How Fortune 500 Brands Dominate AI Search Results

Hello everyone and welcome to today's webinar on how Fortune 500 brands dominate in AI search. Today I'm joined by Josh Bliskll who is a brilliant mastermind and very well known in AI search circles from profound. Hey Josh, how are you doing? Doing well, thanks for having me. Fantastic. So today we're going to be talking about AI search and we know already that people are definitely searching in AI. Around 800 million people are actually using chatbt on a weekly basis and that number doesn't even account for all of the other LLMs out there. So I'm talking about Gemini, Copilot, Claude, Perplexity, Grock, whichever your favorite one is. What we know for sure is that you can make changes to influence how AI speaks about your brand and you can make those changes quite quickly. So, we had a client who wanted to be top dominating in AI search for best iPad Air cases. Some of our PR strategies got them there almost overnight after we published articles on some choice places. We were also working with a client in the finance sector who they figured out there was a lot of negative sentiment around their pricing. They made some changes, restructured some of that, removed a registration fee, and we saw AI shift a couple of weeks later to be much more positive about how they price about their business in general. The thing is though, you do have to take action. You can't just like have a wish and a prayer and hope that everything's going to change in AI. So CEOs, CMOs, marketing VPs, smart marketers, they're all thinking about how to actually tackle this change in search behavior. And we know that AI search optimization is going to be a big part of 2026 strategies. So just a refresh, I'm Charlie, CEO at Exposure Ninja. We're a search marketing agency specializing in SEO and AI search. And I have the absolute privilege of speaking internationally on AI search, as does Josh, who is a full-time strategist and researcher at Profound, which is an enterprise level AI tracking platform. And Josh has been the brains behind researching many fantastic strategies being used by Fortune 500 brands to defend and grow their visibility and AR answers. I am so excited to finally get to pick your brains, Josh. It's exciting. Yeah, I'm excited to be here because I mean the uh the space is definitely moving quite quickly and it's I mean it's an interesting moment I think for marketers as well because it really is kind of the coalescing of like PR com social search all those things are once again incredibly relevant but all at the same time in the same place with AI search as a channel at least. It's an incredible shift. I think it's the biggest shift we've really seen in online marketing for quite some time. I feel like it exploded onto the scene and now 2025 is the year where marketers everywhere are thinking, okay, what are we doing about this? I guess my first question for you though, Josh, is Fortune 500 brands, do they have an upper hand in dominating AI search results or is it all just luck? Well, I think they do. I mean, I think in inherently Fortune 500 brands have the opportunity to leverage a lot of the training data about them in answer engines, which normally when we're thinking about answer engine optimization, we actually don't worry about training data too much. But the beautiful thing is that you kind of have this basis of information that yes, it is a blessing. It can be a curse at times if you're trying to rebrand or drastically change what it is your company's known for. Uh it is interesting to see like for tons of Fortune 500s though there'll be like articles from 2018 pulling through. But for the most part, you know, if your company is telling the same consistent story, the same way, continuing to release products that are aligned with your products, which most companies are, yeah, it's really interesting because you have this opportunity to lever up and use your content um you know, your your just your total content or your content engineering or your your pillars of content to really hypercharge and start ramping up that flywheel of content specifically for AEO, really refactoring, optimizing, and thinking about, okay, We might have this library. I was looking for the word library of content. What do we do to make it really um really impactful and hyper specific? It's hyper specialized for answer and AI search. So I think it's also a disruptors market too, but I think it's it's interesting for Fortune 500s. I think a lot of people listening are going to find it quite scary when you say things like AI is referencing their articles from 2018, particularly if there's been people in corporate level roles, SLT level roles, marketing VP roles trying to shift some of that positioning and narrative about their businesses or about specific products or services. How much of a danger is that going to be? Is it always the case that AI looks back through like their whole database or can they actually make changes that are going to shift that narrative now? Oh Josh, I've lost your sound, I'm afraid. [Music] How's this? Is this better? Back. Perfect. You're back. You're back. Okay, beautiful. Strange. All right. Maybe the AIs are trying to take me down. Um, the interesting thing, when you're telling all their secrets, Josh, well, here here's the deal. If you're chat GPT, you're looking for high utility content. You're looking for content that directly answers the question which the user has posed. And so, and actually by virtue of being chat GPT, you don't have access to any of Google's data, which means by extension, you don't really know what sites have like higher domain authority, more backlinks, rank for more keywords, get more traffic. You don't have the whole picture. All you really have is a list of web pages and you're going to pick the web page that you think has or contains the best answer to the specific question. A lot of the times what that ends up resulting in is citations and influence from pages that yeah, there might not be the the newest page. Yeah, they might not be well trafficked, but for whatever reason, the on-page content is structured, is detailed, and it offers the answer engines the exact specific answer that they're looking for. uh maybe not the most recently provided answer but it seems to be the most specific out there in the web and so we see that shift towards value pure value and really pure utility really taking hold within AI search and it's a great differentiator between SEO and AI well that leads into my next question because I think a lot of people listening going to be really familiar with SEO many marketers know SEO pretty well particularly our webinar and podcast followers but AI search I think feels like a new game from your research Josh what has been the biggest difference between SEO and AI search that you think marketers need to understand is definitely the collapse of traditional authority signals. So when we're looking at content that ranks like you can go to Semrush, you can look at the top 10 cited pages given an industry like I love the uh fast food is a great example. I love it as an example. The number one cited source for a few months in fast food was like Bob's like it was like John's Burger Blog. It was like some random guy who wrote an incredibly detailed blog of all the different fast food burgers that he'd had. And so the answer, and by the way, not well, not well ranked, not well cited on Google, but the answer engines specifically, ChatGpt, latched on and was like, "This content is absolutely incredible. I'm going to reference this as the starting point for learning about the nutrition facts and the calories and the weight and the contents and is it real beef? Is it flame grilled patties?" um for hamburgers or for hot dogs or for French fries um in the fast food industry when our customer was a pretty major fast food company looking at this being like we would never have known that this was dictating so much of the narrative around these products in chat GBT. So like I think there is this flight towards quality and flight towards contextualization where the the new mandate for every brand in the world is to provide answer engines with context on what their products are where they situate in the market and really the the facts and figures behind what they what they genuinely are as an up-to-date direct way in a direct format as possible. That's where we have structured data starting to take hold. That's where we have HTML tables being a really important messaging signal. That's where we have dates in the title tag or even in the first few hundred characters of the content really being important. And you're going to say like, yeah, are those fundamentally similar tactics to what we do in SEO? Absolutely. Those are similar things. The way I contextualize that though is that, you know, at the core, we are we are using the same muscle groups. We're still playing with the same units, right? URL slugs are still important. Title tags are still important. metad descriptions. They're actually less important in SEO, more important in AEO, but they're still units that we use, but we're doing different exercises with those same muscle groups. Now, we're using those muscle groups in a different way to optimize for a different outcome. Uh, and there are I mean, there are some really cool plays you can make with an optimization. We can get into Yeah, we can get super granular in time. Exciting. And I agree. I think anyone who works in SEO, content marketing, they've actually got the skill set already to be working in AI search. Some of it is just the confidence, the experimentation, the understanding the nuances that are different between just doing Google or Bing optimization and now actually thinking about CHBT and all these other LLMs. Let's get into a bit more. Josh, last year you did a fantastic talk at Brighton SEO and that was your findings from analyzing I think it was 10 million AI search results. It may have been even more than that. It's actually 40. Yeah. Oh, 40 million. Okay. Quadruple it. Can you give us can you give us a top level view of what your findings actually were from all of that analysis? I think biggest find a few of the biggest findings. One there is this technology that lives inside of chatgbt called webgt that is actually responsible to be the decision maker for chatbt search. It's not just like chatgbt vibing out deciding what it's going to site. It's actually so when you pass your query into chat GPT imagine like a 10 paragraph query. ChatGpt reads in that 10 paragraph query and is like, "All right, let me restructure that 10 paragraph query into like five to 10 keywords and pass that over to my search API, which is the actual thing that searches and scans the SERs." And then when we grab those search results, I'm going to pass those search results to web GPT, which its only job, its literal only job is to decide out of these pieces of content, given this particular prompt, what article or what source should I actually be citing? So, it's been an really really incredible opportunity to start thinking about that technology because there are things that it implicitly does. There are ways that it implicitly chooses what content to site, what units it looks at more closely than others. Uh, another I'm going to go with like three main insights. The second deal is that listical and comparative content was the most cited type of content period in the world within AI search. So, that was 32.9% of all content was listical and comparative content. blog content was sitting in second place at 9%. So the bend towards comparative listical content within answer engines was absolutely incredible then I know this will be maybe a little bit of an inside scoop for the audience here that percentage for listical comparative content I'm about to go to Brighton next week is down to 25% but it is still the largest share of content blog content went up to 12%. So we're starting to see the field narrow a little bit there but it is still pretty disperate. I think the other thing that we found uh that was really interesting was Reddit was starting to become relatively important in AI search. So Reddit was really starting to make its first kind of ascent steps in being quite relevant as a channel play within the area of AI search which really brings the social teams you know aboard and we have to start thinking about what we're going to do with Reddit and how Reddit plays into those strategies. Yeah, I've seen so many studies where Reddit is absolutely huge in terms of the number of citations that it actually makes about businesses. And I think a lot of the time businesses don't even realize because in many cases they haven't actually been able to try and control conversations in Reddit. It's not been part of the social strategy. They've been focusing on Instagram, Tik Tok, wherever else. And then suddenly there's all these like messages, forums about their products or their services. And then now they're seeing that come up in AI search more. I think there is almost a bit of a a panic mode about bringing that into the strategy. I bet Reddit are very pleased with themselves at the moment. They Yeah, I think it's an exciting moment. Um, also for anyone who is listening thinking, "Oh, Josh is going to be at Brighton SEO." Josh is going to be at Brighton SEO in San Diego on state side. I think you're not UK side this October, unfortunately. Not this fall. Not this October, but then back in April. Yeah. Okay. So, if you're state side, you can see Josh in San Diego. If you're UK side, this side of the pond, then you can meet him in April next year at Brighton SEO. Highly recommend checking out his talks because they are really good. Also, if you have questions, I can see that some people are are starting to have questions. Good. That's great. We're going to make sure there's time for Q&A at the end. So, feel free to drop all of your questions into the chat whenever they come into your mind. Uh, one of our colleagues is in the background. He's going to sort through and raise the best questions for us to answer before we finish. So Josh, Fortune 500 brands then we know that they're starting to take action. We know that Profound are working with some already who are deciding, okay, what do I need to track? How am I managing my AI search? How am I managing all of this traffic that's coming through from it? What are these brands doing about AI search? I'm talking about brands like Bosch, Expedia, Chanel. Yeah. The cool thing about what we're doing for Fortune 500s and what we're really showing Fortune 500s is first of all where and how high their their on page and owned content can actually fly. So most people assume that you know when you have a listing page or like even e-commerce let's talk about like let's talk about e-commerce pages for a second. If you have a product listing page most people assume that's going to be pretty nent like it's going to serve one purpose. It's very utilitarian and you know maybe you expect it to rank for like navigational or transactional type queries but we know within answer engine search and AI search that people are asking informational questions and we need to contextualize our products early and frequently and in a quite detailed way to try to get those products in the consideration set at the onset. So, one of the things that I think is most interesting is really starting to add rich detail and machine readable context to listing pages or product pages. That's been massive on the e-commerce side. If we think about brands that publish a lot of content, it really comes down to thinking about our partnerships and our like even there's a few things. One, does ourformational content work in a way that's structured for answer engine pickup? So if we're going to describe an industry, can we describe that industry in a way or a space or an industry, a topic in a way that offers many different types of answers, many different FAQs? Obviously FAQ schema is a normal natural thing that we're going to think about when we're releasing blog content or releasing kind of FAQ schema is kind of taken the world by storm in the last like 5 10 years. But you know, when we're thinking about the kind of content we're creating, can we go further than just a few random FAQs? Can we kind of hyper contextualize those to be semantic prompts or using rich prompt examples in our content in our H2s asking full questions. Um I think a lot of it is really bringing out like semantic search and semantic HTML as well. So thinking about our on-site technical implementation things like JavaScript JavaScript is not viewable by answer engines. So, it's absolutely insane when you go to a company and they've they're like, "We're pouring so many resources into a blog or onto this on page experience and you block JavaScript, you reload the page and the whole thing is blank." And that's the experience the answer will see right now. So, you got the smartest people in the world. They've made these incredible AI systems and the scraping stops at JavaScript. It's really only looking at the on-page HTML. Um, so site structure is just as important as ever. I think really thinking about the structure of your content though, I think that's quite underrated. So really thinking about the semantic chunking behind things, really thinking about the way that you're going to be building out that point, saying, "Okay, what's the average rent in San Francisco?" The average rent in San Francisco is $2,700 for a studio, $3,400 for a one-bedroom, and $3,900 for a two-bedroom. And that's increased 7% year-on-year. That's infinitely better than, you know, in many, you know, San Francisco has many neighborhoods. There are so many great places to rent from and the different rental prices across San Francisco, like the amount of intros I've seen that start with like in a world of or across the world or there are many things. It's like we really are thinking about shifting from telling narrative stories to trying to convey key pieces of information to these AI systems which are crawling. It's kind of like just not being around the bush in what you're saying. Going back to your point about e-commerce, I saw a fantastic analysis today about how different personas, different customer personas buy different headphone types. And so AI search when it's actually become personalized when you know you've made however many searches in chat GBT and it's got your history. It knows if you're like a gamer, if you're super into your audio, if you're actually just a commuter and you need to bosch on a pair of headphones on the train, and then it gives you different recommendations based on what kind of person you are because it knows that certain kinds of people like prefer using Sony, certain kinds of people prefer using both. And I think what you were saying there about e-commerce pages and actually it's not just a like, okay, here's the buy now button, like come along and this is our newest headphone set. There's so much more context that can actually be given around why those headphones are great, what features they have, the benefits they have, the kinds of things that they know their audience and customer personas actually care about, which is it feels so much more than just uh SEO because now we're having to think about brand. We're actually having to think way further about our content and what we write on those pages. I imagine some people are thinking like, right, I'm just going to write everything on that page. I'm just gonna write so much information on that page. What would you say if that was the thought that just popped into someone's head, Josh? The there's there's two modes right now. One, I think I think both of which are not ideal, which is crazy, right? So, how how can there be two things that are both not ideal? I will I'll kind of explain them like this. Right now, in the short term, the answer engines are they're they're kind of forcing brands to adopt a strategy of kind of having one URL or one page per distinct topic in a very granular sense. Like yeah, obviously one page per topic makes sense when you say that verbally, but what I mean is like if we're asking what's the best CRM for small businesses, that could ostensibly be its own cited page different from like best CRM overall versus best CRM for healthcare versus best CRM for fintech. The crazy thing about that is obviously if you're an SEO, you know crawl budgets exist. You know site structure is important. Like those things work to a point, but like taken to their extreme, there's no world where you can have, I don't know, infinity many pages that each contextualize a different answer, right? Like even though the reason that that actually happens is because a chatbt or a web GPT will look at the URLs and they want to find the URL that's the most hyperontextualized to the specific ask or the specific question at hand. So they want to find the URLs, but finding those URLs is actually incredibly tricky. It's incredibly difficult if those URLs are just like best CRM um you know uh josh.comcrmest. You're not going to win the WebGPT game. The person who's going to win the webg game is going to be josh.combestcrm-4-healthcare-2025. Even including 2025, it's dumb for SEO, but it actually works really well for answer engine search because if you're redirect, you don't want to do it. But in but like staking your flag that directly and matching the the query fan out that way because the query fan out imbuss queries of 2025 inherently biasing the model towards pages of 2025 in the title 2025 in the first few hundred characters 2025 in the URL. Yes. Um it's like it's an immediate thing you can do but I wouldn't recommend it in the long term. I think answer engines are going to get way better at scanning big pages, pillar pages, like end all beall type pages that really offer megat tons of context inside the package of a I guess how do you say just like an overall wrap-up of of a huge topic. I think it'll be more like reading a book than reading like individual paragraphs if there like what what's like the level of a page. Um, I think in the very long term, maybe there's just going to be a standard where by which we can just give answer just plain textual versions of websites just like like an LLMs.ext, but LLMs.ext never really panned out. It was never really as like sexy or exciting as people assumed it was going to be, which like totally fair, but I think eventually there's going to be a world where the answer engines are going to need some sort of base level access to sites to find just basic information. Do you think as well it's going to look like a different experience for human users who come to websites versus AI? Because some people there like you're saying, "Okay, I think pillar pages are going to become big." Some people are thinking like, "Oh, barely would read a 600word blog post these days, let alone anything that's longer than that." Is is it going to be that AI is just summarizing all of that and people rely on AI to do that for them? I think AI is going to be summarizing a lot of the top ofunnel content that exists in the world. But I think one of the cool things AI is also really interesting on the human side of websites because I think websites for a long time have been forced to be like websites look very sy right now. I feel like we're in an age of SEO where like there is a known playbook. Things are very distinct and everyone knows kind of what they need to do to make their website work. I think AI presents an interesting opportunity to make websites kind of weird again to make websites really the front end of the website should be like a functional like a functional human experience more artistic than you would assume like I was working with an interior design company and they had like a chair that was shaped like cheese and their like URL was the blog URL was like you know architecture company.comblog cheese and they're like why is this not getting picked up for like chair designs and I'm like yeah it's sending all the wrong signals it's a really cool web page too, but it's not going to work. Um, I think there is a world where where the raw information should be easily accessible to the answer. It's like a two-lay two-level highway. Raw information, boring, business in the front, but then on the top party in the back 100%. Let's have websites that are artistic, that are weirdly interactive, that are not necessarily easily parsible by machines because we have this whole machine layer existing at the bottom like almost like um just like a word document basically of context. Maybe maybe it's slightly more interactive. Maybe you can parse through it. But I I believe that very strongly that like we do have this opportunity. Now whether we choose to take it, I don't know if we choose to take it, but I think it it could be a very welcome change, at least in the web dev and web optimization world. I'm sure web devs would love to be doing something a bit different from the kinds of uh site structures that they do at the moment. The thi the thing that I think will be big here is there will be certain kinds of businesses who want to have fun with their websites in that way and there will be kinds of businesses who just don't where that just isn't the vibe that they're going for as well. So, it's interesting that you say you think there could almost be like a two-part of a website. So, that that would allow that like portfolio fun type of front end for the kinds of websites. Actually, a lot of things in the marketing sectors like social media agencies, social media companies, architects, anything visual, I feel like that would appeal quite a lot. And then there's like the nitty-gritty stuff where I think like people are thinking that no, I'm just it's a SAS product and like it's just going to be what it is. we're just going to try and get people to buy it. Um, thinking about that kind of thing though, Chanel is a really interesting Fortune 500 brand in that they dominate quite heavily in pretty much anything to do with like latest fashion trends, third party coverage everywhere. They do so much PR and they also have owned editorial. They have like a whole heritage of storytelling that they can do as well. If you were a brand that doesn't have that, what would you do? I think it's going to be about telling that story as many times in as many places as you can. Like your brand story and your product story and your reason for existing is the I mean that's the raw material. That's like that's the ephemerra of what makes your brand yours and whether or not it exists on the web is going to be the critical factor for AEO. Like the I think it literally like you could graph like the number of kilobytes of information that exist about a brand online probably at some level correlates to the AEO performance of that brand in a raw sense. I would bet you there's like some basic correlation there. There's ways to obviously that's like the most like brute forced way to do it. But I think that's the practice of answer optimization is to make the most out of the fewest kilobytes, you know, to like literally like hit the dart on the bullseye every single time. But at the most basic level, you know, for an answer engine to understand what your brand is across different channels, across affiliate, across earned, across your own, across your socials, um, and just to be part of so many stories, that's the ideal world. And so I think the beginning step tactically speaking is to start by blogging by start by starting starting by creating content blogging working with affiliates working with owned um if you know what's being cited in answer engine search about your category or your industry and you can do that with a Google doc like you can use profound profound does it scalably at thousands of queries every day millions of queries across the whole product but you can take a Google doc you can write a few prompts down you can run them every single say, take an index of all right, what sources are actually getting cited? Do a tally. All right, these pages, these sources, these are going to be important to me and my brand's future and the way my brand is contextualized. Getting on those sources or making a source that's like those sources is the future of how you get discovered. And it's again because authority doesn't matter so much because backlinks because traffic isn't really a factor for these answer engines, the gate is open to people who can recognize that that data discrepancy is right in front of them. I also think this is a pretty exciting time in terms of what we know as like SEO link building, which is what I would think of really as more of an early 2000s way of building links. Like you went, you were really focused on getting a hyperlink. You tried to point it to a specific page on your site. There's been a bit of an evolution and we have digital PR, which is not just that sort of guest posting and link building type, but more of an editorial type. You're going after publications where you actually want your business to be spoken about, where you think relevant people will be reading. If you've been doing that already, you're probably already going to start to see yourself doing well. This is almost the next evolution of that for AI search where actually now we're talking about citation building and not in the same way you don't have to go and get like the direct hyperlink anymore, right? Because you've got so much more context. AI can read so much more context about what's going on in the brand. But I think it's a great shift in the sense that PR has been quite scattergun like let's just like go after all of the publications that we can go after that are sort of related to our niche. Citation building. When you've got a tool like Profound or like you say, if you're just doing it on a very small scale, making some searches, seeing what's get cited, tracking it in a spreadsheet, the manual way, it actually gives you a strategy behind what you're doing on your PR function, which I think is huge because then you don't have to have bazillions in budget for doing PR. You can actually be very focused about the kinds of places you want to be published in. And Josh, how influential do you think third party publications are? This kind of PR or citation building we're talking about. Yeah. Compared to I was gonna say on-site bigger than onite. It's different. It's different. All right. On-site is always your fortress, right? So on on-site is the place that you control. The thing about PR is that the sources that are being cited by the answer engines are changing constantly. anything earned, anything outside of your site. What we're seeing right now for like Reddits, for even I'm going to mention Wikipedia is earned. That's insane, but it is true. Um, rear Reddits or Wikipedias, you can call them UGC, you can call them earned. Um, we're going to I mean, we are continuing to, but we're going to continue to see tons of volatility in those sources as we figure out where their role is and how they play a role in answer search. like Reddit's up 80s something percent in the last month. Wikipedia is up 67% in terms of citations in the last month. Um the the important part of all of this is that they're like different media orgs and different media plays, they're going to be they're each going to be their own fair weather golfer for their period of time. Like there's going to be months where certain, you know, publications are doing great and then another publication will come out with a higher utility piece of content or a better listical or a better comparison or a better write up or a better review and you're going to want to move or shift in time. You don't have to do it on a monthto month. It's not like, you know, binary one zero, but that race is not really won yet. there isn't really a place to stake your flag if you're if you don't have the agility and you want to just kind of one and done kind of set it up in the immediate environment right now. You can't do that with the owned environment. You can start building up a fortress. You can start laying bricks. You can start putting the mortar down, building a wall around yourself and starting to think, okay, how do we contextualize our brand? Because at least our brand is known for doing XYZ product or service and when the answer engine looks for those things, it will hit our brand's website. So, I think that's kind of how I'm thinking about it. There's a little bit of like wasteland versus what's controlled. Um, but 80% of the world is earned. Still earned makes up a lion share of everything that's going on, right? So, it is like the value trade-off of like prevalence versus control. You know, you have something that's 20% of everything, super controllable versus something that's 80% of everything. But kind of wild environment, totally crazy ecosystem. And how much volatility like have you seen with the kinds of brands and publishers that get cited in AI? Like do you see that change often? Because we've like you've spoken about Wikipedia, Reddit, they're pretty staple, but is there movement with the other stuff? On a page by page level, it's right around 50% turnover a month in terms of the citations. That's huge. So 50% and just raw terms um of citations will move to another page the next month. And so like in reality what that actually means is that like 50% of someone else's citations will go to 50% of US News's citations will go to Time, but then 50% of Times will go to Forbes and 50% of Forbes will go to Business Insider. So everyone's kind of like we're all playing like musical chairs with the citations. Like everyone's kind of shifting around all the time, but that volatility is shifting. Those citations are absolutely shifting. So the sounds are really uneven. Um, yeah, it's it's definitely not a fixed environment, especially with the recency bias that the answer engines have. So, if we're thinking like what what's one of the major causes of that, it's definitely recency coming to play. Uh, it's just got it really like it's really underrated if you wanted like I don't really have any better word to describe the factor of recency and answer engine search just being massively underrated uh as a cause for seeing shifts. And so I guess what advice would you have for businesses who are thinking like oh my goodness how do I start actually breaking into these conversations like I they're thinking I need to get on this genuine agility is your best mode like agility is the best mode like accepting the the speed of the space is actually something that will set you apart from your competitors like saying okay this is crazy this is nuts we're going to make a few educated bets in this space and we're going to ride it out and we're going going to see a lot of volatility and it's going to be okay. That's like a top 1% mindset to have in answer engine optimization. The second thing is executing quickly like saying, "Okay, we have a huge process for content. We have a huge content backlog. Let's sneak in a blog post. Let's do one experiment. Let's do two experiments. Let's launch a Wikipedia page. Like let's let's push something through to move the needle a little bit." The number one factor whether or not someone's going to be successful and profound, I'll tell you right now, is that first experiment. If you can do your first experiment, which my job is to enable people to do those first experiments. So I would say like 90% of our customers are successful in doing those experiments and getting some really cool results. I think the other 10% really it's about process and getting over process and jumping over that hedge and getting around the content team or getting around the PR team. Um we can get some of those experiments through. You're going to see a tangible gain in your visibility. You're going to want to do another experiment. You're going to want to test AB across two different things, two different blog properties. But I also think if you're thinking about moes and you're thinking about what really sets you apart, it's also kind of dispelling with the necessity of certainty. Like it's really annoying to say. I mean, I was an SEO early on. And so I know sitting in this chair, I would have hated what I'm about to say, which is that like being certain about anything right now is something is a luxury we really can't afford. It's something we really can't do in AI search. And so the kind of old way of running a test in SEO, rerunning it, being really sure before we want to jump in on a certain keyword strategy or before we want to jump in on creating a pillar page or something kind of has to take a backseat to the momentum of the space and the importance of exerting some sort of pressure in this environment or this ecosystem. We just have to exert our brand power. We have to literally move our massive limbs, steer the Titanic. That is the that is the most important part of it. Josh, can you explain what a mode is for anyone who's not familiar with the concept as well? Yeah. I mean, a a mode is going to be your competitive advantage, at least in this space. It's the thing that you're going to be doing that the competitors won't be doing. Okay. So, when I say agility is your moat, it's it's agility is what'll set you apart. Okay. I guess what a lot of people are thinking now is, wow, I've got to get this past my board. I've got to get this idea past my senior leadership. I've got to get it past CEO, whichever senior stakeholders. And some of that pitch is going to be really difficult to make because they're saying, "I don't know for definite that this is going to work. We need to do some experiments." What would you advise them to say to actually get that buy in to start working on AI search? Because we know how important it is. Like we as marketers know it, but actually communicating that to other levels in the business can be really difficult. I mean, we we know behaviorally that um we have a bunch of different cool stats. We're going to release a survey about this soon, which is just like a we surveyed like 2,700 American shoppers on how they use AI in shopping. I can I can give you a few examples of things that like let's I'll start with the most boring granular tactical thing which might actually be cool for the listeners like the like it depends on the audience, but one of the things that like if if you if you're facing someone who's like we don't want to test things on the site, we're worried about the site. Well, we can we can no index or we can block certain search engine crawlers from seeing certain pages and we can ideally have answer engine crawlers only access the page. We can get pages indexed only by answer engines. Now that depends because uh chatbt now the rumors that chatbt is using SER API which is really interesting but maybe that means that you do have to have it indexed on um traditional search engines. But if you're looking at perplexity or Gemini or a Microsoft copilot there are things we can do right. So for thinking about alternative strategies just to show some citations in those spaces. Hey, we de-indexed it on Google but we made sure it was indexed on Bing and we got tons of citations from Microsoft Copilot. Maybe that builds in some momentum to say well look at this like we're influencing one answer engine. Let's start extending that strategy. If you're dealing with a situation where there's just not buyin or belief that answer engine optimization or AI search is happening. Um, we know that 80% of people who bought something basically using an answer engine at some point attributed more than half of their decision-m to AI search. So when people are buying things, you know, looking at things using AI to do it, they're doing it in a massive way. We also know that of shoppers who are using AI search in general, just like people who use chatbt and have looked up products about it, they basically self-reported that 63% of their decision-m was actually made by this these AI search models. We have other data and stats. Forester released a great study, 90% of B2B buyers have consulted AI search at some point during the buying journey. Um, six and 10 consumers are using AI search. Um I mean we know that entire user journeys are happening in these platforms and the really broken part is that the attribution loop is just not there to show it. But it's like this is really like Tik Tok in 2018 or social in the early 2010s. It's a it's a big deal. I have to say I'm impressed with your ability to remember these figures and reel them off like that. It's this is my life. This is what I do. I also think though like it's true how many of us have actually spent time on Chachib or an LLM when making a purchasing decision like if we actually analyzed or thought about our own purchasing decisions we probably rely on LLMs a little more than we would like to admit compared to if you think about how you were buying products three years ago that in itself if you if you if you think in those terms and then you think about those terms in relation to your business it's quite easy to start making the connection of like oh wow like these habits really are shifting for us. Well, here's the user modality, right? You find the thing you want to buy in chatbt. You copy the name of that thing. You open a new tab. You paste it in Google. You go and buy the thing. And so, everyone is attributing that as um as just direct traffic. Oh, I guess direct traffic is all right. Cool. We're getting a lot of direct traffic. We're getting a lot of search engine traffic. Um, one thing that we saw that was really cool, this is for I mean am I going to tell this this one to like the CMO of like Nvidia, like the world, like am I going to go to like the world's biggest companies and say this example? Probably not. But this one's been shocking for us in our how did you hear about us form. This is so benign. It sounds so obvious. We added chat GPT and we've recommended some customers do that as well. We've been so surprised. it has been upwards of 10 15% of leads coming in are saying chat GPT when our sales team hops on we're like all right what prompts and the interesting thing about that is that when the user when we're having that conversation at the user when the customer or the prospect or the I guess the the potentially profound user is like yeah I found you through chat GPT they'll pull up the actual conversation that they had where we showed up and we get to learn like okay this is how someone found it it's a cool thing to do it's a cool thing for the sales team to So it's been massively helpful for us. We we actually just love it. It helps us inform what we're going to track and profound as well. We have also added this to our form like the how did you find us form when people contact us and are also seeing trackbt coming through. So I think it's pretty common. If you haven't added that and you are a service based business, definitely add it to the form because you'd be surprised how many people actually come that way that isn't necessarily in your referral traffic in your analytics reporting that you're using. Um Josh, you also mentioned that you support businesses to do like a first experiment if they're sort of dipping their toe into the waters of AI search. They're trying to figure out how to use profound, how to get tracking, what they even need to be tracking. Do you have some examples you can share of the kinds of experiments that you've done? Yeah, absolutely. Um let me think. There's some really cool ones like not gonna name them. Yeah, that's what I'm I'm debating in my head like naming. Um, so I'll give you an example. We were working with a company that does professional and corporate training and they do it with AI agents and they had roughly they had 3% visibility. I wish I could show the graph. It would be really cool. I could screen share into their product, but you'll have to just picture a graph at 3% out of a 100. That's their AI visibility for that specific topic, right? Yeah. Very small, lots of room to grow. Um they didn't have they really did not do any blogging really no blogging. It's a new company. Um they really focused on like product pages, services pages, about us, contact, careers. Like that was the extent of the site with a really flashy like video as the header. Awesome, cool, respect it. But we need to have context. We need to have data on the platform. So we run a few prompts. We see that they're at that 3%. And the recommendation there is just because all their competitors have large libraries of content really contextualizing their product. What's the best corporate training that uses AI agents? Best AI agents for corporate training. All the major competitors in that space are answering those questions either through dedicated pages or FAQ schema or FAQs just like all littered throughout the blog content whatever they're doing. So we recommended and built a single piece of content just one. We like to keep it simple one piece of content. best corporate training 2025. He released it on August 19th and their visibility went from 3% to 39% within a week. Wow. Putting them in the top four within that industry. They were in like position 17 at that point earlier on. And then with the citation share, which is how often they're influencing it being used as answers, they were at.3% citation share. So it was really only hitting their homepage, we brought them up to 11% citation share. Um that was also just completely astounding as well. So we saw more than I don't know that's like a 30x on the citation share front. It was really really compelling and it was exciting to be able to share that success with them. It was literally again one blog post one like literally one URL changed. That's absolutely incredible. What an incredible result. But also you said that's not uncommon by the way. That's that's like the recent one. It's one blog post but also a lot of thinking goes into what are we actually going to target here? What are we going to write? Oh my god. Yeah. Yeah, I mean that's that's where the work comes. That's the manifestation of probably like 20 hours of work for that one blog post really. Like it is like playing darts. Like people are like well AEO and SEO the tactics are so similar. My my I always I like to clap back with this which is that like the tactics at the base level are so similar. Sure, but it is not answer like it's not search engine good enough. It's not answer engine good enough. We are trying to optimize these systems. We are trying to juice every single last percent out of these systems. So yes, like having comparison content is good, but what's the optimal? What is the literal like peak of what you can do within AI search? That's what we're doing here. The funny thing about this is when AI really exploded onto the scene, one of the earliest use cases probably in businesses was, okay, how do I speed up blog writing now that I've got AI tools? And the reality is like, yeah, okay, you can speed up blog writing using AI tools if you're not worried about what you're actually writing. The reality is blog writing should be actually taking so much longer now because you're really trying to be competitive and thinking about what you're writing. You've got a structure and a strategy going on in the background that you did first before you even decided like, I'm going to push out a piece of content on this because there's no point putting it out otherwise. then it's just going to go into the ether and be completely unaffective. And I think sometimes businesses get really caught on like, well, I should be able to pump out like 100 pieces a month. But if they're all useless, you may as well have done one that actually had an impact very similar to what you've just described here, Josh, and spent that time on the one. The way to think about content creation right now is like when you're doing it with AI at least my mindset is that basically it's not basically the best content in AI. Everyone's always like can we use AI? Is AI going to ruin our our content? Like should we use AI to write AEO optimized content? My recommendation is that your AI written content should really be it should be like stringing beads onto a necklace. Like you have all of these beads you've actually collected. Like the beads are the value. You found data that no one else has. You've collected that data. You've put that data together. Like imagine a listical piece of content to do a really good listical comparison content of the best CRM. You should really do some research into all of Salesforce's offerings, all of HubSpot's offerings, all of Zoho's offerings, how they stack up, what are the different pricing tiers, how many seats, what does each tier come with, what kind of support does it have? combining all of those things, looking at them in an opinionated way, thinking about like taking all the like I'm literally going to keep using the example of beads, taking all these little data beads, all these data points, giving them to an AI and say, "Okay, I did the hard work of putting all these beads in one place." Your literal only job is to just thread the string between and add the ams, the ors, the its, the job. Fill the narrative out. But the piece of content and the value and the structure should have been there the whole time. The real problem we run into is when people want to have the co-pilot or the flight attendant fly the plane. That's not good. We have to be the pilots. That's everyone's I mean like we say that all the time in AI. Everyone kind of knows that. But really take that to heart. Like AI is fine as long as you're flying the plane. Yeah. Absolutely. And your point there about proprietary data as well or just having something interesting that you're writing about something that makes your piece different is incredibly important. Actually, I love the profound blog and many of those posts actually written by you personally, Josh, because that is exactly what you're doing. Like you're doing proprietary research. You have interesting data. You're finding a good hook that people would actually want to read about that. but it is also adding something to the conversation about AI search, about whatever it is you've been specifically looking at. Um, let's get into tracking. Let's talk about what people can actually track in AI search. Tell me about the top things you can track in profound and why people should be looking at them. The the coolest thing you can track, the coolest the best, the most utilitarian, maybe they're all the same, maybe they're all different. Um, one of the best things to track is referral traffic. Even if referral traffic is being crushed and destroyed and clicks are, it's like, you know, the philosophical question is like, we're in a zero-click future, but we're still all tracking referral traffic. Are we setting ourselves up for disappointment? I don't know. I don't know. I'm not going to get in that war. That's a whole war in the world right now. Everyone's like, how do we track? Yeah, it's a separate that's a separate podcast. I could probably write a book on that. Um, referral traffic is still good, though. Like, let's let's use it as a barometer. And really what that barometer tells us is how often we are showing up even as that little chicklet within AI search. The literal number of inline links our website is getting in these search experiences. And and forget how many people click them. Legitimately forget the number of actual clicks. just use it to understand kind of the magnitude because let's if we assume that 1% of people actually click the thing but you're influencing or getting inline links in 99% of other situations you're doing incredibly well depending on what that referral traffic number is like even if I had if I built a personal website and I saw one referral traffic from a search within the first like week or month I'd be like holy hell that is awesome that means I'm showing up in a bunch of different prompts and a bunch of different questions around the world and now people are clicking through that's cool. So like drastically drastically expand what a single click from AI search means first of all but still track the referrals I think. Secondly it depends on what you're tracking like branded questions for branded questions like um I don't know uh the Google pixel like Google pixel how many megapixels is the Google pixel camera if you're a fortune 500 and you just want to know is Google pixels like help center being authoritative in that question citation share is the best thing to track then like maybe I had this great example. I love this example. Converse. Let's just say um someone asks how to tie your shoes. Converse can work 24/7 to create listical content and blog content. There will be no world where the answer engine says the first step of tying your shoes is to buy a pair of Converse Allstar sneakers. It's never going to happen. But if Converse cares about that query so much in AI search that they're like, you know what, forget the brand visibility. I want to show up as the inline link. I want to show up as the chicklet. We believe owning the little citation button on that query is so valuable that we're willing to influence that query without any expectation of brand visibility. Then you want to track your citation share. For brand visibility, that's where you're going brand agnostic. What's the best this? Should I buy what like where do I get a great sweater for Christmas? Where do I get an ugly sweater? Where do I do this? Where do I do that? How do I do this? How do I do that? Um that's where you want your brand to start showing up and getting contextualized. And that's where visibility becomes your key KPI. Um, the really annoying thing with AI search is that like it's an influence channel. Like I I I I've been debating for so long whether or not I should live for a week or a month and do like a video of just like I'm going to live and let AI make all of my shopping decisions to illustrate just how interesting. Yeah. Right. Wouldn't that be crazy? Just to illustrate just how interesting that is. Like I think I do it already. Like my wife is already kind of laughing at me of like she's like, "We need to go buy X." And I'm like, "All right, let me go look for it." And she's like, "JABT, right?" I'm like, "Yes, it is." I just love using Chacht to find things. I think the the trust factor for consumers, and you can ask a lot of consumers about this, there is kind of this like Reddit effect of like it feels as trusted as Reddit, but as personalized as, you know, a unique kind of search result for you. And so, it's combining those things. I think ads will be very interesting because I think they're going to erode that trust factor. So, it it depends where and how they want to do ads. But yeah, the question was on attribution and tracking and we got all the way to talking about uh me wanting to buy things and living like a AI hermit. But referral traffic, citations, visibility, those are the best KPIs right now. Perfect. I actually think a LinkedIn video series of you living in that world where you just let AI make your decisions is is what I'm here for. Right, I'm going to get us to Q&A because we're coming close to the end. We've got like way too many questions. So, I think we're just going to pick out some of the top ones and have a look at those. So, first one, do you know a rough percentage of tracking coming from AI versus traditional search to brand websites? Any kind of ballpark figure on this? I heard the AI traffic is still low, says this person. I'm going to try to like I'm going to try to like go quick and and hit hard with these. I think it's between two to 5%. I think that's like and that's like optimistic. I think for certain brands if they're like I don't know if you're a plumber or like waste disposal or like different industrials I know like we've worked with a few industrials it's so low it's like 2% it depends how tech forward your your market is also like we worked with the GPU company that rents cloud GPUs it was like 10% it was crazy um but that was their that's their audience like that's their audience is using those AI engines much more so it ranges but like I would be 2%'s fine great great question thank you Reena Right. Next question for this quickfire round. This webinar is for Fortune 500s, but do you have any insights that you feel would apply to local businesses? Content that speaks to the locality is going to be key. So like everything I just said, but like literally giving local focused answers. What's the best hair stylist in Boston? What's who's the best pet groomer in Charlotte, North Carolina? Like thinking about those kind of questions. If anything, they should actually be easier to win. like you should actually be able to take everything I just said, but being more hyper specialized. Like the hard part about working with like I don't know, a Nike is it's like, "All right, who's your market?" Well, everyone everyone needs shoes, don't they? And it's like, yeah, that's cool, but like all right. Um if I work with like a dog my dream client is like dog groomer Charlotte, North Carolina. We could do some create we could we could like rock the world. It would be really cool. We do some fun stuff. Yeah. Yeah. I mean, it's it's not authority based, so it'll be cool. Yeah. Right. Josh, you mentioned earlier that AI is changing usual authority signals. Is link building still going to be the main thing that you think needs to be done for authority or if not, what do you think it is going to be? Link building is not going anywhere and people are like there's this whole narrative that like everyone all these AEO professionals are like link building's dead. All you need is like mentions in the text. Typically, when you're mentioned and and that's good, like you can get by just by by being mentioned in the text, but typically when you're going to get mentioned in a piece of text that you're going to like partner with or pay for, you're going to get that inline link anyway. Um, like my hierarchy of belief is that like an inline link and a brand mention is awesome. Best like best case scenario. We'd love that. Um, the only thing that is like slightly more valuable now is just like a semantic mention of the brand without the inline link. There's actual value there for answer engines. don't you don't need the inline link anymore for answer engines but it's obviously a good to have but I think what people are missing the point on is like the semantic mention is is more valuable it doesn't take any value away from backlinks I think backlinks still to summarize that that answer back links are still especially in traditional search one of if not the single best way to really build up an authoritative mode I don't like other than time and lots of effort and lots of content nothing really attacks authority like a backlink Yeah, I honestly see all these things working together like semantic mentions, brand mentions, however you want to think about those backlinks, citation building and your more general PR when you do things like sponsor charities or do corporate away days and all of that sort of thing where you've actually got partnerships. All of this is about like what does the brand care about and what does the business want to do? What do their customers read? Where do they need to be seen? And all of that comes together. All of those things incredibly important for authority online and offline as well. Exactly. Um this this is a long one. The crux of it is chatbt it uses different metrics to perplexity has its own way of functioning. Right? So how can we optimize for all of these different LLMs if they're using different ways of working? This is a big question. So the different LLMs each are pulling from slightly different indexes and are are using slightly different sources of truth. I my perspective is that there is there's like four answer engines right now that are really important and they and the rest are good but like very situationally. My perspective is you've got your Google models and they they sit over in Google camp like camp Google is still using a lot of Google SER stuff. It's actually nice because if you want to win an AI search in like Google AI overviews and Google Gemini, you can kind of play the Google game a little bit at least like way more alignedly than you would a chatbt. So you've got your Googles over there. Google AI mode is the most weird of the Googles. Actually, people don't assume that. They think Gemini would be the weirdest of the Googles. Google AI mode is actually stranger than Gemini in terms of its search. Um, and then you've got your chatbts which are using like a thirdparty search provider. I would lump kind of I would almost lump like Claude in with ChattBT even though Claude's using Brave. Claude's actually really unsophisticated. I'll give everyone the tip here. The tip is Claude is pulling from the Brave browser and it pulls results one, two, and three. That's it. Uh if you want to be mentioned on Claude, just be one, two, and three on Brave. At least that was the that was the initial thing the test and quad initially did search. And then C-pilot's kind of weird. Copilot has a like a a secret version of the Bing index because I'll see I I won't submit a page to index now. It'll be a day old and or did I say cla co-pilot I should say Microsoft copilot. It'll pick up that content the next day before it should be indexed by anybody because I think it's using some pre-production or some dev instance of the Bing index. So copilot's very Bing focused. Um perplexity is very vector focused and they're they're caching pages a lot. So bet you know perplexity is just like basically saving pages to one big database and doing basically vector search on those pages and incrementally updating their cache. How do you do it all in one? I would focus on chatbt. I would focus on Google AI overviews AI mode Gemini. I would also focus on perplexity and a little bit on claude Grock meta Microsoft copilot. All are excellent but it depends on the business you're doing. Like Grock I work with lots of startups who are like we do so much work on X like X is so important to us. All right, let's optimize for grock. Otherwise, yeah, I don't know. I'm going to go towards chat GBT. I would say that's knowing where your customers are. So, if you know that your customers have to use like Microsoft Copilot or work, okay, fine. If you're actually not sure, it's probably Chat GBT because it just is the most dominant. So, it's probably the one to focus on, right? I didn't plant this question, Josh, but this is the the segue you need to explain what it is profound actually does. So, how do you test all of these things? How do you track them? This is getting a bit complicated. How do you check what comes up in AI search engines and is there an easy way to do this? There is an easy way to do this. This is um I mean this is why Profound was started. So Profound was started about 11 months ago to give marketers the opportunity to understand, take action on and tangibly improve their ad visibility. So we are our job is to interrogate answer engines, ask them millions of questions to find out what sources are showing up, what brands are showing up. We can show you what prompts people are actually putting into the answer engines. We have the first data set of its kind that's actually prompt volume focused. So it's almost like keyword prospecting but for prompts which is crazy. And we can track in live like minuteto minute the number of referrals indexing indexing events crawling events that these answer engines are actually having. So you can go to chat like in our platform you can go to chatgbt.com in a different tab search for your company open up our platform and within 10 seconds you'll see it hit that someone searched for your company somewhere in chat GBT and it'll show you immediately we've also got workflows it's it's a really it is the full platform for marketing in the age of AI especially AI search so it's it's extremely exciting yeah we're at trib definitely worth a check out we got self-s serve and plenty of cool stuff going It's a really fantastic tool and if you've not heard of it before definitely go and check it out. We love it at Exposure Ninja. We use it ourselves. It's very advanced and it's got a really really cool tool suite. So I'm sure you can go and have a look around then. I guess they can get a demo or just chat to you Josh if they need a hand. Yeah. Yeah, absolutely. We've got I mean we've got our we have an awesome sales team. We've got awesome account executives. I'm always around. So yeah, plenty of excitement there. We also have a an event ZeroClick 2025. That's on October 8th. the industry's first uh it's the first real AI search focused event in the world. It really is um it's really an opportunity for the leading figures in AI search to come together, learn, share strategies and a really kind of limited exclusive environment. Please register, check it out. Uh seats are extremely limited. We're trying to get a good headcount by the end of next week so we can send out acceptances. Uh but it's in person in New York. Unfortunately, so many questions about this. No virtual option, no recording. It is that kind of event. um in every way, but we're exciting. Yeah, it's exciting. That's incredible. Josh, you have been so fantastic. Thank you so much for sharing all of your amazing insights, so much detail. I know that people would have really appreciated it. If they want to follow you personally and find all of the insights that you share because you share so much stuff from your research, where can they do that? What should they do? They can do that on LinkedIn. LinkedIn is my place. Um LinkedIn, Josh Bliss. I think it's literally linkedin.com/joshuabbliskll.l. I don't know, but I'm always on LinkedIn. It is the best place to see me. The best of my research goes out on LinkedIn. Um, usually there's something every week. So, it's it's exciting place and we give you the raw data. So, it's very cool. Amazing. Perfect. So, follow Josh on LinkedIn. If you need an AI tracking tool, if you're thinking about where to get started, what you actually need to track, try profound.com is the place to go for that. If this is really overwhelming and you actually need a bit of a plan and some support with your marketing, you're getting into AI search, come to exposioninja.com and we will be able to help you out. We offer a marketing review so we can help you understand what prompts you even want to be tracking before you get into profound so that you're ready to go when you get there and figure out the plan going forward for actually dominating in AI search like some of these brands, then that's what we're there for. Amazing. Thank you so much, Josh. Thank you so much, everyone, for joining us today. All right.