary, but when it's faced with a prompt or a request on something it hasn't seen before, then it breaks down um and therefore um it. You know the hallucination um rating goes up to, you know to to more than 50 or you know to a point where it's not not useful. And the clear is in this uh, it's in the acronym GPT. It's generative.
Speaker 2:Now it just so happens that the LLMs are trained on billions of documents, so the hallucination component of the generative output is very small. But when it's on areas of knowledge that is unknown, it doesn't think where it can respond and respond smartly in the face of situations or information that it hasn't faced before. And this is what some might call symbolic knowledge-based, you know, like critical thinking type AI, and that's going to be truly exciting, and I think we're probably five to 10 years away on this. We've seen Meta hire some uber smart AI scientists for $200 million speculatively, so that just shows you the race is on.
Speaker 1:The race is on and you seem to be at least two or three meters ahead of everyone. Congratulations on that and I can't wait for an update in a few months and see where we are. Really appreciate the time and everyone you know. Reach out to Andreas and he's a great guy to talk to and get help. We all need collaborators and help navigating this brave new world. Thanks very much, andreas. Thanks for your time.
Speaker 2:Thank you, Evan, it's been a privilege.
Speaker 1:Thanks everyone for listening. Thank you, and check out the new TV show at techimpacttv Now in Bloomberg and Fox Business. Take care everyone. Thanks.