The #1 Reason Your Brand is INVISIBLE to AI (And How to Fix It)

Episode Description

Is the era of the search engine over? Are we now living in the age of the Answer Engine?

In this episode, Frederick Vallaeys sat down with Kasim Aslam, a trailblazer who went from running the #1 ranked Google Ads agency to becoming a pioneer in the new field of Answer Engine Optimization (AEO).

Kasim reveals why he sold his 8-figure agency at the “exact right time,” warning that the old PPC agency model is a “hot potato” doomed to fail in the new AI-driven landscape. He argues that AI tools like Performance Max are making button-pushing obsolete and that the key to survival is to “own the prompt” by providing strategic leadership. Kasim also unpacks groundbreaking research from his new venture, aeo.co, including:

• The “BRAIN” Framework: A new methodology for getting your brand mentioned by AI like ChatGPT and Gemini.
• Google’s Secret Strategy: Discover why Gemini’s results have almost ZERO overlap with traditional Google Search and what it means for the future.
• The Most Valuable Website for AI: The surprising platform that accounts for 20% of Google AI’s citations (Hint: It’s not what you think!).
• The Death of the Keyword: Why clicks are becoming fewer, more expensive, and dramatically more valuable.

Episode Takeaways

Kasim and Fred unpack the realities of marketing and advertising in an AI-first world. Kasim reveals how his agency went from 200 clients to 130 in nine months post-sale, why Performance Max “ate agencies alive,” and how the entire client-agency relationship must evolve.

Whether you’re an agency owner fighting to stay relevant, a brand trying to get discovered, or a marketer preparing for the future, this episode offers a rare glimpse into the seismic shift happening right now—backed by real data and actionable insights.

The BRAIN framework

Kasim Aslam introduces BRAIN (B-R-A-I-N) as a working theory to improve visibility within large language models (LLMs). It is built on a foundation of substantial data and experimentation.

There are three components to it and each of them is essential for optimizing how entities are surfaced and referenced by AI models.

“The BRAIN framework breaks up into—it’s a three-parter. BR is brand, A is authority, IN is index. Your brand has to be strong, your authority has to be strong, your index has to be strong.” explains Kasim.

Here’s how the core components work:

• Brand: New form of link-building in the AI era. Mentions matter more than backlinks. Being referenced across platforms like Reddit, Quora, IMDb, and TikTok contributes to a brand’s AI awareness, even if those mentions aren’t hyperlinked.
• Authority: LLMs prioritize individual thought leaders over corporate brands. Personal authority carries more weight than company names, because people are more likely to be discussed, cited, and viewed as knowledge sources.
• Index: Technical accuracy and consistency matter. Factors include schema markup, directory alignment, and consistent contact details—ensuring that structured data feeds LLMs the right signals.

Google’s secret strategy: Why Gemini has almost zero overlap with traditional Google search

When analyzing overlap between AI-generated answers and traditional organic search rankings, Kasim Aslam reveals a surprising trend:

• ChatGPT overlaps with Google/Bing search 30–70% of the time (higher for local queries).
• Perplexity overlaps 40–80% of the time.
• Gemini/Google’s AI? Almost 0%.

This suggests Google’s AI is not referencing its own search rankings— clearly different from what others like OpenAI and Perplexity are doing.

Unlike other LLMs that scrape or simulate organic results, Google’s Gemini isn’t built on traditional Google Search infrastructure. Instead, it’s forming responses from a completely separate knowledge system, independent of search rankings.

“Here’s what that means. I think it means that ChatGPT, Perplexity, all these other LLMs are building on Google search. They’re just searching Google faster than you. Google is building their own knowledge graph. They’re not using Google search.

Google’s agentic AI-driven information isn’t being pulled using what Google ranks websites for because we’ve known Google organic ranking is flawed and gameable. We’ve known that for 20 years.” shares Kasim

Kasim feels that this strategy makes Google uniquely positioned for long-term dominance in AI search. While competitors rely on “antiquated” systems built atop Google Search, Google is discarding its own legacy search engine to grow a new, AI-native model.

The most valuable website for AI: Reddit accounts for 20% of Google’s AI citations

Another interesting insight that surfaced during the discussion is that despite its modest size and niche history, Reddit has unexpectedly become the most influential site in AI-generated knowledge.

“20% of citations from Google’s AI come from Reddit and 18% come from YouTube.” said Kasim.

LLMs value Reddit because of real human insights, topical depth, and the upvote/downvote system which creates a natural quality filter. This helps LLMs prioritize high-value information.

If you want to influence AI outputs, building authority on Reddit is emerging as a powerful AEO (Answer Engine Optimization) strategy.

Another point to note is that although YouTube dwarfs Reddit in terms of traffic, content volume, and user base, Reddit still leads in AI citation share indicating that AI models prefer structured, text-rich, context-heavy sources. Something Reddit offers in abundance.

The death of the keyword: Fewer, more expensive, more valuable clicks

The digital landscape is moving from search queries based on keywords to conversational prompts driven by large language models (LLMs). With AI agents that retain memory and answer queries with increasing sophistication, the role of the keyword in initiating buyer journeys is rapidly diminishing.

Chat-based AI interfaces lead to fewer but more decisive clicks. This shift radically alters click economics:

• Consumers engage deeply with AI tools like ChatGPT and Gemini before clicking.
• As a result, each click is more informed and closer to purchase intent.
• This reduction in volume creates scarcity, driving up the cost per click (CPC).

Today’s AI-trained user doesn’t click casually—they click only after exhaustive research via AI chat. This means clicks now signal readiness to buy rather than just curiosity.

“If somebody’s already problem aware, if they’re solution aware, you’re kind of screwed. Unless you can pay to play and most people can’t. If they’re problem aware, it’s already a little dangerous.

You need to find a way to get your hooks into people way before they need what you got. And that’s a different type of marketing. That’s like the Michelin star old school. It’s weird how everything that’s old is new again.

It’s like old school Don Draper style thinking marketing. It’s like how do I get in front of somebody before they’ll ever need me?” said Kasim.

As clicks become rare and expensive, pre-intent engagement becomes the new battleground. Marketers must focus on capturing attention pre-intent, building brand affinity before the buyer realizes they have a need.

This requires storytelling, community building, and brand-led content.

Episode Transcript

Frederick Vallaeys: Hello and welcome to another episode of PPC Town Hall. My name is Fred Vallaeys. I’m your host. I’m also the CEO and co-founder at Optmyzr, a PPC management software. So today, we wanted to talk about something a little bit different than PPC. Specifically, we’re going to talk about the evolution of how searches are done. Are people still searching or are they now looking at answers? Are we talking about answer engines instead of search engines?

Now to help us develop that topic, we’re going to talk to someone who actually does have quite a bit of a background in PPC, Kasim Aslam, and he has made the transition from PPC specialist to answer engine optimization specialist. So with that, let’s get rolling with this episode of PPC Town Hall. Kasim, welcome to the show.

Kasim: Good to see you again, Fred. Always a pleasure. Thanks for having me back.

Frederick Vallaeys: All right. Well, so for the people who haven’t seen you on one of the previous podcasts we did, remind them who you are, your stature in PPC. I believe you were once ranked—was it number one or number two?

Kasim Aslam: No, you know it was number two because I was right behind you. That was a sneaky way to bring it up. Yeah, I was number two in the top 50. I think I built the number one ranked Google Ads agency in the world though. Our ad agency, if you searched anywhere in the English-speaking world, we were the first agency mentioned in the organic listings. I sold it in 2022 for an eight-figure sum of money. I had $100 million in ad spend under management. We got as big as an agency can get when you’re dedicated to Google ads—we kind of tapped out that market and then I jumped ship at the exact right time, to be honest with you, because it’s been an interesting ecosystem and I know you probably know that better than anybody.

Frederick Vallaeys: Well, let’s jump right into it. So, you said a good time to sell given all the changes, but let’s unpack that. What do you mean by that?

Kasim Aslam: My agency died on the vine. When I sold I had 200 clients. Nine months later, I think we had 130. And it wasn’t the acquisition party’s fault at all. It was the model that we were running which was an old school model. It was fee plus percentage of spend and then very hub-and-spoke client manager and ad specialist. It didn’t really work any longer.

And the reason—and I’m about to slander our collective environment so you can stop me if I say anything that you feel is either not true or just stupid—but what organizations started to realize was I actually don’t need you to press the buttons in the dashboard anymore. I need somebody outside. So it’s pre- and post-click. And even if they wanted and needed us to press the buttons, in so many instances the buttons kept going away.

It was right as Performance Max was really reaching a pinnacle. I think we were the first agency to adopt Performance Max. I still had the highest traffic Performance Max YouTube videos that existed. I think we cracked the code there in a lot of ways and Performance Max ate ad agencies alive because it took so much of our ability to optimize. It was a black box and so much of the black box was reselling our own traffic back to us, but in ways that you couldn’t see unless you were playing things like attribution and post-click and follow-up—and probably a lot of the stuff that you do at Optmyzr.

So for a standalone agency that wasn’t playing an AI-driven game, and we probably should have been, we were caught so flat-footed. It was nuts. And people were bringing things in-house. They were bringing on full-time marketers. And they were looking for just more of a holistic solution. And I just lucked out. I was playing a game of hot potato. I didn’t know I was playing. And I happened to hand my hot potato over to somebody else before it kind of blew up. So I’ve seen other agencies that were kind of caught where I was caught.

Frederick Vallaeys: Well, congratulations on making that move at the right time.

Kasim Aslam: Yeah, I’m just blessed.

Frederick Vallaeys: But many of our listeners today, they still hold the hot potato themselves. And it certainly has been challenging for agencies due to the reasons that you’re explaining. Now I think when it comes to Performance Max, it actually has gained a lot of the settings, a lot of the insights reports to give the advertiser more control over the whole thing. But I think the level of control that you have and the level of insight you need to have is quite fundamentally different from the very simple PPC of the old days.

For example, when it comes to the new channel reporting that Google does, you can’t just look at how is YouTube performing versus search because they are two different beasts. They should work together to give you better results. So if you split it out, you’re not getting a holistic picture and then you might end up making some bad decisions. But that said, more of that data is back. So the smartest advertisers and the smartest agencies, they can absolutely do something with this if need be.

But I 100% agree with your point about people don’t hire agencies to push buttons for them. That button pushing should be done by agents though. These are AI agents. That’s the level of work that they’re going to be really good at and that’s where you should look to automation. But explaining to the agent what is your business goal and what are the constraints of your business and how should we approach new experimentation—that’s really where the value still lies. So I think that’s hopefully where agencies still have a little bit of an opportunity to distinguish themselves and maybe stand out from the pack. Any thoughts?

Kasim Aslam: I think an agency in order to be valuable in this ecosystem needs to—I’ll phrase it this way—they need to control the prompt. So, we’re all being taught to be prompt engineers, which I think is superfluous and going to go away very soon. And we’ll use that framework because it’s something that’s fresh in everybody’s mind. The thing about prompt engineering is the value in the entire endeavor, the entire sequence is in the prompt. Everything that happens after that is a trillion-dollar AI mechanism that neither you nor I have any influence on. And as AI gets better and better and better and better and better, that will become more and more and more true.

So the value is in the thoughtful human who decides there’s an end goal that I want to reach and I’m going to prompt this goal. If you as an agency are the one being prompted—so if the client’s coming to you saying we need to find a new service line or we need to improve the conversion on this product or we need to figure out how to increase our retention—you are going to be replaced by AI. The question is when. Like guaranteed 100% you are superfluous to the process and you exist only because they can’t prompt the computer to do your job yet.

So if you are being prompted, it’s forced obsolescence. If you’re the one prompting, that’s where the value is. That’s when you come to the client and say, “Hey, listen. We need to do this. We should look at this. We’ve been examining like that. Who’s the one who’s pushing the endeavor forward?”

And so, whoever controls the prompt is—and I own a staffing agency and what I—we staff EAs for companies. What I tell my EAs is if you’re the one being prompted, you’re going to lose your job. You have to be the one going to the client saying, “Hey, I really feel like we need to improve our brand continuity visibility because I just took this new AEO course from Kasim and he showed us that the brand continuity and visibility is the one thing that’s the key performance indicator for AEO mentions and so I’m going to go do this, this, this, this, this. I own the prompt.”

So, if you’re an agency owner and you’re watching this, own the prompt. Be proactive. You can’t sit back and be an order taker. And that’s the thing that I—because I’ve done a lot of agency coaching—I see these agencies that are doing everything right and losing clients and they’re like, “We know what to do. We do everything they ask us.” It’s exactly right. You only do what they ask you to. It’s worthless. What is that worth? Especially when I’ve got ChatGPT. You have to do more. You have to do more and be proactive. Own the prompt.

Frederick Vallaeys: Yeah. No, I love that. And I’m going to take that idea of if you’re being prompted, that’s worthless. I’m going to take that to my whole team. I’ve been thinking about similar concepts, but the way that you put that I think is really succinct and spot on. Now, let me take that idea that you had and I’ll spitball something I’ve been working on.

So, Jamie Stevens, who’s a senior or chief scientist at Microsoft, she gave a really cool talk at the Cannes Lions marketing festival and she basically talked about the future of work and much like you, she’s saying it’s about the prompt is the thing that sets you apart. It’s not about actually doing the work. It’s knowing what you’re asking for that the work should be.

So, it’s not about coming into the office and saying, “Today, I’m going to do pivot tables.” You never do that. But you do pivot tables for a reason. Like, are you doing pivot tables because you have budget questions? Great. But now, instead of doing the actual work, formulate to the AI what it is you need to do. But let’s take that one level up and let’s imagine now that when you have a client meeting, why couldn’t that client meeting be the actual prompt development?

So the half hour you spend talking to the client, figuring out what is the client wanting to achieve? Well, ask them about the constraints that they have. Ask them about what they’ve done in the past. Ask them about things that have worked in the past. That transcript from that half hour call with the client, that could be a beautiful 10,000-word prompt, which is the best prompt you’ve ever had. And now, unlike typical agencies, you come out of that meeting and like, now I got to actually go and do stuff. I got to do the work. You give it to the AI and the work is done and you move on to the next one.

And your value was in asking the right questions and building that beautiful prompt.

Kasim Aslam: Yeah, I think that’s beautifully stated and that’s exactly right. The truth is clients don’t know what they want or need. And those are two different categories. They don’t know what they want or need. And it’s up to us to kind of pull that out of them and as much as is possible to also inform them of both. You can walk in and say, “This is what I think you need.”

The risk that you run that agency owners aren’t used to—you have to step into an environment where you force the loss of a client. And we’re because we’re so obsessed with retention. So it’s like I’ll do anything you want, just keep paying the bill. And I think that’s wrong. I think if we assume this new paradigm where we’re the ones that are leading, it’s like here’s where I think we should go. And the client goes, I don’t want to go there. And then we get to say that’s—if we don’t go there, I don’t think that we should continue because I’m the one that’s driving and I’m going to be held accountable to these results on a long enough timeline.

And so the way to look at that is if you tell the truth, whatever happens afterwards is the ideal outcome. If you lose the client, you’re losing a client that you’re going to lose on a long enough timeline, but you do so within the confines of a controlled environment. And then the clients that you retain, you retain so much stronger because once somebody follows you once and it works, they will follow you to the ends of the earth. Like you’ve proven this model and this concept.

Frederick Vallaeys: Yeah. And you also have to have a client who’s willing to do the experimentation and sometimes fail. But the key is in failing quickly and learning from it and moving on to the next thing. So if you can get your client to follow you along the way with that, that’s fantastic.

Kasim Aslam: Yeah, that’s brilliantly stated. It’s teaching them what failed. The classic example is an email marketing campaign that fails, everybody immediately goes into copy and they’re like, “Fix the copy, fix the copy, fix the copy, fix the CTA, fix the landing page.” Well, did the open rate fail? Because then we just need to fix the subject line. But so often we just zoom in on this global failure and say like, “Oh, email doesn’t work.” No, it never landed in the inbox, deliverability was bad, or it landed in the inbox and it never got opened, the subject line was bad, or it got opened, they couldn’t pull it because your images aren’t loading.

There’s so many things along the way. Figure out exactly what failed and then inform them of the failure and then decide what’s next. And that’s what agencies aren’t doing. They’re like, “Hey, email failed. Subject line didn’t work. What do you want to do?” And it’s like, well, let me go ask ChatGPT then, and I’ll ask myself the whole time why it is I’m paying you.

Frederick Vallaeys: Yeah. No, and listen, I think knowing when to use ChatGPT or when to go to Claude versus OpenAI versus Gemini and then knowing how to prompt it, that’s really the magic right now. And I think a lot of clients will eventually pick up on that. But I’ve always tried to push the envelope with this stuff and build my own implementations that connect certain systems together and run them through Zapier and you kind of learn a lot of stuff along the way and it’s really cool what you can do.

Kasim Aslam: Oh yeah. Yeah. Does your audience know you invented conversion tracking? Have we talked about that?

Frederick Vallaeys: Oh no, but thank you for mentioning it.

Kasim Aslam: Yeah, because it’s funny. We figured that out on—you were on Perpetual Traffic and we figured that out live on the call. Like you were so humble about it. You were like, “Well, when I was at Google, we were building this thing that did this thing that allowed this thing.” And I was like, “Hold on. That’s conversion tracking.” You’re like, “Yeah, I guess it is.” So, like you invented conversion tracking for anybody who doesn’t know this about Fred. Like, probably the most important part of paid advertising came from you and your time at Google.

Frederick Vallaeys: Yeah. Well, well, thank you. I think you’re slightly overstating it. I mean, I think conversions have definitely been understood. I just put together a technical implementation before that feature was available in AdWords at the time and then some product managers were like yeah that’s pretty cool maybe all advertisers should be able to measure conversions and conversion tracking. So then we launched that to everyone but I won’t take sole credit for that but thank you. It just you reminded me of it when you said I like to build my own implementations because I remembered that like that was your own thing that they basically—oh gosh yes obviously we need this.

Kasim Aslam: Well and the one that I’m building right now is the notion of using multiple custom GPTs to help me improve blog posts that I write. So rather than giving it one GPT and saying hey can you make this better? I’ve built different personas. So one of them is a skeptical reader, one of them is a fact checker. So I have a long list of that. But the problem is you have to go to ChatGPT and you have to sequentially put in the blog and then say can you give me feedback skeptical reader okay now a fact checker can you go and fact check this and then at the end of thing you can say okay consolidate all the feedback into three key bullet points that I need to know but that still requires me sitting at the computer and doing this.

So I used a little bit of AI coding to build a custom implementation of that so that I can just say here’s the blog, here’s the six personas that I want to involve in checking this and then it just runs and a couple minutes later comes back to me and says here’s kind of like the consolidated feedback that all six personas have agreed upon is the most important to make this a better blog post. Make sense?

Kasim Aslam: It makes sense. I love that. So two things are you—what are you using? Lovable? Bolt?

Frederick Vallaeys: Yeah, I was using V0 for a long time. So the first version I built in V0, the second version I did in Lovable, and it literally was two prompts.

Kasim Aslam: Dude, Lovable’s blowing my mind right now. So I love the idea of the different personas. The thing that occurs to me, not that you asked for my help or my input, but I wonder if there shouldn’t be a moderator. So you make a custom GPT that decides because the thing that scares me about the personas is you’re always going to get a persona that overwhelms. But if you could get a moderator that decides based off of the persona input like where—because that way it’s not decision by committee. But I think that’s really brilliant. I’d love to see that.

Frederick Vallaeys: Yeah, I really like that because and again I think what we’re illustrating here is take a simple concept that most people have heard about like using a custom GPT to do something but then remove the constraints of why can I only have one custom GPT? So now I got six GPTs but I had given myself the constraint of these six GPTs need to sequentially talk to each other when in fact like you’re saying maybe it should be more like a human round table conversation that they’re having and that requires a moderator and that’s absolutely possible.

Kasim Aslam: That’s really cool that you’re doing that, man. My buddy Perry Belcher has a—I don’t know what to call it—it’s like a council that he’s built that consults with him on business endeavors. He builds businesses. It’s his and so he bases it off of real people. So he has like a Rockefeller and he like loads it up and he’s like, you know, you’re shrewd, you think big, it’s large scale, it’s always hostile takeover, whatever. And then he’s like, and then he’ll have Rockefeller like audit his business plan. And they’re really mean. The GPTs will come in and like it’ll say something like, you know, rolls eyes and questions why you haven’t killed yourself yet. And then gives like really caustic, you know, because he’s trained them so well. But it’s interesting. He’s he hasn’t done it with content. He’s doing it with like business and marketing models. But I think that’s really cool.

It’s to use these things in a way that’s not commoditized. You have to go a layer deep. Everybody can prompt ChatGPT, but not everybody’s going to take the time to create six different custom GPTs and then get them to talk to each other and themselves. So, I think that’s really cool, man. Good for you.

Frederick Vallaeys: Thank you. Yeah, let’s maybe transition here into what you’re up to next. So, your PPC agency obviously seeing the light about AI is changing everything. And the key thing that it’s changing that we want to start with here now is it’s changing how people find things. So talk me through what you perceived and then let’s go into what that means for the next business that you’re working on.

Kasim Aslam: This started with our staffing agency. I own a staffing agency called Pareto Talent and we train executive assistants and we place them with entrepreneurs. And I was talking to ChatGPT about something and at one point ChatGPT came back with Pareto—with my business and it was this moment where I was like oh my goodness. Like I’ve really made it if I’m—I was asking something about like hiring in Latin America and it was like well if you want a high-value EA out of Latin America Pareto Talent.

Now I have found out subsequently because we’ve studied this—I have five employees that have done nothing for the last eight months but study how LLMs source citations and a lot of it is heavy confirmation bias. So I now know it had a lot to do with what I was feeding it. And I realized in that moment I was like, “Oh, this is like ranking number one on Google. I’m just ranking inside of ChatGPT.” And it sent me down this rabbit hole where I paid many thousands of dollars for the domain name aeo.co .

And I’ve launched a community because I’m not ready to play agency yet because I don’t know any of the answers. All I have are questions. I have questions and I have data. And so talking to some buddies, I decided the best way to do this is just take all the smartest people that I can get to have eyes on this, put us all in a room together, and then talk about what it is that we’re seeing in terms of common denominators.

So I developed this thesis through another business that I own, and then I hired some people to—all we do is and I don’t know how much of the screen share we use, Fred, but I can show you the data analysis that we’re doing. I have more data on this I think than any human on the planet except maybe Monik Bond who’s a really sharp guy.

So for the people who are listening to this as a podcast, we’ll try to walk you or talk you through what we’re sharing on the screen, but it’ll also be available in the show notes. So this is one example that I used actually at a recent talk. I have four employees that do the study, one person that oversees the four employees. This is one day in time. So this is June 11th and the gentleman’s name is Matan.

Every single day they do 20 queries. So the first query is “best boutique coffee shop in Kingman, Arizona.” And they take that query and you notice that there’s a couple of specifics here. This happens to be the local brick-and-mortar study. We’re doing local brick-and-mortar, local service, ecom, SaaS, informational queries, etc. So this is a high commercial intent query and we plug it into three separate LLMs. In this particular instance, we use ChatGPT, Perplexity, and Gemini.

So if you’re listening, all we did was take the query, “what’s the best boutique coffee shop in Kingman, Arizona,” and we asked ChatGPT, we asked Perplexity, and we asked Gemini. That’s one query. Well, he does 20. “Where to find specialty coffee in Kingman,” “the top artisan coffee cafe in Kingman,” etc., etc., etc. So we do this 20 different times.

With every query, we look at the answers, but most specifically, we look at the citations, which means the websites that the LLM used in order to develop the answer, and we place those in chronological order, and then we study those URLs based on every level of analysis you would assume that we would study as from a classical SEO perspective. So what’s the URL matching, keyword matching, keyword density, domain rank, Ahrefs rank, traffic value, referring domains, backlinks, page speed, schema validator, etc. And then we also map it against Google position and Bing position organically.

So I have hundreds of millions of data points on how LLMs are capturing the citations that they’re using. Here’s where things get absolutely crazy, Frederick. It’s blowing my mind right now. The single highest key performance indicator for ranking organically has been for the last 10 to 15 years links. Inbound links. Guess what the single highest key performance indicator is for placement in LLMs?

Frederick Vallaeys: Links.

Kasim Aslam: Schema markup. Schema. So the more robust your schema, the more likely it is that you get pulled by the LLMs. And they don’t like JavaScript schema. They like hard-coded schema now. I don’t know what the hell schema is. I had to Google it. I haven’t heard schema for 20 years since I built websites 20 years ago and I remember but LLMs are pulling heavily from schema.

Here’s the other thing that’s crazy. When you look at the overlap, this will blow your mind. I hope it blows your mind because it blew my mind and it’s actually scary and exciting at the same time. When I look at if I ask ChatGPT a question and I look at the overlap between ChatGPT citations, Google rank and Bing rank, ChatGPT overlaps with organic search about 30 to 70% of the time depending on the type of search. More for local, less for top of funnel, but 30 to 70% of the time.

Perplexity overlaps with organic search 40 to 80% of the time. Guess how often Gemini or Google AI-driven engines overlap. Guess.

Frederick Vallaeys: Throw out a number. I’m going to say it’s lower. It’s 20 to 40%.

Kasim Aslam: You—I love you so much. You’re the first human I’ve had this discussion with that actually guessed that. You’re the very first one. Frederick, it’s zero. If there’s any overlap, it’s negligible.

Here’s what that means. I think it means that ChatGPT, Perplexity, all these other LLMs are building on Google search. They’re just searching Google faster than you. Google is building their own knowledge graph. They’re not using Google search. Google’s agentic AI-driven information isn’t being pulled using what Google ranks websites for because we’ve known Google organic ranking is flawed and gameable. We’ve known that for 20 years. It’s for as long as you’ve been able to do SEO.

So when people look at it and say like, “Oh, Google’s not as good as ChatGPT or it’s not as good as Perplexity.” They’re wrong. Here’s why. ChatGPT and Perplexity are showing you a rainforest, but they’re built off of Google’s old sinking island. Google is showing you a sapling that they’re growing from brand new. And Google’s the only one except AWS, but AWS wouldn’t be able to do it because it’d be violation of the terms of service. Google is the only entity on the planet that has the knowledge graph to do this.

So I’m so long on Google because all other LLMs are using an antiquated model which is Google search and Google has thrown out their own search and is actually building delivery completely organic grassroots from the ground up. So who knows how long it takes for them to catch up but I think Google is going to dominate on a long enough timeline.

Frederick Vallaeys: That’s fascinating. Is it crazy?

Kasim Aslam: It is crazy. And listen, I also tell people be bullish on Google and it’s not because they’ve stumbled out of the blocks and were very conservative—it’s not because they don’t have the technology. It’s because they have so much regulation and oversight and they have to be very cautious in what they do. But when it comes to actually what’s happening behind the scenes, they got the best in the world, I believe. And at some point, if they truly do fall behind, they’re just going to say, “Okay, let’s go for it. We need to reclaim the market here.”

And also keep in mind, I mean, they invented DeepMind. They invented the models that drive GPT and the transformers. So, it’s not like Google doesn’t know what it’s doing. And for you to share here what you’re seeing about this sapling coming from a sinking island, I think that’s brilliant.

Kasim: Google historically has let other companies take the arrows. Here’s the craziest part. Meta when the congressional hearings happened about all the privacy concerns, Meta had 55,000 demographic and psychographic profiling factors and they got pulled in front of Congress because they let people choose those profiling factors in their advertising. At that time, Google had 72 million and all that happened is they just stopped publishing that data. The only difference was Google didn’t let people segment and Meta did. But Google was a way bigger threat to privacy than Meta ever had been.

But they do this thing that’s absolutely brilliant, which is they wait to see what happens from a legislative perspective, public temperament perspective. They want to see where everything lands. So I think, not that I know what I’m talking about, but I’ve spent my whole life studying Google—that’s all I did. I think that they’re just waiting in the wings. And it’s pioneers get shot, settlers prosper. Google’s going to come in and they can afford to wait a decade. They can hemorrhage money and wait a decade. I’ve only ever lost money in the stock market and if you want—I think Google’s a strong buy.

They have so much data and they have more data and they have the type of data that’s so perfectly rich for an LLM. Facebook Meta has what you tell it. Google has what you ask it. Like Meta’s data is so artificial in construct because it’s nothing but a bunch of people posturing about what they wish were true. Google has all the information that is true and they have the ability to do it in a peer-reviewed fashion. So anyway, sorry that was a bit of an aside.

Fred: Well, and I think it’s really interesting. So, couple of things. Google has more data points, but if you had 72 million data points as an advertiser, like where would you even start? So clearly we need the help of the machines to tell us what to do with this and then there’s a level of trust like instead of giving you some base level control over that why not just have the machine learning models figure it out and show the ad at the right time at the right price to the right consumer. But what that is premised on is you do still need to tell Google what your business outcomes are that you desire. That’s how they connect the dots in the machine learning models. So yeah, that makes total sense.

And then the other thing I’m basically positing is that the keyword is dead. And the keyword is dead because we were shifting from keywords to prompts. But there’s more to it than that. There’s also the fact that the large language models and the chat agents, they have a huge memory. And so it’s no longer just about the thing that I’m searching right now today, but it’s also about the things it knows from me from when I walked into a store and I was wearing my AI computer that’s constantly listening and sometimes I’m wearing my Ray-Ban Meta glasses and so I can take pictures of the things that I like.

And now when I go to Google and I—it was actually a cool example that somebody was complaining about. They were saying when my ad shows in AI mode for personal injury attorney. Now I see that sometimes Google shows my ad when somebody’s looking for dentist. Okay, dentist is the search. Personal injury attorney is the campaign that’s showing. And I get why that’s concerning and why you think that’s a mismatch. But what if Google actually knew that you got into an accident yesterday, your teeth were messed up, you’re kind of concerned about paying the bills for your dentist, and this search for a dentist now all of a sudden’s like, “Wow, did you know that the personal injury attorney can help you recover some of that money and do a lawsuit?”

Kasim: So to take what you’re saying, one of the most valuable prospects on the planet is a pregnant woman. Nobody spends more money with greater fervor over a longer period of time than a pregnant woman. And fight me if you want to, that’s a data driven truth. So if a pregnant lady is shopping for anything, there are so many businesses that are willing to bid up the cost of that click beyond what most businesses [can afford]. So now we put ourselves in this kind of weird hierarchical structure where the advertiser willing to spend the most for the avatar wins across every level of analysis. And there’s going to be all these advertisers that can’t actually play that game. So they’re pushed off into the periphery. That’s really interesting. The implications I think are going to be pretty staggering there too.

Frederick Vallaeys: Right? Then you take a look at that example now where you’re basically saying that the CPC becomes untenable for most advertisers unless you know who you’re targeting. And then there’s the whole notion and Microsoft shares this and Google shares this, but people do fewer clicks when they use a chat or an LLM. And so the click by nature should become more expensive because the click—the reason there’s fewer clicks is the user, the consumer is having a conversation with the chat mode AI mode in Google or ChatGPT whatever it is and they’re so well informed by the time they’re ready to click.

Now that click is maybe 10 times as valuable as it was before because in the past that click was an indication of like, oh yeah, maybe interested. Let me find out more about your business. Now that click comes after they know everything about your business because the chatbot has answered all of the questions and they’re fundamentally just ready to buy. They’re looking for the place to go and put it in the shopping cart and transact. And even that, by the way, I say landing pages are dead because who in the world wants to go to a landing page to put a thing in a cart and actually put in their credit card? Like that’s annoying. Just tell the chatbot, I’m ready to make a decision, have it delivered to my house tomorrow.

But that also means that these interactions when they do happen, they’re going to become much more expensive. And so that’s something advertisers do absolutely need to be ready for.

Kasim Aslam: There was to the point that you just made there’s an article on Ahrefs about AI-driven traffic. Here’s the title: “AI search traffic converts better. 0.5% of visitors drove 12.1% of signups.” And then to read the take-home message, “AI search visitors convert at a 23 times higher rate than traditional organic search visitors for Ahrefs.”

So you’re exactly right. Like it’s going to be way less traffic and far far far more potent. I think we’re already in a traffic bubble for most businesses. Like that’s subjective, depends on what it is that you’re selling. And I think that AI is going to put us in a position where people who aren’t playing some level of community game, meaning they’re not trying to capture a prospect before intent or awareness exists. I think they’re going to be in a lot of trouble.

If somebody’s already problem aware, if they’re solution aware, you’re kind of screwed. Unless you can pay to play and most people can’t. If they’re problem aware, it’s already a little dangerous. You need to find a way to get your hooks into people way before they need what you got. And that’s a different type of marketing. That’s like the Michelin star old school. It’s weird how everything that’s old is new again. It’s like old school Don Draper style thinking marketing. It’s like how do I get in front of somebody before they’ll ever need me? Which will be a fun game to play on the internet for sure.

Frederick Vallaeys: Well, and so that’s so interesting, right? Because PPC being so low-funnel typically. But then you see demand gen campaigns and Performance Max campaigns which enable you to take those existing assets and move up the funnel to put those hooks in a little bit earlier. But then you also think about influencer marketing where I start following people because I like what they talk about and then at some point, boom, they show a product and now it’s like, oh, I didn’t really particularly care about that product, but I will take a look at it because I trust the person or I like the person who’s hawking it.

So let’s talk a little bit more about aeo.co and the community you’re building around that and the advice that you’re giving people as far as how they maybe put their hooks into the customer earlier on. I think there’s some methodology that you were explaining. So talk about that a bit.

Kasim Aslam: Yeah. So we’ve developed a framework that we call BRAIN. B-R-A-I-N. It’s the BRAIN framework. And the BRAIN framework is our thesis to ranking within LLMs. The disclaimer I’ll make Frederick is nobody knows. All these morons coming out saying, “I’m an AI expert.” It’s like, “Really? Are you? You got your 10,000 hours in?” Like you just jumped on the rest of us. So this is best guesses, but it’s best guesses based off of immense amounts of data.

The BRAIN framework breaks up into—it’s a three-parter. BR is brand, A is authority, IN is index. Your brand has to be strong, your authority has to be strong, your index has to be strong. Indexing are things like schema markup, making sure that all your directories align and the contact information is the same on your website, that type of thing.

Brand, interestingly, feels like the new link building. You need to be mentioned, but it doesn’t have to be a hyperlink. If somebody’s saying Frederick Vallaeys on Reddit and then Quora and then IMDb and then a TikTok comment, that is being compiled into kind of this repository of like, oh, Frederick must know what he’s talking about along this level of analysis.

And then the authority of course is the content that you’re producing. But the content is solution agnostic. It actually reminds me of what Google+ was trying to build 15 years ago. You remember that when they wanted everybody’s +1 URL added to whatever it is they wrote because they wanted to disassociate from the website and move into authorship. That’s what the LLMs have been able to do.

Interestingly, and this is good and bad, LLMs are massively prioritizing people over brands. So if Frederick Vallaeys publishes an article that means something. Optmyzr publishes an article means a lot less because people are going to mention you in Reddit, they’re not going to necessarily mention Optmyzr at least not the same way and they’re going to reference you because a person can get a PhD and a person can be an authority but a brand is a brand is a brand is a brand. Rand Fishkin sold Moz. So like at a certain point it’s like I trust Rand, I don’t trust Moz and that’s the—it’s going to be a harder game for people to play I think. You have to be willing to be visible. You have to be willing to establish authority.-N. It’s the BRAIN framework. And the BRAIN framework is our thesis to ranking within LLMs. The disclaimer I’ll make Fred is nobody knows. All these morons coming out saying, “I’m an AI expert.” It’s like, “Really? Are you? You got your 10,000 hours in?” Like you just jumped on the rest of us. So this is best guesses, but it’s best guesses based off of immense amounts of data.

The BRAIN framework breaks up into—it’s a three-parter. BR is brand, A is authority, IN is index. Your brand has to be strong, your authority has to be strong, your index has to be strong. Indexing are things like schema markup, making sure that all your directories align and the contact information is the same on your website, that type of thing.

Brand, interestingly, feels like the new link building. You need to be mentioned, but it doesn’t have to be a hyperlink. If somebody’s saying Fred Vallaeys on Reddit and then Quora and then IMDb and then a TikTok comment, that is being compiled into kind of this repository of like, oh, Fred must know what he’s talking about along this level of analysis.

And then the authority of course is the content that you’re producing. But the content is solution agnostic. It actually reminds me of what Google+ was trying to build 15 years ago. You remember that when they wanted everybody’s +1 URL added to whatever it is they wrote because they wanted to disassociate from the website and move into authorship. That’s what the LLMs have been able to do.

Interestingly, and this is good and bad, LLMs are massively prioritizing people over brands. So if Fred Vallaeys publishes an article that means something. Optmyzr publishes an article means a lot less because people are going to mention you in Reddit, they’re not going to necessarily mention Optmyzr at least not the same way and they’re going to reference you because a person can get a PhD and a person can be an authority but a brand is a brand is a brand is a brand. Rand Fishkin sold Moz. So like at a certain point it’s like I trust Rand, I don’t trust Moz and that’s the—it’s going to be a harder game for people to play I think. You have to be willing to be visible. You have to be willing to establish authority.

So, we’ve got this community. I don’t think people should join it yet. I’m not here to sell the community per se because it is so new. So, if you want a blueprint with a checklist and a promise that this is going to work, like hold your horses. That’s 90% of people. If you’re in the 10% of cowboys that want to come in and play and just shake this thing up and figure out what comes out of it, that’s who I’m looking for.

So, the reason that I’m kind of on this tour is 90% of folks, hang tight. I promise we’re going to tell you what works when we know. But this isn’t for you. For the 10% of you that are willing to be guinea pigs, willing to see 99% of what you try fail, but get excited about the 1%. That’s who I want to join the community.

We’ve published a book. You can have it for free. It’s aeo.co . And I’m going to as soon as you download the book, my promise is every time we update the book, we’re going to give that away for free. And the book is going to be everything we’ve learned so far. And some of it’s really actionable. I’ll give you an example. Here’s the craziest part. Guess what the most valuable website on the planet is as it pertains to being mentioned?

Frederick Vallaeys: Reddit.

Kasim Aslam: Reddit. You’re so good at this, Frederick. Gosh, you just keep knocking things out of the park. Reddit, get this. 20% of citations from Google’s AI come from Reddit. 18% come from YouTube, 20% come from Reddit. Reddit, which used to be like the ugly stepchild of the internet. It’s where all the incels go to complain about things. Like, how did Reddit become the most important site on the internet?

So, what that means is if you really want to play the AEO game, go start establishing yourself as an authority on Reddit. Start answering questions in a very deep and meaningful way and get yourself upvoted. Like Reddit is an extraordinary repository because the LLMs value what they value—information from real people that is housed in a hierarchical structure that allows them to see the meritocratic environment. That’s what Reddit is.

So Reddit to Google is more important than YouTube. And that 20 to 18% doesn’t feel staggering. It feels like Reddit’s just barely out in the lead until you think about the size of Reddit and the size of YouTube. Reddit, I don’t know what the numbers are. YouTube has to be a thousand times the size of Reddit. It has to be, probably many multiples of that. And Reddit is outpacing YouTube massively. I can share the graphic with you if you have any interest.

So those are the things that we’re coming up with. Schema markup is really important. There’s a tag inside schema that I think is sameAs. So what you want to do is you want to have hard-coded not JavaScript hard-coded schema on your website that links to every piece of content you ever create. So if you have a Medium article and a YouTube video and a guest podcast, you want your schema to say, “Hey, this Frederick Vallaeys is the same as that. Same as that, and same as that.” Schema markup important. Reddit, really important.

Multimodal content. LLMs love the fact that you’ve got a podcast that’s also a YouTube video that’s also a Pinterest post, that’s also an Instagram carousel, that’s also a Twitter thread because they get to go, “Oh, expert.” And they kind of capture some of the social signals. So, we think anecdotally speaking, I have some data to support this assumption. They’re capturing the social signals and then they’re bringing it back to you.

So, those are the things that people can do to increase the likelihood that they’re ranking in LLMs. The thing that’s crazy, though, is there’s no such thing as rank. There’s no such thing as rank. Because if you’re a 50-year-old conservative heterosexual white male and I’m a 20-year-old homosexual black female, and I ask for the best ice cream store in 85258, the zip code that I live in, you and I are going to get very, very, very, very, very different answers. We’re used to contextual search inside of Google, so we’re used to 10% variance. This isn’t a 10% variance. They’re delivering custom. So, it’s hyper hyper hyper hyper hyper specific.

So, the idea isn’t to rank. The idea is to increase the likelihood of mention or citation. And I think small, nimble advertisers are going to win this game because they’re the only ones that can create the content that will achieve the depth that LLMs want. Like stupid example, actually it’s a great example. I love books. I have 3,000 books in my home. I happen to know that the Scottsdale library system has used bookstores inside of all the libraries that sell these books that are donated and you can get them for $1, $2 and $3 depending on the color of the sticker of the book. But I also know that the Arabian library closed their store after COVID but the Mustang library store stays open and the Civic Center library is where they send all the antiques. I know that. But Borders, Barnes & Noble, Amazon doesn’t.

So the depth of knowledge on a level of granularity that a major corporation would never even be able to produce no matter how much you’re willing to pay for. That’s the type of content LLMs want. So small, nimble, deep dive content, content specificity is going to reach an apex that we’ve never seen before in the history of humanity. And it’s already happening.

Frederick Vallaeys: Yeah. It’s this whole long-tail notion again. It’s Will Reynolds gives a great keynote talk recently about as a brand you got to stand for something and that something cannot be generic. The bland marketing message is basically what you’re saying, right? It’s the Borders who’s not going deep that just have all the books, but if you’re the brand that has the specific—like antique books or antique maps or whatever it be—that’s what’s going to stand out.

And then I think that’s sometimes a scary proposition for advertisers because thinking about having a more focused key message implies that you have a smaller audience that you would show your thing to. And that I think is true in the old ways, right? So yes, with that more specific message, if somebody did a search for something generic like books on Google, well, who is Google going to show? It’s going to be the big brands. It’s going to be the Barnes & Noble. It’s going to be the Amazon. It’s going to be the Borders.

But now like we’re saying it’s that that search for books has so much more inherent meaning about who’s the persona, what have they searched for in the past, what are their tendencies? And so that could now very specifically show them that one really niche product, bookstore, library book sale, and actually have a tremendous conversion rate like you were saying the data shows from Ahrefs.

So yeah, I mean I think you connect all of these dots together and it’s a scary new world, but there’s a lot of opportunity there.

Kasim Aslam: It’s fun. Category kings are going to be usurped. I used to do lead generation for one of the largest hospice care lead generators in the world. You’ve heard of them, I promise. They spend hundreds of millions of dollars a year just on leads for the silver tsunami. People that need to go a place to be housed, let’s say. Well, here’s what’s crazy. If, and I just did this because I know their business model inside and out. When you ask ChatGPT like where should I put my mom let’s say they don’t come up. They don’t come up. It’s local small local houses with very specific local mentions and great reviews and like they don’t come up. The aggregators are going to get nailed and so if you haven’t played the SEO game because you just thought like how could I ever win now’s the time and I think that there’s a huge opportunity for paid because what paid does still to this day better than anything else is capture intent in an echo chamber.

So maybe it’s not necessarily the acquisition model for the traffic, but once the traffic is engaged with any of your properties, people should be running omnichannel remarketing, brand targeting, they should be melting the airwaves. If somebody’s engaged with you in any way, they need to see you everywhere.

Frederick Vallaeys: Yeah, that makes sense. So that’s—I would absolutely do that. So awesome name rhymes with awesome. How do you remain awesome? How as a personal brand what are you doing these days? Because you obviously made a brand for yourself through some amazing YouTube content and videos and fantastic job on those. You’re saying that personal brand matters tremendously into the future. What are you doing today?

Kasim Aslam: I’m rebuilding my entire personal brand after learning all this. Actually me and Julian who runs aeo.co for me, we talked about it. I need to be the flagship. I need to show people like this is the way that this looks. The thing that’s really frustrating, I own—I have a portfolio of 15 businesses. After I made my exit, I did the thing that all poor people do. I thought I was an investor. And what LLMs don’t like—if Frederick Vallaeys is a professional skier and an opera singer and a soccer coach and owns Optmyzr, they don’t like that. They like you to be one thing, one person.

So, I own a staffing agency out of Latin America and I own AEO.co and I own SaaS.com and I own syndications.com and I own techstack.com . So like I have all these businesses. I own Driven mastermind. I have all these businesses and all these tendrils and it from a personal branding perspective, it’s massively dilutive.

The people that I think are going to win this game are the ones that will plant their flag and say, “I’m an SEO or I’m a graphic designer or I’m”—my friend Dr. Kate Shanahan is the mother of the seed oil-free movement." It’s such a micro niche that’s so important, but she’s going to be the one that gets those mentions. So, you have to kind of get myopic in scope.

Frederick Vallaeys: Cool. Okay, Kasim, this was amazing. Thank you so much for sharing your vision of the future. And like you said again, not all the answers or not many answers are there, but really just questions.

Kasim Aslam: I just confused everybody.

Frederick Vallaeys: Well, you know, but it’s good to start thinking about these things early on. So, and again, you said people shouldn’t join the community, but it is out there. It is eventually going to have some answers or at least it’s going to have good conversations and I also think even Google doesn’t have answers today. Like everybody’s sort of figuring out this seismic shift together so I think now more than ever it is important to be at the table to be part of these conversations to be willing to share what you see and that builds your expertise, it builds your brand but it also gives you more information to make you smarter and eventually better at your job.

So with that, any final pointers where people can find you and catch up with you?

Kasim Aslam: Yeah, you can go to kasim.me—and all my links are there. You can have the book for free. Would love to keep this conversation going and for anybody out there that’s cracked the code in a meaningful way, I’d love to hear it. So help me because we’re all I think climbing this muddy hill together.

Frederick Vallaeys: Well, good. Hey everyone, thank you so much for watching this episode. If you want to get updated when there are new ones, please hit the subscribe button and also like these. Kasim, thanks again for joining us and we’ll see you for the next PPC Town Hall.