Owning AI Search: Actionable Strategies to Move from Schema Markup to Knowledge Graphs

AI is transforming the world of search, and brands must adapt to stay visible and relevant. In this webinar, Emanuel Petrescu hosts Martha Van Berkel, CEO of Schema App, for a practical session on moving from basic schema markup to building powerful knowledge graphs.

Learn how structured data and entity linking are becoming essential for AI-driven search, not just traditional SEO. Martha shares the latest industry insights, actionable strategies, and real-world examples to help you make your website AI-ready so your content is recognized, cited, and trusted by both search engines and AI systems.

Key Takeaways

• Why schema markup is now critical data for AI
• How to connect your content and entities to build a knowledge graph
• How to prepare your brand for the future of AI search and digital agents

Whether you’re a marketer, SEO professional, or business owner, this session will give you the tools to future-proof your online presence and ensure your content stands out in the age of AI.

 

Emanuel: Hi everyone and welcome to yet another session of the How about Some marketing webinar series. My name is Emanuel Petrescu, founder of How About Some Marketing? I’m an SEO digital marketing consultant from Toronto, and I can’t be more privileged and happy to have a special guest today: Martha Van Berkel.

She’s the CEO of Schema App and an authority on the topic. I’ve turned to her articles, presentations, and everything that she shared for the past couple of years, many times, and I’ve applied successfully in my work and I’ve seen some results. I suppose Martha, on behalf of the entire SEO and digital marketing community, welcome thank you very much.

Thank you.

Martha: Thanks so much for having me. I’m super jazzed to be here.

Emanuel: Excellent, We’ll just give a few more minutes to give everyone a chance to join. It’s the first webinar I’m hosting at noon. I’m experimenting with a different format. The turnaround seems to be okay so far.

If you wanna drop in the chat wherever you are joining us from and what is your job title, and what do you do? I already seen some people putting in some, comments, but anytime, feel free to do that. And I was just looking to see where the audio is to play a background music, but it’s already too late.

Based on that Andrew from LA 9:00 AM 9:00 AM here. Hi Andrew. Martha is joining us from Guelph Ontario, Canada student Italy. We have somebody from Italy. Nice.

Martha: Ciao.

Emanuel: SEO specialist from Montreal. We have Amin, a mutual friend.

You both shared the stage a few weeks back at the Seo OIRL and a former guest.

Thank you so much, Jason from New York City on the same time zone. Excellent. I think we are ready to start, but before that, I wouldn’t be much of a marketer if I wouldn’t ask anyone who’s here, if they’re not already.

Go to howaboutsomemarketing.com and subscribe to the newsletter because that’s where you’ll find previous recordings, this current recording and any other webinars that will put together for. How about some marketing and essentially what’s how both the marketing? It’s a place for marketers to get better at their marketing.

I wanna let you introduce yourself, Martha, if you want, and then just simply jump right into the presentation. And I just wanna remind everyone that we have a chat. So if you have any questions fill the information there. And Martha, you wanna address them at the end or as we go? I suppose it depends.

Martha: I’ve got the chat up, so put your questions in there. If it’s relevant to what I’m talking about. I’m happy to jump in and then if I miss any Emanuel, I’ll make sure that we catch up at the end and, address any questions or comments that the team has.

Emanuel: And we give a shout out to Kevin and Hassan who are here on the backstage and who will be helping us monitor and will be putting the information.

On the screen, the questions that you have as you go. And my suggestion to everyone here is ask questions. This is how you this is a unique opportunity. I can be more privileged. And with that, Martha, turning it over to you.

Martha: Amazing. Thanks for having me. Hopefully you’re enjoying your first morning coffee or your lunch and excited to learn more.

As Emanuel mentioned, I had the chance to speak at S-E-O-I-R-L in Toronto. And I’ll be sharing a little bit of what I shared there around how do you own AI search and how do you learn about knowledge graphs. And don’t worry if you don’t know what that is. We’ll go ahead and cover it.

So hopefully you can see my slides.

All right. So when I introduce myself, I use actually my knowledge graph to introduce myself and a knowledge graph is how things are related and specifically what the relationships between those different things.

So you can see on the bottom, I’m a rower, so I live here in Ontario, in Canada. And so I go out in a single or a double rowing in the summer. It’s just turned to fall here. So my rowing season has just come to an end. And I actually have a technical background, so I got the chance to attend MIT for strategy and innovation.

But I also studied engineering here at Queens University, also in Canada. I am the founder and CEO at Schema App. We’re an enterprise schema markup solution. We build knowledge graphs for a living. That’s what we do. But I actually spent 14 years at Cisco. And so I’m an alumni of Cisco and was really fun while I was there, as I got to be an intrapreneur.

I got to start up some new things there and work with large enterprise and really understand the importance of agility and innovation. And so a lot of what I learned working at Cisco is what we apply at Schema App. I am Canadian and I used to drive a 1965 Austin Healy Sprite, so this is a picture of my actual car.

I used to bomb down the streets of Toronto. And my actual car was in the movie called Losing Chase, that Kevin Bacon directed. Now in North America we have this game called Six Degrees of Kevin Bacon that when you meet someone you can figure out your six degrees of Kevin Bacon. And so good news for those of you that play the game you can now win.

Because I am two degrees of Kevin Bacon. My car and then Kevin Bacon used to drive my car when he was filming the movie, and I love this example of inferencing. So when you can figure out that you’re gonna win the Kevin Bacon game, you’re actually doing that same type of inferencing that search engines or AI does when it actually reads the context of your schema markup that’s on your website.

Today I’m gonna introduce and kinda revisit schema markup, which is the language of search engines, which is what I’ve used to describe who I am here. To help you also understand how you can use this in this new kind of realm of search, whether you call that AI search or like with the emergence of AI platforms.

And so I always like to start with who does schema markup today? So maybe in the chat I know Emanuel, that’s something that you’ve done. There is actually a certification. So someone’s asked about training. If you wanna know more about training, we actually have free training on our website, so training dot Schema App.com.

It actually introduced you to schema markup 1 0 1. It’ll also introduce you to knowledge graphs. So if you wanna learn more you can go learn about that. Great question there. So schema markup’s been around for a long time. It’s not new. And it’s data that goes behind the website in order to explain what it’s about.

And it really is putting it in a structured format. And historically people used to only think of it as driving rich results, so getting kind of visual changes in search. But now I really see it as this bridge, between sort of your brand and AI understanding. So I’m gonna share how you can do it today.

I’m also gonna share sort of some news from the market that don’t just take my word for it, but let’s see, what the big, sort of owners of search, Microsoft Chat GPT and Google are saying about. Now I wanna take you on a bit of a journey because schema markups like my world, this is what we do every day, all day.

And in 2023 there was big changes when they took away some of the rich results. So FAQ and how to, we remember that very clearly. But starting this kind of February, we started seeing them starting to talk about structured data again. And it was really interesting because this is of course following a year of ChatGPT being introduced and consumers starting to search differently. And in February AI chatbots drive higher quality traffic than Google, right? So we saw people and higher engagement from people who coming there. Then in March we started seeing I’ll say general kind of commentary from John Mueller and other spokespeople saying, yes, like we’re using schema markup for generative AI or in Fabrice Canal was talking about like how it’s helping them with LLM understanding, but there wasn’t anything like hard documented yet.

This was showing up at conferences. John Mueller then started talking about structured data, improving targeting. And then finally in May there was official announcements around Google and Microsoft saying that structured data helps specifically with their AI experiences. And Google actually put it on their document of, to like how to perform in sort of their AI experiences.

Again, they didn’t talk about AI overviews specifically. Or talk about AI mode, but just like AI experiences and they were saying like, do structured data, like it’ll help you actually show up. We then saw Bing start to talk more about structured data mostly in the context of Index now, so those of you that don’t know index now is where you notify, it’s a conglomerates and open project to say Hey, this page has changed. So they’re saying now we don’t know just what has changed or what pages have changed, but what on the page has changed. And this of course, I like to think about as like around allowing search engines to be smarter about their crawl budget. Everything’s about dollars, right? It’s just even more expensive now with AI. So again, like how do we inform them sort of what information on our websites are new? Now I always love to bring in the latest information.

Just last week, October 23rd in Dubai, Google hosted their Google Search Central Live. One of the slides was about structured data. Oh, by the way, if you hear me use the word schema markup and structured data interchangeably in the SEO world, those mean the same thing. So structured data can mean other things like putting in tables, but in this context I’m gonna talk about it in the context of search.

And so what they talked about was they reiterated not just that it’s important. But structured data is critical for modern search features. They talk about like you can find different ways and types to do it. Again, schema, APTA has a lot of in, but the last part was new.

They talked about structured data being efficient, easy for computers and machines to understand and help them be very precise. Good for AI systems. Now, I’ve been talking about this forever because to me it makes so much sense. Again, it gives you more accurate answers. It allows them to be more efficient as they’re crawling to get the information they need.

And Google’s now documented it. So if ever you were in doubt, if you’re like, Martha, I don’t do schema markup, I don’t do structured data, I don’t think it’s relevant. Good news. Lots of evidence that it does. Oh, how do I check and verify the schema markup on my website? Good question, Kevin.

So the, there’s like different ways you can audit it. Often a lot of SEMRush and other tools will tell you if you have it, but the easiest way is to just use the schema markup validator. So that’s like probably my favorite tool. So if you go to schema.org, which is actually the vocabulary that describes everything they have a validator and it will tell you if it’s there on that page.

Often if you’re checking a large site. I would look at one or two pages for every page type. And then typically you can start to see what it is. Again, if you’re an enterprise customer and you’re looking to do it at scale, my team would do that for you as part of our sales process. But again, the validator is a great tool in order for you, to do that.

Google Search Console will also tell you if you have rich result eligible pages, but sometimes those won’t show no all the schema markup on it because it’s just looking at things that are rich results. All right. I always like to ground us on who’s dominating search, like who do we care about, who do we need to listen to at this point?

And everyone has been talking a lot about Chat GPT and because that’s like the emerging technologies. And yes, Chat GPT is leading in the aI kind of market and the AI platform market. But Google is still dominating in search. And in fact there’s some latest research I was just looking at from AHrefs, where they were looking at 40,000 websites to look at where traffic is coming from.

And traffic is also still primarily coming from Google. And I’m talking not just like a little bit more, but a lot more like 40% of traffic on websites versus 2% or 1% coming from Chat GPT. And you can verify that on your own site. I know the organizations I work with are seeing very similar, like under 5% of traffic is still coming from AI

Now, how does structured data implement these AI features? And so I have some research I like to share that sort of talks about what is like the measurable impact that we get? So this is some newer research that I’m citing from Relixir from July this year where they looked at no schema compared to different types of schema. Again, I would, we’re gonna talk a little bit about schema being connected and creating a knowledge graph, but you can see they’re actually seeing the citation rate increase. And this was specifically in AI mode different from AIO.

Now when we talk at Schema App about structured data, we’re thinking of the big picture about how it has an impact on your organization and schema markup, if you think about optimizing a page, think of that as this foundational layer at the bottom. So this is when you’re translating or putting this page is about a product or if you’re A hospital, maybe you’re talking about it being a hospital or you’re talking about the authors on your website, the people that are writing. So this is just basic markup on a page. I’m gonna talk a little bit today about the connectedness because you saw in my introductory slide how when I connected and defined the relationships between me, my car, and Kevin Bacon, that you’re able to do that inferencing.

And so the reason I believe why. Google and Microsoft and everyone’s talking about the importance of schema markup in building I’ll say robust and healthy schema markup is because it can do that inferencing. And that comes from doing this thing called entity linking. Don’t worry if you don’t know what it is.

I’m gonna explain it to you today. When we do this entity linking this connectedness between the content on our sites, and we bring that context, we’re actually building a content knowledge graph. And that’s really, think of my introductory slide, but for all the content on your website. And when we elevate to this sort of like I’ll say more robust data layer or more robust data that’s coming through our structured data, we unlock different rooms.

So I call this the value house. We unlock different values. So SEO is the one that we know very well. It’s the one I’ve just talked about, sort of rich results, non-branded queries, AI overview citations. We’re seeing those increase when we do really good schema markups. But there’s actually other areas you can unlock as well.

One is around content insights. Right now people will talk about you need to have topic authority or depth of, content. Good news. There’s actually ways that you can use your schema markup to tell you about that. I’m not gonna focus today’s talk specifically on that piece.

But it’s, I would say like an emerging area. And if you wanna learn more, we have tons of content to read on our website under resources. The last area, and I’m gonna talk a little bit about this, of as I’ll say, like evidence from market that we’re seeing that is around reuse of your schema markup or your knowledge graph to help AI.

And so at the end of today, I’m gonna talk a little bit about the agentic web. I’m gonna talk about how Microsoft is talking about them using structured data, this SEO tactic or this SEO strategy that we’ve been working on to help them prepare your site for agents. So that’s what we’ll talk about today.

But I want us to really think broader, especially if you’re talking to leadership within your organization. We’re getting a lot of questions on what are we doing to prepare for AI or what are we doing to prepare for agents? And really good news, like the work you’re doing in SEO is actually setting you up perfectly for that.

So let’s talk a little bit about what does that look like? And I’m gonna give you some tactical pieces around the, Kevin asked the great question of am I doing schema markup? What I wanna give you today are some examples of, am I doing good schema markup? Because I really think there’s a differentiator.

There’s a piece of content I’ll share in here of research of saying. Like having basic schema markup isn’t actually gonna deliver the results that you need in this new AI place. And so these advanced schema markup practices are gonna help you go from basic to advanced and really help you prepare for this new age of search.

Everyone good? Everyone with me? Okay. So the first one I wanna talk about is depth and breadth. So we wanna think about kind of optimizing and, why this is important is because, if you think about, again, my intro slide about like connecting the dots, like your website is very similar, right?

Like we’re trying to connect the dots between all the content on your site, and so it’s not just about who you are as an organization or the products and services you sell, but like the people within your organization, the amazing content, the case studies, all of these tell a story and answer your consumer or sort of your prospect’s questions. And so we wanna make sure that you’re telling that whole story. And so historically people thought about schema markup just for getting rich results. Again star’s price in the search. It’s much more about that because if it’s about understanding, we need to explain all the pieces so that we can answer those, especially long tail, very specific questions that are coming out of AI.

The second piece is around depth, and I’ll show you an example here. But depth is around, again, like we used to do the minimum, right? So on a website, I would see if you’re using maybe a YOAST plugin, it would say, oh, I it’s a blog, and the blog has this title and it has this author, and it was published on this date.

And maybe it’s this is all the blog content, but there’s a lot more context to it. What’s the blog actually about? What is the blog mentioning? And so the schema.org vocabulary was actually created in order to describe all of these things. And I really encourage you if you’re maybe just, maybe you’re a general marketer and you do SEO as part of that job go to the schema.org website, schema.org.

Has been around since 2011, and it basically defines all the things that Google, Microsoft, Yahoo, and Yandex at the time wanted to know about the topics. And it’s a living, breathing project about what they wanna know. And so my whole point is do you think they would’ve been this detailed? They would’ve put in this much work to give you the language, the vocabulary to describe the pages if they just wanted you to describe one or two things on the page. I think the answer is no. And I think as I talk about where we’re going with agents you’re gonna see some examples of like why that’s important. So I wanna give you an example. It’s like nice to talk in concept, but let’s actually apply it to a page.

So I’m gonna actually use the Schema App solutions page. So this is where we describe what we do and we are a software application. So we would first go and say, okay, we’re gonna describe this page and it’s, ask the philosophical question, what is this page about? It is not a webpage, I’m sorry, but web webpage markup’s actually good.

If you wanna get the favicon, like that’s still a good reason to put webpage markup, but like these are bots that are answering advance questions for us. So what is this page about now? If you can’t give an answer on what the page is about. It’s also maybe oh, is the page specific enough, right?

Because it maybe I need to be more specific about what it is. So in this case, like we’re talking about our software application, this is what this page is about. And so we go to schema.org and we say, okay, what are the things then that I can use the properties? So the type software application, what are the properties I’m gonna then use to fully describe what’s on the page?

And it won’t always be everything, but you can see on this page, like I’m gonna describe the creator of our software application is Schema App. It’s got a name it has a description. There’s an example of how it works through our video. We have a review from one of our customers and then we also have an FAQ.

And the FAQ is the the page is the subject of that FAQ. So you can see I’m connecting all the dots. I’m connecting to our homepage, which is actually where our organization is described. I’m connecting to the FAQs, making sure that’s not on an island. Everything is in context. So just again, like I introduced myself I was describing if I just talked about the Kevin Bacon game, you wouldn’t know that you can win it because you know me. Again, we’re pro trying to build context between all the things that have to do with your company so that when people ask questions, they can get the right answer.

Alright, so schema markup drives rich result. So just with the depth and breath, like you still want to try to make sure that you’re getting a rich result. Rich results are found on Google’s website. So there’s a page that describes all of the different rich results.

They are evolving. In fact, there was an announcement this week that like some of the less used, rich results, so dataset was one of them, book was another one. They’re going to retire, but there’s still lots around jobs, events, products, recipes, software applications. So there’s lots of different, there’s about 42 different ones.

We are still seeing these drive higher click-through rates. So I would say these are like low hanging fruit to drive results to your site. Now, they might also inform you to add things to your site. So let’s say you have a product page, but you don’t put a price, but your, team is thinking about maybe putting a starting price or a price range.

Those can actually drive higher click through rate. And I’ve been sitting in quarterly business reviews with our clients and they’re still driving higher click through rate. And I think in a world where clicks are disappearing because of AIO and because of AI chat this is a great sort of easy win within your organization.

Now how do I go beyond rich results though, and stay on this topic of inferencing? And so I wanna talk a little bit about building that content knowledge graph. Going from just like depth and breadth, explaining what things are about to really unlocking these other areas of value. So let’s first start talking about what is a content knowledge graph?

Knowledge graph, you’ll see people talking it about a bit more. It’s in the area of semantic technology and it really is where you’re defining relationships between entities or things using a standard vocabulary. So in this example, we’re using the schema.org vocabulary as that standard vocabulary.

And when you do this, there’s tons of research. Knowledge can be gained from inferencing. And this is, I think sometimes helpful to see it, a visual. If you were to look at this is one of our customer’s websites when you connect all the dots, all of a sudden it’s this network and it looks like our brain, right?

Like when we think of like our network of our brain there’s all these pathways, all these ways that things are connected and it is those relationships that unlock that value. Now it’s not just me talking about knowledge graphs. Gardner also has done research saying that knowledge graphs are on the rise.

As we look at this next hype cycle of artificial intelligence that knowledge graphs are a technology that we wanted. That we wanna look at. In fact, one of my favorite pieces of research from 2024 was a Bullseye, and it was saying like, in the year of 2024 gen AI and knowledge graphs were the key research that you should look at.

So one of the things that we do is like, how do we do this then? We’re talking about schema markup today. We need to connect entities on your site in order to elevate from just doing depth on a page, right? Breadth, getting coverage of the different pages. But now we need to do this connectedness.

This is where we unlock all this value. And so let’s start with an entity. So an entity is a bit of a weird word, right? So think of it as a thing and you’re describing it like with details. So those properties that we just talked about in the schema.org vocabulary whether it be price or color or parent organization is like thinking of it as taking it from like a one dimensional word into kind of a three dimensional word.

And the importance of this is I, sometimes describe my shoe. Now, if I were to tell you like, I have this favorite shoe, and my shoe is amazing I’d be like, okay, Martha, great. You have a shoe that’s very exciting. But if I tell you I have a hot pink, sparkly high heel, that’s my favorite shoe, all of a sudden you have a totally different understanding of what my shoe is.

And so that’s what we’re thinking of doing with entities. Like we’re using schema markup to add the hot pink, to add the sparkly so that there’s a much deeper understanding, a much more contextual understanding. And here the example is if you say the word apple, like how do we also give the understanding that I’m not talking about the fruit apple, but I’m talking about the company Apple.

And so schema markup is really this language to describe and define. Entities or things on your website. And when we think about entity linking, now, we’re not just describing it, but we’re actually connecting. So we’re not just wanting to describe it, but we also want to connect it. And when we do this, again, we’re having this really high quality descriptive knowledge graph.

And then it also provides. Understanding and clarity for AI. And this is I think why Microsoft and Google, when I started off keep saying, do this. And that’s why again, Google kind of adding like AI systems like it, right? It helps them understand this is why, because we’re really being deliberate about those connections.

So there’s two types of entity linking: external entity linking. So this has been something I’ve talked around for a long time. You might hear other sort of thought leaders talk about it. And this is where we’re clarifying something, using external data sources. So whether that’s Google knowledge graph, wiki data, Wikipedia Wiki data, think of that as the non-language dependent version of Wikipedia. Elon Musk just announced Grok PDF, so that might be another kind of way that you can disambiguate or clarify things. And so again, here we’re just trying to say, like on our page we mention knowledge graph. Why? because we’re a knowledge graph company.

And so we’re saying like, what do we mean by knowledge graph, we mean these specific definitions. So you can do this for, again, you don’t, I wouldn’t say you wanna put like a hundred things on your page, like through external definitions, but what are the main things, right? So just like when you’re planning your content what are the main topics or things I wanna talk about?

You wanna make sure you’re being clear in your schema markup and you’re being clear on the relationship. So about is a really strong relationship. Like this page is about this thing. This page is mentioning this, or you might be defining other relationships depending on how it’s defined.

Now, the relationships might also be internal references. So in the SEO world, we know this as interlinking or backlinking like within our websites. What if we can do that with context, right? So we’re still in this case now pointing to that location. So you can see my at ID in my schema markup is saying, Hey, this page, the highlighter is defined over here, but I’m also describing the relationship.

So in this case, I’m talking about how the Schema App solution has a part, an element of it is our Schema App highlighter, and hey, machine, go over to this webpage to learn about that software application with more detail. Go find out about that entity on that other page. And so we are seeing internal and external entity linking, having an impact on results, and I’ll give you an example right there, but this is a how we move from this old world of rich results to this new world of building a data layer or building a content knowledge graph. We have seen in research, so BrightEdge did this great research earlier this year saying that when they did entity-based SEO and they made these connections it was three times more likely to be cited in AIO or AI overviews and AI responses.

And so let’s see, like what that looks like on your website, because I do find it’s great Martha’s talking about like the software application page, but again, now we’re connecting the dots not just on a page, but excuse me, across the whole website. So in this example, you can see I’m describing the page that’s talking about my software application, but I’m also defining a relationship saying, go to my homepage, right?

Which is where my organizations described, right? And if I was Apple I might describe it with that same as that external entity linking there. I’m also saying that, hey, this page over here talks about my service output. So go look at that page in order to learn more about my service output. Or my example work, it’s actually described in this video, but by the way, the video also has schema markup describing the video, right?

That entity. And then I’m also saying. My software application has an offer. Alright, great question. Can external entity relationships be faked? Can someone effectively create fake connections to make famous person doing the virtual name drop? I wouldn’t say like you can, just like authority, I would say, like this is just another way of adding authority.

I think just like keyword stuffing, people can overdo it. But I don’t know if you can like fake relationships. So it’s it should still be like within like your own content as to what you’re describing. I don’t think unless like that content is an authority, I don’t think really like having more links is gonna make it be a different authority.

That would be my answer on that. So I would say just like keyword stuffing. Don’t overdo it. Google has everything, like the schema markup has to show what’s on your website. And so you have to make sure that your content reflects that. So often when we’re talking about this, like someone wants to be more we work with a lot of hospitals, a lot of they’re like, we wanna be an authority on this type of surgery within cardiology.

And we’re like, great. Like, where are your pages talking about that? because we would and and then it’s oh, we don’t have pages talking about that. Great. Like it starts with the content. And actually like. Entity linking. We actually report on what entities show up on what pages to help give data to teams to have those conversations with the content teams.

So if you want to be a topic authority, like your it starts with your content. So I would say that’s how this can inform, like if you want to make a connection to. An internal page that’s talking about maybe a pillar page. We would talk about it in marketing. That talks about a specific topic that’s like the expert page on this.

If you don’t have that’s a great opportunity then to go to the content team or to get that page created. I mentioned this is having an impact. So we are seeing that when we are deliberate about adding entity linking within the schema markup, we are seeing it have an increase in clicks in I’ll say search performance for things related to that entity. And this makes sense, right? If we’re thinking about non-branded traffic and we’re being very clear about how its authorities on certain topics when Google’s trying to figure out like what’s the best page to talk about that topic and it’s tied to an authoritative brand, it would make sense that if they could understand it better.

That it would then drive results. And we are seeing this this is BrightView senior leaving, which is in the senior care area. We’ve also seen it working across healthcare, tech, all different types of content. And so your content knowledge graph, what we just talked about that we’re building with schema markup, right?

Which is this SEO tactic and tool is this data layer. And I wanted to switch gears and talk about like how it’s important to manage it because most people really think about how. I’m gonna do schema mark. I’m gonna turn on the plugin. I’m gonna hope it just works, and then I’m not gonna revisit it.

But if this becomes the data, like what we’re putting into the machines in order to drive visibility, and we know that consumers and searchers are changing their behavior to look more this way. Like we need to be managing it. And some of you’re asking like how do I think of my strategy?

Like what should I optimize and with what schema markup? That’s that top step one. So good news get started there. What is this thing? What should I optimize it on? Then you have to actually create the schema markup. So whether that be with a plugin, an enterprise solution like Schema App, or you’re manually doing it.

If you’re working on a, maybe a more, a smaller site, you have to author it. The, webs, the markup then has to be on the website. So they have to crawl it in order to consume it, and then you can manage it. So that’s like adding things like entities, like adding external entities, internal entities.

And then as your content changes, you’re also gonna need to manage it, right? So again, sche app is a dynamic solution, but. Perhaps you’re working on a plugin, like you need to go check to make sure that it’s like optimizing, like the next things that coming out. Because again, it’s that connected story right across your content that we really wanna make sure is optimized.

And then the last piece is around measurement. And I love measurement because measurement’s changing, right? Historically we would look primarily at rich results and higher click-through rates. And now we look at as we deploy it are we seeing a change for certain query groups or keyword groups?

You can also look at whether it’s you’re getting more traffic referral traffic from different sources. And then my favorite one that I think all businesses should use as the ultimate measure is conversions. So what does a conversion look like to you and how is it impacting your business results?

And this is I think ties to solving one of the biggest problems that a lot of large organizations, again, who we work with is. Is saying they wanna do AI but they don’t have the data to do it. Good news. If you do a really good job of your schema markup, you’re actually building a content knowledge graph or think of it as a marketing data layer in order to prepare that data for AI

So I just talked a little bit about per like data layers and how does that unlock? And I wanna now step back out to the 10,000 foot sort of market view of the world. And so in 2001, Tim Burners Lee, or SLA and James Hendler wrote a pretty famous article called The Semantic Web in Scientific America.

And if you haven’t read it. I really highly recommend going rea. It is lot lengthy, but it talks about exactly what’s happening right now. And they talked about the definition of the semantic web, about machines understanding, right? Like semantic documents, like semantics, all about the context of things and able to then do sophisticated tasks.

And so when we talk about now agents or the agentic web, it’s really. Machines taking now that they have understanding and moving next to execute on different things. And so what’s really interesting is like what they pontificated back in 2001 is actually happening right now. And I think what’s interesting is I’d love to say like the agentic web, like that’s a future thing, Emanuel, let’s talk about that in 2035, but it’s actually happening right now, and we’ll see. Chat GPT and then I’ll talk a little bit about Atlas, that was just announced, like they’re creating partnerships building integrations so that you can literally have a conversation and make purchases and at some point like I look forward to, because there’s things like I’ve run out of my protein powder.

Could you just go to well.ca and order the protein powder? The one I love, like I’ve ordered it a million times, it has all that details outta my email and it just goes and get it done. And, I think what’s interesting is that when, microsoft presented., So Krishna Madhavan, who’s the senior product leader for Bing, when he was talking about how to plan for AI search and agents, he talks about using structure and using structured data in order to help machines understand.

And there’s actually a great article, Emanuel I’ll, dig it up if it, if you haven’t seen it, but they actually produced it on the ad part of Microsoft. Where he actually talks about how to prepare for ai. And he talks actually a lot about what he was talking here, and it’s not just about structured data.

He talks about using headlines in your content making sure you don’t have hidden content on the site. So don’t hide things behind the click. This is all basic SEO. This is like technical SEO 1 0 1. Good news, it’s gonna help with AI and agents, so all the work you’ve been doing.

But it’s a great article in order to kinda give you further evidence. But structured data plays a role here and what’s interesting about Microsoft is that same week, remember I started my presentation talking about how there was a week in May where there was like all this confirmation from Google and Microsoft about using structured data for AI experiences.

That same week, they also made an announcement. About a project called NLWeb. Now, NLWeb is led by RV Guha. RV Guha is the creator of RSS, and he’s also led the collaboration work to build schema.org. The vocabulary that we’re using in order to do structured data and schema org and structured data was always built to be semantic and building these relationships.

If you look at the vocabulary, it’s. Built like that. And what’s really fun is they reached out to us about a month ago and asked if we would actually collaborate with them because NLWeb is actually dependent on structured data. Why? Because it helps them reduce hallucinations. It’s gonna, they’re gonna use it in order to inform vector databases.

Then their vision for NLWeb is to be this agentic endpoint. So I talked about how the agentic web is these bots going and doing tasks for us. And so Microsoft is working on this open project called NLWeb to have a way, a standard way that we can actually put something on our website and make it ready for agents.

And what’s super cool is that they’ve already built integrations with it with Atlas, which is the new, Chrome, or sorry, the Chrome competitor coming out of open ai again, making it easy for us to all play ball, right, in this kind of very fast evolving work. And so the work that like you’re maybe learning today and hopefully the action items you’re taking in order to get your structured data to be more than just explaining a page but bringing context is actually preparing you for agents. And the other play I wanted to just talk about is that there’s also a lot of talk about model context protocol. And I put this in here more about just get yourself educated about it because this is actually like an API for agents. So think of it as a way for people to interact.

At Schema App, again, we have an MCP connector, so people who do schema markup with us in an enterprise can use it to connect to their Copilot, connect to their chat bot. But we should be thinking in this way, like, how are we gonna make sure that our data that we’re producing for search and for AI, building a marketing data layer.

Is really ready to be consumed by whomever is going to be that new consumer, right? Whether that be agents whether that be Chat GPT whether that be Google’s next Google mode. And so that’s really what I prepared to share with you today. We have a ton of content at Schema Apps, so if this is interesting you Emanuel, thanks for sharing that you use those resources.

There’s one here around mastering AI search essential strategies. And insights. But there’s a ton of ton of research. Again, the https://training.schemaapp.com/ is a good place to start if this is new to you. Or if you wanna read how it’s moving the needle like evidence of how it’s driving clicks, conversions.

We also have that all there.

Emanuel: Thank you so much. I’m gonna leave this slide, on the screen. Maybe. Let’s see if I can make it just a little bit smaller. I’m pretty sure I can, but I don’t have one predefined. Okay. No, that didn’t change a lot. Honest. I’m gonna leave it like this so people can can scan and

Before we dive into the question. And by the way, please add them there. I like to give my review and my 2 cents unsolicited, 2 cents about what was discussed. I’m pretty confident now that it was you who give the definition that Schema is data about data. I don’t remember exactly.

Martha: Oh, yeah.

That’s funny. Yeah.

Emanuel: But I’m pretty sure I’ve heard it from one of those webinars that, that is from you. Now, I did do take a lot of notes, which I encourage everyone to take and lots of great information out there. One thing I struggle with is I have a friend, and maybe you know that friend, his name is ChatGPT and he’s good at everything apparently.

And if I ask him to build me some schema, even though I, he doesn’t know what business it is or what the business is all about, he’s gonna go ahead and build it. And the first part of the question is no schema better than the wrong schema?

Martha: Know, like I, I think like you have to make sure your schema describes the content.

That’s been around, I’ve been doing this for a decade. And I would say ChatGPT is you should still be looking at it and is it describing the things that you want? Especially for people on WordPress, like I would say general schema’s better than nothing. If you’re at least describing what it is.

But I think like now more than ever, there’s a need for us to think in more advanced ways. And I Emanuel people don’t do the harder way to do it, the way we do it at Schema App because it’s hard and it’s hard to maintain I think that speaks a little bit to Alex Meg’s question, right? Like it’s, not something that you can just do once and then if your content changes, the website changes. Like you have to maintain it. That’s why my third tip isn’t just just don’t do the work. Like you have to be thinking of this as a process. Just like you check on the health of your website, other, web vitals, like you need to maintain it and it’s something called schema drift. So it’s like where you would actually build the schema, but not necessarily maintain it or have it like reflect what’s on the page. And that’s why having it as it’s oh, I’m gonna go do it in ChatGPT and I’m gonna cut and paste it on the webpage isn’t the best way to build it.

Like it should be dynamic with your content if that content’s changing.

Emanuel: And that kinda like addresses Alex’s question. Yeah. For sure. Is there a better non-linear way to, or is there a better non-linear way to organize?

Martha: Yeah, and I like the, other question, Alex, I go to is will Schema exist without the website?

That’s something we have debate on. And I think the answer is no. Like it’s still based on content, right? Like we’re still trying to answer questions for humans. Like Marie Haynes who presented also at S-E-O-I-R-L is like the human is still at the center of the experience. And so again, like what we’re trying to do is still answer the questions for the humans.

And then those still need containers on where they exist. We just now need to think of like a new consumer. Like it’s not just about the humans. Like we need to think about the machines now more than ever. And before that used to just be search. Now that’s, now that’s also AI I have an old t-shirt, one of our very first t-shirts at Schema App.

It actually even has our old logo on it. And on the back it says, I speak to machines. And I think that one’s probably needs to come back. It has to come back on the way back machine. because it’s, more relevant than ever.

Emanuel: And before moving to the next question, to address your point and what Kevin asked earlier about being able to impersonate, I know that, and you mentioned that some of the large organizations actually don’t do these things like the schema because it’s hard and because they’re not aware, but they dominate kinda like the search. So this is an opportunity for smaller organizations to win on this on the search. Yeah.

Martha: Yeah. Where I’ve seen it really fun. We also work with, I would call challenger brands or like smaller organizations that are trying to stand out against some of the big brands.

And yeah, it’s absolutely a differentiator. And if you can do it well, if you can actually do you know the connectedness that I shared about the external and internal entity linking, we’re seeing it move the needle. Like you saw bright VVC living saw 25% increase in clicks. And this, with this world changing, I would say like rich results, table stakes, get those done at the very least.

And Mek your question around like rich results are when like maybe you don’t have all the content on the page. Use that as a, to inform your content strategy then. We need to get that on the page, because that’s gonna drive a higher click-through rate. Or again, you can use the schema.org vocabulary to inform content.

So content and SEO more than ever need to be working really tightly together because again, these things work in, work together.

Emanuel: And I do feel like cautioning again about asking GPT, sometimes it invents attributes that don’t exist but those, you can kind catch up if you use the schema validator, it’ll tell you there’s an error.

Don’t use that.

Martha: Right. Yeah. And it’s an iterative process, right? So I think especially if you’re learning how to do it it’s a great tool to help teach you how to do that. And again, like if you’re, we started off a long time ago working with smaller businesses where the websites weren’t huge.

Che cheche, bt and you iterating on it, validating, editing it is maybe not a pla bad place to start if you have a relatively static website. It’s when we get into like more complex or larger sites where there’s changes going on to content that you really need to be thinking about, generating J-SON LD off of a JavaScript or working with a technology like ours, or again, a plugin that sort of works with the components of the website. Those are good options that you can have there.

Emanuel: And one last caution because I’ve seen it many times is with the address. So if there’s an attribute that doesn’t exist, you might be able to call it, but if the address is wrong, that can potentially create the issues, address of all the address or phone numbers and one whatnot.

Yeah. So use Chat GPT carefully and I’m just looking to see if I missed anything. So just Ali’s comments. Schema markup is basically defining your content and help search engines understand it.

Martha: Yeah, and I just extend as not just search engines, it’s gonna be chat like AI platforms, chat, GBT, and then agents as well.

Emanuel: I took a lot of notes. I’m gonna come back to this recording and take some more notes. Marta’s site is a great example on how to implement schema, so I recommend everyone to go visit. I believe you’re pretty active on LinkedIn and you shared all the links. I believe we’re also close to the time, so if there aren’t any more questions.

I just wanna take one last moment to thank you again, Marta for

Martha: Oh, you’re very welcome. Thanks for having me

Emanuel: being live now and for putting all the information that you’ve shared throughout the years. It made a difference for many businesses and at the end of the day for everyone, because we’re doing SEO in the SEO space to provide people the best best answer for a service or for a product. So we’re helping Google, in a sense.

Martha: Amazing. We’re happy we wanna teach the world how to do schema markup properly. We’ll keep producing content and share as much as we can to help people get that done.

Emanuel: Thank you so much. Thank you everyone for joining. You’ll find a recording of this session at howaboutsomemarketing.com. That being said, see you next time.

Martha: Thanks so much.