GMT20260507-150648_Recording (1).m4a: Audio automatically transcribed by Sonix

GMT20260507-150648_Recording (1).m4a: this m4a audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.

Speaker 1:
Check. Check. One. Two. Check. Check. One. Two. Hello? This is when it looks good, I guess. Hello, everybody. Today we're going to talk about generative engine optimization or what we're calling geo. It's the new search engine optimization. And basically we're going to talk about how to get your website ready to be findable and indexable and a marketing tool for answer engines, which are essentially your perplexities, your clouds, your copilots. All right, so this presentation will go over how to look at your current state of affairs, your current website and content. It'll talk about how to organize that. It'll talk about the differences between how search geo search engines or AI empowered search engines differ from traditional SEO, but still build on traditional SEO. And then it'll talk about how to create a roadmap for you to optimize your website and eventually generate what is called an LRS. Sorry, LLMS dot txt file, which resides in the root folder of your website domain and is basically a menu or a map of your website that's specifically optimized for engines. Okay, so let's just go ahead and jump in. Um, as we mentioned in the new era of search, keyword search is declining. Traditional search is fading as users shift to AI powered search engines as ChatGPT, Google, Gemini Anthropic's, Claude, Perplexity, copilot, etc.. Um, this is known as the rise of conversational AI. And because the, uh, at least one of these AI powered search engines leverages a large language model, we're able to ask, which means basically, instead of just looking for search terms, very short phrases, we're able to have a conversational query to an AI engine for it to go out there and find specific information.

Speaker 1:
It's just a much more comprehensive understanding of, of all languages and, and therefore meaning of, of what someone's looking for. Ai engines are better at finding the intent, the true intent of the user, even if it isn't an educated user. And we need to be able to optimize our websites for that purpose. Uh, by next this year, there's a 25% basically of search traffic is going to answer engines or AI powered search engines, and that's going to continue to increase significantly year over year going forward. Uh, AI chatbots are becoming a primary discovery tool. And when we talk about AI chatbots, we're talking not necessarily about AI chatbot on your website, but we're talking about the perplexities and the clods and the others. So, um, one thing I would encourage you from the beginning is this presentation is built on a much longer white paper, which is about 35 pages. You can access it on our website at tech b2b.com here forward slash generative dash engine dash optimization. And it'll take there with a quick overview. And then you can download that white paper for more detail on all the topics we're going to cover today. So what is geo.

Speaker 1:
It stands for Generative Engine optimization. Um Geo prepares content to be discovered trusted and cited by AI engines. Unlike traditional SEO. And if you haven't used a lot of AI engines, you type in a query into Cloud Gemini or to I'm sorry, Cloud or Gemini, etc.. And, um, basically, um, what it does is it gives you citations and it gives you mentions. So that's what we're going for. In addition to in the previous days, it was all about search engine return position, SERPs. And what we're looking for now are numbers of mentions and citations in AI answers. And if you've used, say, Gemini, which is the most common, or Google, you'll see the Gemini response often at the top. And each one of those paragraphs often has a hyperlink. And that link is linking back to a site that it pulled that information from, and then is offering that up to the surfer, the web surfer, or potential customer to say, this is a place you can go learn more information. So that's ultimately what we're going for. Geo optimizes for AI discoverability, as we talked about. Now we're targeting citations and not clicks in this particular case, although ultimately we do want the click. We want the backlink back to our website. That's what we're going for. And of course Geo is designed for LM driven systems, this presentation presents a four pillar blueprint for mastering GM. This framework is designed to position our clients as the definitive, authoritative source within the industry, ensuring the expertise is not only visible but becomes a foundation for AI generative answers and responses.

Speaker 1:
And we're going to go a lot more into that, that what we're going to talk about is rag retrieval, augmented generation, which is the way that search engines go out and find information in the great big world. The pillars of the strategy are one verifiable authority, which are built on the principles of expertise, experience, authoritativeness, and trustworthiness. Eeat. This is a commonly used. Google has been using this for a number of years to help evaluate who is an authoritative voice when it comes to covering or has information about a particular topic. The second pillar is topical dominance achieved through the strategic development of of comprehensive content hubs. And the third element is machine comprehension, which is enabled by meticulous content structuring of our website. And we're going to talk a lot about spoke and hub strategies. That's a key concept when we come to optimizing our website. So keep that in mind. And then fourth is the knowledge graph imperative, which is executed through the technical implementation of structured data. That's structured data, both on our website as well as it incorporates certain external website components such as backlinks, which is something we've all traditionally kind of put on the tail end of the dog, so to speak, when we're positioning a website, websites today are basically judged by coding content and backlinks, backlinks being other websites that point back to us.

Speaker 1:
And the more authoritative and the higher the traffic of those websites that point back to us, the better our rankings go. In traditional SEO optimization or Serp positions. There's a similar paradigm at play when it comes to geo, and we'll get to that in greater detail. All right. For SEO to geo SEO is all again about rankings and traffic. Um. A E or answer engine optimisation are short answers and search return pages. Recall that I mentioned today. If you're doing a Google search, you'll often see the AI response at the top. Um, that's what we're talking about on the short answers. Um, back up real quick. Geo is a holistic approach that includes SEO a e o and LLMO. So search engine optimization, answer engine optimization, and large language model optimization to influence the full AI answer pipeline. Fancy way of saying the AI summary that comes back to the searcher. To address this, do AI driven conversational search paradigm, we need more sophisticated and holistic strategy. Generative engine optimization Geo is the discipline of structuring and optimizing a company's digital content and data to be effectively discovered, parsed, understood, and cited by generative AI answers. If you're familiar with the concept of schema, which is a markup language that is applied to the best websites today, that also provides a lot of contextual information for traditional algorithm driven web search.

Speaker 1:
And what we're talking about here is, is similar in many ways. Um, schema is heavily leveraged, but basically by connecting various pages and topics built around a series of cluster topics. And these cluster topics are often core to our business. So our service groups, our industries, our products, whatever it may be, right? We're going to usually build these around our revenue centers. Um, the links of the, all these subpages from knowledge base articles, how to articles, videos, blogs, you know, we're going to move more to a question and answer format instead of long form factor, technical white papers and things for websites. We'll talk a lot more about that. But it's this cross linking these internal webs on our website that really form the basis for our LMS dot txt file. And that's what AI needs to see to be able to offer us up as an authoritative voice. Um, so as we mentioned, search engine is about rankings and traffic and search engine. It's a subset that talks about the direct answer where the LLM is not only using its knowledge of language to help determine the intent, the actual intent of the searcher who may not really know the best answer or the best question to form, but it also is going to use those lmos to generate an answer which is conversational in tone, and much easier for most folks to understand.

Speaker 1:
Which is why answer engines are going to supplant traditional search. And finally, large language model optimization, or LMO. This concentrates on making content easily retrievable and extractable. And we've been talking about that. Um all right. So geo is the overall umbrella all this. Now let's get into how AI synthesizes content. Llms like ChatGPT don't rely on memory alone. They retrieve live data before answering, avoiding the information cut off of search returns based on old or incomplete data. Information cut off means you write something about the latest of autonomous mobile robotics, but it's six months old or 12 months old. Traditional search engines don't really pay attention to when that page was born, but AI does. And we're going to be. And every time we update it, we're going to update the dates on each one of these pages and also things that we didn't care about before. Like we're also going to add biographies of the author and that person's biography and their personal network and affiliations is going to be very much critical to AI to say, hey, these guys know what they're talking about, but we'll get a lot more into that. Now, at the core of the answer engine response is the retrieval augmented generation or rag. Rag includes two primary steps. First is retrieval. When a query comes into an to A to a clot, or a ChatGPT. That's when you pose a question to it.

Speaker 1:
We'll call it the query or the prompt. When the query that searches for an it first searches before it searches an external knowledge base, such as the live web or companies documentation or specific database. Um, I'm sorry. When it first gets the retrieval, it first gets the query, it goes out and it searches for the web and company information, um, that are relevant to the query. The second step in this, in the retrieval augmented generation, the first step is retrieval. The second one is augmented generation. Retrieval information is augmented to the user's original prompt, both providing the LLM with rich specific context and augmenting the search. Return to a more conversational, easy to understand response that we discussed. This model uses its combined information to generate a comprehensive response. The critical takeaway for any geo strategy is that the quality of the final generated answer is entirely dependent on the quality of the information selected during the retrieval stage, which puts the onus on us to make that information readily available in bite sized formats. For the answer engine. The AI system must first decide which documents are relevant, trustworthy, and structured enough to be retrieved. You know, we've been talking about structure a lot already, and we've been talking about some of those elements that determine trustworthiness, the backlinks of high, large volume websites that have high authority, having bios, having date stamps on there to show that the information is fresh.

Speaker 1:
This selection process relies heavily on signals of authority, clarity, and relevance, and the content is ambiguous. Poorly structured. Keep thinking on that. When we talk about there's both content structure. When you're like in a Q and A or an FAQ format, or ten steps to consider or ten steps to implement something, that's that structure that AI can easily understand, as do people online. I mean, you think about the social posts that people click on five ways to make more money. That stuff sells. So basically, AI is reflecting our interest as humans because it's learned from the language, from our how we think, how we construct and how we communicate. Um, it will be filtered out by the retrieval system if, if it's poorly structured. Uh, consequently, battlefield visibility and AI search is won or lost at the retrieval stage, the primary object of his goes to transform a company's website into a highly retrievable, authoritative knowledge base base that AI systems preferentially select as a source of truth. Now, topical dominance, um, to be recognized as an authority by an AI brand must demonstrate comprehensive knowledge of its core subjects. Sporadic blog posts and disconnected topics are insufficient. The most effective strategy for signaling deep expertise is the development of content hubs known as topic clusters. Content hubs topic clusters interchangeable. This model organizes content creation and central themes that are important to the company, their industry, and their customers, creating a dense, interconnected web of information that's highly valuable to AI systems.

Speaker 1:
You sell robotic hardware and software, right? So you're in the business of providing robotic solutions to an industry. So, you know, as we develop these clusters, we think about that, we think about it and say, what's our core element for robotic solutions for industry? Okay. All right. So there's a, there's a cluster topic now as a subunit of those different cluster topics that may also be relevant would be how to design an efficient or how to design your first robotic solution, how to design a robotic work cell for food and beverage or for logistics. Right? That's reflecting both our products and our services, as well as optimized for very specific application, right? So we go from the very broad and general cluster topic to other key cluster topics that represent usually built again, around our revenue centers inside of our company. We focus on core subjects, as we've said. We interlink content logically. This is critical. Uh, I personally, when we first started doing geo analysis for tech, B2B marketing, for example, we used a new tool that we'll talk about in a little bit called joel.ai. It was developed by Kevin Vincent, my head of IT development here and this, which by the way, you can find online geol.ai. Now, uh, one of the things that we did was to do an analysis of our website. And, you know, there was 120 pages existing on the website, right? Because we don't have e-commerce and all this stuff.

Speaker 1:
So it's pretty common size for, for an agency like ourselves. Half of those were interconnected to various topic clusters and services. The other half were not, they weren't orphan pages, which is interesting. And oftentimes they would have been linked unidirectionally say a blog or a knowledge base article or a video to a specific service that the company offers. But now we need to bidirectionally link, right? Our, our schema and our internal links have to be much more comprehensive and not an afterthought. This is going to be a key component to how we organize our content, and then ultimately how we create that roadmap for AI answer engines in the LMS dot txt file, we need to, as always, match user intent. And then when we come to the content strategies themselves, the structure of an effective content hub is built on a pillar page and a cluster content. Now the pillar page is a centerpiece of the hub. Um, again, if we talked about how to deploy your first robotic automation system, for example, that might be a long form pillar page. And then from that there will be elements on hardware and software, on maybe specific industry implementations, etc.. And all those bidirectional links will help the AI to understand what is a key, uh, cluster topic, which also will be clearly defined in our LMS txt file, but also how all the different supporting elements go into it.

Speaker 1:
Now when it sees, you know, someone types types in, how do I develop a robotic solution. It's going to look all over the web. It's going to look for people who are whose clusters. Cluster topics, um, and hubs, how they're the one who really addresses that specific question. It understands variations of it, of course, but it's looking for people who are really focused on that. And then it's going to say, I'm going to look at all. That's our that's the hub of our wheel. Those cluster topics of the hub of our wheel. Then it's going to look for all the other pages that link back to it and bidirectionally to it and say, now I can understand it says, now I understand all the components, all the little parts that go into solving the big question. So, um, the pillar page is a long form centerpiece. It's a long form page gives a broad overview of a main topic, like a complete guide to industrial automation systems. And it acts as a central anchor for all related content, the cluster content or spoke page. The cluster is a series of more specific, in-depth articles, uh, each of which explores a distinct subtopic related to the pillar. Each of these pages links back to the central pillar page, creating a tightly knit structure. Individual articles on PLC.

Speaker 1:
Programming fundamental SCADA HMI. A well executed topic cluster creates a powerful semantic network for an LLM analyzing a website. A dense and logically interlinked collection of pages on a single topic is far stronger. A sense of authority than dozens of disconnected articles. And back to my original point. When we were looking at tech again, half of our pages. They weren't orphaned, but as far as AI was concerned, there weren't bidirectional links. They weren't built around cluster topics enough, and it basically was willing to ignore half of our website. So this is a critical step. We link all of our spoke pages back to the hub. You can see in the illustration here cool visual representation of it. And we use descriptive anchor text. So um, think of this as our keywords in our search terms. Right. Uh, if we're smart, we're, we're optimizing, we're talking about our services and our products in terms keywords or search terms that are commonly searched for and sought out by our customer or customer group. Um, that doesn't change. And but what does change is that, um, we may have a blog that covers 4 or 5 that has relevance to 2 or 3 cluster hubs or hubs in, on our, in our website, right? We need to make sure that we're using the same terminology for each one of those cluster hubs that we're building around and that we're using that same phrase inside this article so that it's, and that those, that text is actually hyperlinked back to the hub on different ways and vice versa, uh, from the hub to, you know, one hub can actually link to an infinite number of pages and, and a, a page, as long as it's relevant to multiple hubs can do the same thing.

Speaker 1:
So a lot more hyperlinked text on your pages. You're going to see a lot of that. And this gets back into the micro schema a concepts that if you're not familiar with, we'll talk more in a little bit more in more detail here. Now as we're building our content hubs, we have to get into the how tos, right? Um, uh, for the, the steps for building a content hub are identify core topics through brainstorming and market analysis. Identify the high level subject areas that are most critical to business and target audiences. And most of your SEO efforts have already identified this. And theoretically, your website should be structured similar and around those topics. So again, our SEO strategy is not going to mirror our existing website, but we sure hope we've developed our website based around customer topics that are critical to our customers. And, um, the people who have done the better, folks who have done that, the less work and lift they're going to have to do to get geo optimized. Uh, but you know, in our experience over the last 30 years, uh, you don't see a lot of good company.

Speaker 1:
First of all, it's, it's tough to find time to build more content and continually add to your website. That's one thing. And the second thing then is it's hard to have 1 or 2 dedicated people who are just simply looking at all these new pages that are being added and make sure they fit into a larger micro schema strategy. So the truth is, is that 98% of companies in the world are really going to be starting from scratch when it comes to geo optimization. The second part is to conduct deep keyword and intent research. Use SEO and AI tools to map out the relevant subtopics, long tail keywords, and specific questions that users are asking related to the core topic. This is loosely defined as search intent, which doesn't just mean what they're searching for the actual words, but since the user may not know the right search phrase to put in. We can leverage the LLMs um, intelligence to be able to identify. All right, we know we're interested in robotic work cells, but how are people out there searching for that? What are the questions? Right. Um, traditional SEO would just tell you long tail and short tail phrases related to this. It really wouldn't go into, um, a lot of hypothetical, the ability to provide conversational questions that people often type into their search engines when they're looking to learn more about a particular topic.

Speaker 1:
Geo and AI search engines and search engines do that very thing and have that capability. Third part map the cluster and the interlinking strategy. Visually plan the hierarchy of the pillar page and all supporting spoke pages. A deliberate internal linking strategy using descriptive anchor text is crucial for guiding users in AI crawlers through the hub, and signaling the relationship between content pieces between our hubs and our spokes. Interconnection between our hubs and you get. Okay, now that's going to move on to the next pillar machine comprehension. Um, while often, well, humans can often infer meaning from poorly formatted text, machines require clarity and structure. Llms have come a long way to be able to understand our intent and who we are, but they are not aware. Let's not kid ourselves. They are not aware. So they're still looking for giant patterns. Is that the patterns that they can spot in a world wide web are much more complex. You know, nearing human comprehension. Um, but a far faster processing speeds than the human can provide. Uh, LLMs analyze the hierarchical and formatting cues of a document to understand its content and identify the most important information. Optimizing on page structure is therefore essential for machine comprehension as a foundational element of geo. And you'll remember our fourth pillar was knowledge graph. Um this is this schema element is helping. It will definitely feed into that fourth pillar of knowledge graph building by using Content Hub with a pillar page that addresses the big question and spokes that dive deep into specific elements.

Speaker 1:
One of the big questions you create a content strategy that needs to be deployed through all of the hub pages. The specific elements of the big question you create content for breaking up long texts into single topic paragraphs using inverted pyramid style writing with our most important topic and largest topic at the top, and then components supporting arguments and components down lower. This is very important for that. Lists and Q&A formats help the AI engine understand what needs to know about specific topic, while also providing bite sized questions with answers which. If you've noticed in your. Your answer engine search returns from Google. That's exactly what it does. It's like, all right, so you're asking this general question and then, you know, here's an answer. And if you went to those websites cited there, you're literally going to see a variant of that question that you initially put into the answer engine search with a short answer. All right. So it's not when we talk about retrieval augmented generative responses or rag responses, um, you know, it's, it's working from what we've already done. All right. Yes. It's generating new phraseology and responses and everything like that to put it into a conversational tone. But it needs us to break down complex topics into logical, smaller questions and components.

Speaker 1:
And that is going to be a core part of everything that you're going to do for geo optimization. And then, of course, we want to lead with key insights, right? So that's the pyramid style we're talking about. Listen Q and A's. So AI friendly content formats. There's a lot of things we can do just with standard structure. You remember we up till now we mentioned coding content and backlinks for traditional SEO optimization. Coding is not. It's about render times. You know, they have to be below three seconds. We need H1SH2S. We need alt text. These are H ones. Are your primary headlines H2S. It obviously gives more weight to the words that are in those formats that using those formatting codes for on the page and on the page, they appear as more bold and highlighted headlines kind of deal. So that was important for SEO, and it's still important for geo, but there's a little bit more we're going to add to it now. We need a logical heading hierarchy. The page should have a single H1 tab for the main title. Standard SEO process. Subsequent sections be organized with logical nesting of H2 and H3 tabs, similar to what we're already doing. And this hierarchy provides a blueprint for comprehension, allowing the AI to understand the flow and relationship between concepts. Using multiple H1 tags is always a bad process in SEO, and it's still a bad process in jail.

Speaker 1:
We need scannable and parsable formats. Utilize formats that break up text and structure information clearly. Bullet points, number lists, comparison tables, and well defined Q&A sections. I've already talked about how answer engines basically appear to be writing complete and original answers to us, but in most cases, it's pulling that little paragraph from a paragraph from a website. And that's why we want to keep those paragraphs short. Key insight front loading. Again, inverted pyramid. We touched on that basic answer. You know, the H1 is our general overview question, and the first paragraph should summarize everything that comes below it. Why? Because the answer engine is going to pull when it finds this page. And if it if it likes your experience and your expertise, your authority and your trust values, your e t values. And it says you rock, you're an authoritative voice. It's going to probably pull that full summary, full first summary paragraph inverted pyramid. Inverted pyramid is critical. Um. Let's see. Oops. Uh, okay. Last point. Conversational language. Question based headings. Write in a clear, direct style that avoids unnecessary jargon. Frame headings and questions that mirror natural language as much as chatbot query. It's okay to say SCADA. Okay, SCADA is an acronym, but it's incredibly. It's very well used, You're right. It's well known. It's a word as far as AI is concerned. Now let's get in more deeper into the structured data.

Speaker 1:
Things that we've been talking about. If content is structure is if content structure is the blueprint that AI engines are going to read, then structured data defined as schema.org markup, which is invisible to most people looking at websites, unless you're looking at the code, is the explicit machine readable language that translates that blueprint into a format AI can understand with perfect clarity. You can. I'm sure all of us in the marketing department know what alt text is, right? You have an image, you have to have a text image. And that's so that blind people can use different machines and software to know what was what the picture is about. Okay. Um, not having that is a knock against your traditional SEO efforts. Um, schema goes much farther than that. It's not just about images. It's about sections. It's about tables, it's about PDFs. It's about everything in the giving this contextual information to every asset or. Important section of content on our. On a page or on a website. Um, this is arguably the single most important technical component to successful geo strategy. If I haven't hammered that home schema markup transforms a website's unstructured text into a structured knowledge graph. Our fourth pillar of interconnected entities that becomes a foundational data source for advancing AI systems. You'll remember Knowledge base again was the fourth pillar. Uh, how schema.org markup works. Schema markup is a vocabulary of tags added to a website's HTML, preferably the JSON hyphen l d format, which is widely known today to be one of the latest, greatest options for for schema markup.

Speaker 1:
That this this this text, this um, this schema information explicitly defines the entities on a page such as the company, what the specific product is, who the author is, or what the event is that the page is about. Stop hitting the doggone buttons really screws us up. For example, instead of AI having to infer that John Doe is the author of an article about product X made by company Y, schema markup can state these facts explicitly, and the link will be linking from John Doe's name to his bio on our team's page to validate that this guy knows what he's talking about. We're going to have that timestamp on there. We're going to link from the product reference in this particular article back to the actual product page. Maybe that's a hub into itself, or maybe it's directing up to a larger hub to show that, you know, maybe this robot controller is part of a robotic hardware solution. Okay, so there's a number of different schema types that can, and we'll go into some more examples about what kind of information is in here? Um, so now the organization defines the company's unique entity connecting its official name, logo, address, contact information, and social profiles to a standard, a single canonical sense of truth. Source of truth.

Speaker 1:
Uh, canonical is an organizational structure where you, you establish the primary, the company name. And then here are all the Subelements, much like our inverted pyramid style product defines product detailed structured information about our products, including features, specifications, pricing, and availability. Again, all this information is not going to be it'll be displayed obviously on the website for our customers, but schema is defining it in Unrendered text for the machine. We don't need to worry about fluff or anything else. It's just the facts, ma'am. And that's why AI likes micro schema. The person itself establishes the identity and expertise of authors and SMEs, linking them to credentials, affiliation, and directly supports EAT FAQ pages. How to. Articles, organized instructional content and to clear step by step process that AI can easily parse and present to users seeking guidance on a task. And then any articles provides metadata about article type schema markup is. It provides metadata about the article itself, about a blog post, about a news article including the author, the publication date, the headline, the reinforcing credibility and freshness. You'll notice that when we update an article with the latest, greatest information, we're going to change that timestamp on that blog or that knowledge base article, right? We're going to have to change in our own micro schema to keep that in mind. Stay on top of answer engines. You know, folks, this is critical. We just saw that 25% this year of all search engine traffic will be generated by AI engines.

Speaker 1:
And we know that Google and others are getting more. Um, they're holding their search position data tighter because it's this giant competitive race to see who's going to win the AI search, um, commercial marketplace, right? So we don't have as much information coming from the big search engines out there about how we rank against competitors, although there's still a lot of good tools that answer that question. Um, so implementing schema couldn't be more critical and it needs to be across the site. And my point to, to answer engines is that unlike traditional SEO, where your rankings or your authority might be determined every few months, something like that AI driven search engines are basically evaluating information constantly. And by the way, when you send a prompt to. So when you send a prompt into an answer engine and you're, um, you're, you're asking a single question. Now the answer engine generates a prompt fan on the back end you're not aware of. It takes your question and it finds hundreds or more of related questions. And it sends all of those questions out to its large knowledge base, graph its database and to the web in real time retrieval. And it's looking for answers for all of those subquestions. And then it's using micro schema and other criteria to then prioritize the people who are, who have information on those various stories, on those various questions, and then it generates a comprehensive summary response.

Speaker 1:
And that is how retrieval augmented generative responses work. That's how answer engines think. They don't just take your one question, shove it into their database and go search the whole wide world for it. No, no, because statistically speaking, there would be. The answer would be poor quality. It's much better to ask the similar questions. Hundreds of versions of the similar questions. Take all of those search responses and generate and distill it into a summary. And the likelihood that the AI engine is going to be on target is going to be much more effective. That's how answer engines work. Okay. Um, now we've mentioned knowledge graphs and artificial intelligence many times at this point. Uh, while some analysts suggest that rag systems may not parse schema org markup in real time for every query. But again, don't think that it doesn't already have AI training database as a constantly evolving model of the World Wide Web, or at least segments of it that it believes are the most important for answering different types of questions. As we go farther, you're going to realize that different large language models are better at different types of questions and queries, and it's totally based on what that model was trained on. Was the data trained on Reddit and Quora and Wikipedia? Was it trained on, uh, corporations? Was it trained on the news media? Was it trained on social media profiles? Each one of those is very much going to skew the operations of that particular large language model.

Speaker 1:
So again, some folks would say, well, the micro schema not rendered rag search from an AI answer engine, it's not going to read both everything on the page and all the micro schema markup. However, this is overlooking the point because it is using schema when it's constantly scanning the web and constantly optimizing its model for who's the smartest, latest, best authoritative voice on any particular topic. Uh, micro schema is multifaceted. It significantly boosts the page relevance and authority signals for traditional SEO, making it more likely to be selected by the rag search. Second, and more importantly, micro schema markup is the primary mechanism for feeding accurate, real time information to the persistent, large scale knowledge graphs like Google's that AI systems use as a foundational source of truth. By implementing robust schema, a company is effectively pre validating its own information within the AI's long term memory, and this increases the trust score of the content and the probability of citation making schema. A powerful long term authority built asset, not just a real time parsing instruction. And one of the most important takeaways here is that when someone queries an answer engine about a topic, okay, it's not learning about that topic at that moment. All right. It's already considered this as a topic and many, many thousands of related topics to it.

Speaker 1:
And it's already been filtering and searching the web for people who are very knowledgeable about that, and it creates this large knowledge graph that helps it to understand commonly asked questions, commonly searched for topics, etc.. Okay. Um, our team also reduces hallucinations and hallucinations are basically responses from the AI when it didn't have a big enough search model, when it didn't have enough data and therefore it was fooled. It was fooled by thinking that a couple of social media posts, for example, on quantum computing, were actually, you know, the end all be all answer. Right. And that's why you want to really look at that answer engine. Which one? Where is it being trained on? What are the sources of information on the web that it's being trained on? One that's really good in social media might be incredibly great at crowdsourcing information about customer intent, customer desires, what's hot? Um, things of that nature. Is it the best source of scientific truth? Huh. I'm gonna let you guys decide on that one. All right, now let's go into part three, a phased implementation of our roadmap and building your AI optimized content ecosystem. Okay, so transitioning to an AI first content strategy requires a structured, methodical approach. As you can see, there's a lot of moving elements. We need a plan in place and we then we can measure against it and we can continue to judge our progress.

Speaker 1:
Right. Um, this is not a gut solution, folks. All right. This is a very methodical technical approach to how to organize the content on our website. This five phase roadmap is designed to build momentum, deliver incremental value, and ensure that all foundational elements are in place for long term success. It integrates technical readiness, content strategy, authority building into a cohesive, manageable plan. So phase one, we develop, we audit what we have, develop a strategy. Phase two, we evaluate our technical readiness of our team and our and our capabilities in-house. Phase three, we develop those primary content hubs often built around our revenue centers. Phase four, we've got authority building. This gets back to the EAT thing, and this is the Bios. This is the cross linking. This is the micro schema. Um, you know, structured data Q and A's um table comparison tables, bite size information that AI engines can easily assimilate and include in part of the answer. And then phase five monitor and improve our efforts. Right. So it's one of those things we must monitor. Otherwise we cannot manage. And we certainly can't improve. This seems like a good point to talk about the custom tool that tech B2B is Kevin has generated. For us, it's called we call it Joel. Joel. Ai and if you go to geol.ai or scan the QR code on your screen, you will go to one of the world's first tools for doing competitive analysis on answer engine search results.

Speaker 1:
This thing will look at your current website. It'll talk to you about where your your coding is good and what you're missing at a foundational level. It will allow you to put in, um, to build a series of questions like a library of questions that you think your customers are most interested in related to your products and services. It'll suggest optimization for those questions. It'll manage whether how many citations and mentions you get from those. And then it'll also show you who is getting more of them. Our competition. Right. And ultimately, this thing is going to be able to, uh, not only it'll compare us against the competitive landscape out there, but it'll also suggest ways where we have gaps in content based on where we don't have clear question answer formats for commonly asked questions. Okay. And then in the end, ultimately, it'll also be able to help you generate content from a variety of sources original unique content using AI power as a starting point for how you might develop content that would fill that gap. Okay. So Joel is going to become hopefully what we consider. We hope to be one of your closest companions as you go into the answer engine age. Okay. Phase one you recover was discovery and goals. So our actions comprehensive SEO and content audit. I think that's, you know, pretty self-explanatory.

Speaker 1:
You look at our current website evaluating organic search performance, backlink profiles, existing content inventory. Think about all the PDFs, the individual pages, the knowledge, the videos, etc.. Um, and then technical health, of course, you know, like, for example, Joel or your SEMrush RF will tell you right now how many orphan pages you have, how many times you're missing schema or alt text or more Similars missing H1's. H2's, etc.. Competitive geo analysis. Um, manually query every major AI platform like ChatGPT, perplexity, Gemini, etc. using core commercial informational keywords relevant to a given industry. So basically, we come up with a series of questions that we think people would ask if they were going to find our products and services, and then we go out and we ask the engine all these different engines. I'm going to ask this question where, you know, who's being cited and am I being cited at all? And then but that's what Joel's going to do for you, by the way, that it automates that whole process. Otherwise, imagine that you've got it's not a common. The baseline is like 50 questions. You could have hundreds of questions that you think are relevant for your company. Imagine if you had to query them every single week, because that's how fast the AI models change 50 or 100 times, times 7 to 10. Primary AI answer engines. We need a tool. Hence we can make a joke.

Speaker 1:
Um, so we're going to benchmark the AI citations from our geo analysis. Now we're going to do a team capability assessment. We're going to look at you know how many folks do we have for generating content. How many, what assets do we have for folks to find all those bidirectional links? Do the micro schema. We have to reformat some of our most important pages to make sure that we've got long form pages for our core hop topics, our core hubs or topic clusters, right? That's in some cases, that's going to be one thing that, you know, you're going to have to build because we're really good as technical B2B providers, you know, companies, manufacturers, service providers. We're good at knowing our business, we're good at knowing what we sell. And we know it's important, but we also assume that the customers who come to us already have a general idea of why it's important and how the parts fit together, right. Um, in some cases, we'll write blog articles about those. Well, we might be taking some of those general primers right from our blog. Moving that over a pillar page, looking at all those pages, seeing what topics are deep dives, adding a ton of bidirectional links between those articles in our knowledge base and the large tiller cluster topic, taking some of some of those best of those blogs that might do individual that are deep dives, taking some of that information and putting it onto our long form so that it is comprehensive, like top ten steps when developing a robotic solution for an example.

Speaker 1:
One, two, 34567 ten. And maybe we've covered all ten of those topics in our existing knowledge base. So, um, that's part of I went back into content, but we need to see what human assets we have to make this happen. We need to set measurable goals. Now these are, you know, pretty common increase brand citations and AI overviews in Joel, for example, uh, for industrial sensors topic by 50% within six months or grow AI referral traffic from perplexity by 20% over the next quarter. You get it. And by the way, all of our GA4 and capability. That's where we measure the referral traffic. So we know if search engines, cloud and others are sending us traffic, that's so on the you know, and Joel can tell you that too. If you connect it to your GA4 situation or actually, um, that's a whole nother discussion, but you'll be able to, with your existing infrastructure, know how much traffic you're getting from these referral engines. And you can go look after this meeting and see if you're getting anything. Remember, 25% of all search traffic right now is coming from them. It's 25% of your search traffic coming from answer engines TBD milestone. So, uh, we need to finalize our SEO strategy document. That includes our baseline metrics, our competitive analysis, our defined KPIs, and a detailed project plan for the subsequent phases.

Speaker 1:
Phase two technical Readiness. With a strategy in place, the focus shifts to preparing the technical foundation of the website and beginning a process of content alignment. What are our specific actions? Uh, technical SEO remediation. We need to fix our site errors and our performance, uh, crawlers. And you can use a SEMrush or an RF or Joel to help identify these challenges. We need to do initial schema deployment as we've talked about. Remember Json-ld. Content audit and triage. So we're going to have a lot of long form content that right now has a bunch of h2's that we should probably start putting some of those h2's as questions as subquestions to that particular section of our article. And now we've got the structure of a commonly asked questions or FAQ. Right. So we can turn white papers into something more bite sized. And as long as we're using proper formatting, you know, an overall topic header for H1 reflecting of our content cluster topic. Right. Right. We want to have that, make sure that those words align with its primary content, um, hub. And then each section in our FAQ are going to be, uh, hopefully offered up by our, by answering engines as part of an overall, um, overall, uh, retrieval, augmented generated response. Okay. Uh, and we're going to need to restructure some of that content as we, you know, the formatting, the questions, make sure we've got our H1, H2, and H3 in order.

Speaker 1:
You get the idea. Phase three content execution. This is the primary execution phase focused on building out of the deep, authoritative content that will establish topical dominance actions. Pillar page creation. We've already talked about that cluster content. Systematically create and publish detailed spoke articles that form the content clusters. Each piece must be high quality. Address specific user questions include robust internal links to the pillar page and other relevant cluster content. When we see other relevant cluster content, think back to the schema definitions about what our company is about, what a particular product is, about the Bios for whoever the byline person is. And let's hope that bio person is someone who is part of I, triple E or A3 or other primary associations that prove they have authority. And let's think about this people. All right. If we're we've got engineers on staff. Traditionally, a lot of wisdom has been. I don't want to publish the fact I have this. Engineers. People will poach them. Have you heard of something called LinkedIn? Okay. Those days have come and gone. What's more important now is if you're a member of an association, and that association has the opportunity to have a member directory, you need to fill that out. And it needs to be braggy without being obnoxious. And it needs to link back to your website on your team page where homey has his bio.

Speaker 1:
And then we're going to link all of the articles that these people engineers write back to that bio page in the micro schema document. Bingo. Now you're starting to see how the GI strategy is coming together. Uh. All right. We said that's the advanced schema types. Um, which we've gone through the different types that exist. And we need to create those and add that content. And we need to start and finally create and deploy our LLMST txt file, which summarizes our cluster topics and all the spoke elements, right, as well as other critical micro schema assets that we talked about before. The first complete AI optimized content hub is live on the website. At this point, it's fully enhanced with advanced schema and guided by corresponding LMS txt file. Um, as I think I mentioned, I'm going to add this in there. The LMS acts as a table of contents essentially for our website that tells answer engines this is who we are, this is what we're about and what we do. And here are all the things that are important to the those things that we do or content hubs. Okay. All right. Number four, build external authority. We already talked a little bit about this by making when we suggested that our key personnel, uh, who are key because they're well educated and trained and that they're, they're members of certain associations and contributing back to the larger community and that increases their authority.

Speaker 1:
Hey, don't forget, I mean, it's smart in your business to have to have some of that technical. Those people work on various committees and boards and mean something, and it's going to mean something more now with answer engines. Right. So those outside links is the point are critical. All right. So now that was back links back in SEO days. It's still important although it's going to be even more important, um, for us to, to really be serious about this and to really realize that, for example, sending a press release to a small trade publication that maybe has 100 or 200,000 site visitors per month is not going to be nearly as big as being able to somehow get your corporate company onto Wikipedia, for example, or to have a Quora or a Reddit profile, for example. These are the three most commonly used websites and platforms today to train LM engines. And therefore, if you have a profile there, the LM engine is going to AI answer engine is going to give you preferential treatment. And don't kid yourself if you haven't tried to put a Wikipedia page for a corporation online, it ain't easy. Okay. Um, you drafted it goes through a whole editorial process. You have to have all kind of external validation. You need to be giving awards by primary industries.

Speaker 1:
Can't be paid. Um, prefers to be like charitable standards development organizations or industry associations over even traditional media. But if you're going to have traditional media, it's not so bad. But AI is going to say, all right, was it, um, was it e online who has 80,000 site visitors a month? Or was it Forbes who has 1.2 billion? Right. Um, and just for the record, we were just talking with Forbes the other day about because you can use custom publishing to get various profiles and get coverage and some of these main publications, 25 grand, 25 grand for a single citation on Forbes to link back to your company. So, um. Okay. Linkable assets. All right. So digital PR and linkable asset creation, unique high value content assets that are naturally due to attract backlinks. We're talked about. Now we're just talking about backlinks and external authority strategic outreach and inclusion. This is the create linkable assets. Proactively seek inclusion for the company, its products and its experts in high ranking list articles. Uh, Quora, Reddit. You get my drift. Now, third party platform optimization, LMS use information on trusted third party platforms as part of their evaluation process. This is the Reddit and Wikipedia we talked about. And then the milestone here should be that we see a measurable increase in the number of referring domains from high authority sites. As we do this outreach, and the AI engines realize that we have.

Speaker 1:
And they will we are, by definition, going to see increases. We should see commensurate increases in referral traffic from those AI engines to our website. So that's our KPI for this type of activity. One of our key KPIs for that activity. Phase five monitor and improve. Now Geo is not a set and forget the landscape is dynamic requiring continuous monitoring adaption. And I think I mentioned in the world of SEO, we had it easy. Okay. There really wasn't a big changing in recognition of a particular page or topic. There wasn't this, you know, super comprehensive search of hundreds of related questions to answer a single topic. They're just looking for phrases and similar phrases that appear in nearby context. That's the pattern that the algorithm was searching for. Now it's searching for much more complex patterns. So point is, is that the geo knowledge graphs are constantly changing. The AI knowledge graphs are constantly changing. Um, so this will not be something that we do once a year. This needs to be built into our regular ongoing activities if we really want to dominate the space, it's TBD the business consequences of that, because this is the first time that AI has been widely used. Again, we're basically two years into AI search, and in year two, 25% of all traffic will go through. So we're going to track those KPIs that we identified. We're going to make sure that we have planned for a content refresh cycle, especially on key spokes and content hubs.

Speaker 1:
If there's new stuff there, we need to add it. We need new date codes, guarantee that we stay ahead of the wave. We need a quarterly strategy review to see how our KPIs are reviewing where our gaps are, prioritize what kind of content we need to fill those gaps and move forward. Um, the milestone here is that you should, at the end of this, have an established recurring process for reporting and analysis and strategic refinement that demonstrates ongoing ROI and ensures the geo strategy remains effective in a rapidly evolving environment. All right. I think we're into our final section here. Part four measuring success in navigating the future. Successful geo strategy delivers tangible, measurable results that directly impact business objectives. This section provides evidence of effectiveness through world examples, defines KPIs for measuring return on investment, and offers a bunch of good case studies for you to look at. These will be in much more detail in the white paper that I mentioned earlier. Okay, so we mentioned that we want AI citations and mentions. Those are our new criteria for success. Um, geo performance isn't just traffic, it's visibility, trust and influence in AI ecosystems. And there's four measurable metrics that you should use to monitor your progress, as well as how well you stand against your closest, strongest competitors. Citation frequency direct traffic from LMS using our GA4 connections, reviews, and backlinks.

Speaker 1:
We've you know, it's there's a deeper dive on how you establish expertise and experience and authority and trust when you're doing your E8 optimization. Reviews. Traditionally, B2B companies have actually discouraged commenting on their knowledge bases and resources and not on their social platforms because you can't do it there. But every one of you who has a blog or a knowledge base in your article, generally, that's that's software came with an ability for people to comment on it. And we discouraged it because we didn't want people who might have been trolls in some cases, or we didn't want to spend them. Basically, we didn't want to spend the time and the money to have someone monitor those pages for things that needed a response. Someone had a problem and they were upset and we needed to respond to them or, um, so we just figured it was easier to turn off. Ai is all about giant data assimilation to make sure it has a better truth. If you and I are going out shopping, what's one of the first things we do? We look at the reviews. How many reviews? And then what's the score? What's the average score? This is super critical for most of us when we're going shopping now. It's just as important for our website and us. Now, I don't care if we're selling encoders, right? The least consumer, one of the most non sexy topics on the planet earth.

Speaker 1:
But we need to be encouraging our customers to comment on our website as well as on our social media. Okay. And if someone has a problem or answer, we need to be able to we need to respond publicly. We need to give them a link to the spoke or the content hub that's going to, uh, have the solution to their problem. Ai is going to lap this up like a kitty with milk. All right, so super important now knowledge graph a graph accuracy. This gets back into, you know, times and regular updates, etc.. Now real world success stories. Our white paper includes several case studies on how companies have used geo approaches to achieve greater website traffic, specifically in regards to AI search results and associated KPIs. Um, case study one is an industrial manufacturer had strong SEO. They were positioned well at the beginning. Company lacked visibility in AI generated by enhancing content for AI readability and managing brand presence on AI rhythms. 20 platforms I'm sorry, 2,300% increase in AI traffic and began ranking for 90 key AI overview terms. Now don't you may say 2,300%, you know. Okay. All right. Whatever. Don't be surprised that it's that big because we're starting from scratch. Okay? When everyone's starting from zero, any traffic resents a massive percentage increase. Point is, is that the point here is that progress can be made in a relatively small time frame if you approach it methodically and with intent.

Speaker 1:
Okay. Study two, um, software as a service company, e-commerce, they were struggling to rank for competitive keywords. They used the AI approach and geo approach, push their page from 10th to first position in AI search results, greatly boosting visibility and traffic. Case study three e-commerce brand Rocky brands. By revamping their content and keyword strategy, they got a 30% increase in search driven revenue and a 70% year over year growth. Again, go to the white paper, get some more details there. Uh, geo ROI, we're going to just summarize real quick some of the ROI metrics. We've talked about, a lot of them before visibility and citation, frequency of citation and AI overviews how manual tracking of target queries, um, manual, how, how do you, how do you look at the frequency of citations and AI overviews? You can manually track target queries in AI platform. I hope you don't have another job. Um, but you specialize AI monitoring tools. We mentioned Joel. Ai glimpse AI is another one that can be effective manual querying of chatbots. Chatgpt Gemini again takes forever. The next one is traffic and engagement. We've talked about that. Keep an eye on our GA4 referral traffic from these these platforms. That's important. Um, we've got direct traffic, we've got click through rates from citations. All of these things are relevant when we're looking just purely at numbers. Uh, and then finally, authority and trust, uh, review volume and velocity, how we monitor the number of frequency of new reviews on key third party platforms using clutch Google business profile and other places.

Speaker 1:
Again, it's time for us to embrace our customers publicly talking about us. All right. Faqs. Top ten reasons or ways to do something. That's all fundamental. You don't do that stuff. If you don't organize your house and your website, you will not succeed in an answer engine competition. But if you want to be a standout, we have to embrace the web way beyond our primary domain. Those backlinks, establishing those company profiles, encouraging customer engagement and monitoring those things and responding accordingly. Super critical. You don't need to worry about bad reviews if you respond positively and help people find a solution. Actually, that's more likely to generate more business for your company than simply having no reviews at all. Right. People expect there to be a problem. They don't. They're not surprised by there being a problem. They're surprised when they're disappointed when the company doesn't try to make it right. Keep that in mind. All right. Geo challenges to navigate. Embarking on a geo strategy involves navigating a new and evolving landscape, acknowledging the potential challenges as the first step towards mitigating them as a strategic partner. Our role at tech B2B is to guide clients through these complexities with foresight and expertise. So what are the challenges we have? The evolving nature of LMS.

Speaker 1:
We've already talked about this. These suckers are changing constantly because that's how they work. All right. There's a reason why we're building data centers at, you know, ten every day across the world. It's there's a reason why they're the most high power number crunchers in the planet based on using GPUs. Okay. These guys are just consuming knowledge at a rate that is unheard of in all of human history. Exponentially. And it's not going to change. The risk of AI hallucinations goes up as AI takes engines. Unless we can structure our data better and make it easier for it to get to the truth. That's why it's important to have EAT. If you you will not be part of a hallucination if you do the strategy right. And if you clearly, you know, go step by step to establish your expertise and authority and trust, you are not going to be cited. There's not going to be a negative consequence, right? That's you don't need to see AI as someone who is going to malign your brand. It's not going to do that, but it may boost someone else who's full of junk. If we don't teach AI that it's full of junk. Here's the real truth there is a long term resource investment in this. As we talked about. It isn't going to be over tomorrow. This is going to be ongoing, just like our SEO efforts should have been.

Speaker 1:
This is going to be a little bit more complex, but after the first build. It's not a big deal as you think, right? It's just about identifying key topics, making sure we're reviewing them on a quarterly, bi, bi annual basis to make sure that they're timely relevant, using our tools to identify gaps in our, our, our, our, our, our, our cluster hub, you know, with our content hubs and our spokes, that whole structure need to always be looking for new questions that people are asking and then generate content for those. And then finally, data privacy concerns everyone's, you know, about trackability and everything. This is no, no, no more than GDPR. There's nothing new here from the California Consumer Privacy Act. It's just something that you want to be aware of. And if you've been doing your website, right, with preference centers and everything else, you know, if you're in the US and if you're in Europe and the rest of the world, you're not going to have any other additional worries and concerns there. Okay. So the path forward, um, the transition from a keyword centric web approach to a conversational, AI driven information ecosystem represents the most significant opportunity and threat for brand visibility in a generation. It's not an overstatement. The question is no longer whether AI will change how customers find you, but how quickly you can adapt to this new reality. Proactive engagement with Geo is the only way to secure a competitive advantage in this new landscape.

Speaker 1:
Systematically building verifiable authority. Topical dominance. Machine comprehension and knowledge graph. Imperative authority. Dominance. Comprehension. Knowledge graph. Those are the things we're going to. And we can't really say comprehension without machine comprehension. Right? But those are our four pillars. And the roadmap outlined on this document provides a clear, actionable path to achieving these goals. The journey requires commitment, expertise, and a strategic partner who can navigate the technical and creative complexities of this new frontier. We are ready to be that partner here at tech B2B. Let's begin the work of future proofing your company and your authority today. So I really appreciate your time today. Uh, if you've got any questions, I'm happy to answer them. Uh, if you are seeing this online, you can reach out to us at tech b2b.com. You can get the white paper again at tech b2b.com/generative, hyphen engine hyphen optimization. And feel free to send us some questions. We're going to be launching some new tools pretty soon. Um, that'll be an ability to do kind of a cursory evaluation of your current geo position. So that's something we're looking at. And again, don't forget to go to joel.ai geol.ai. Um, there's not many tools out there online and your SEMrush and HAH ref environments. Um, and don't forget about glimpse AI also, uh, they can help you to, to baseline where you're at today. Key component, as we talked about in our strategy. Thanks again for your time and I'll talk to you soon.

Why are thousands of creators switching to sonix.ai? Because automated transcription, audio and video translation, and AI subtitle generator finally feel effortless. See affordable transcription pricing for the details.

Automatically convert your m4a files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.

Thinking about switching tools? Our VEED review and Otter AI vs Rev comparison dig into the trade-offs. Also see our picks for best AI tools for audio and video, Otter AI transcription review, and Happy Scribe.

Whether you're focused on scalable transcription for teams, Microsoft Teams transcription, or live transcription, Sonix adapts to your stack. Our how to make English subtitles and Zoom meeting transcription capabilities are a standout.

Try Sonix free today. No credit card required.