The AI Search Strategy That Actually Works

Neil and Eric break down why AI search tools like ChatGPT, Gemini, Perplexity, and Google AI Overviews heavily cite listicles, why Google is cracking down on self-published SEO listicles, and how AI-generated content is causing major traffic declines. They also discuss GEO vs SEO, token optimization opportunities, Google’s AI search evolution, and the future of AI consulting, content marketing, and search rankings in the AI era. Key Takeaways: • AI search engines favor listicle content for citations • AI-generated SEO content can trigger massive traffic drops • Token optimization may become a billion-dollar AI service Chapters: (00:00) AI Search Loves Listicles (00:45) Google Cracks Down on SEO Spam (02:33) Short-Term SEO Tactics Fail (05:11) GEO Tools and AI Content Problems (07:16) AI Content Traffic Collapse (10:19) Google vs ChatGPT Search Future (12:00) Google AI Search Updates (14:21) Token Optimization Business Opportunity (16:30) AI Consulting and Token Control (18:47) The Future of AI Pricing

• 0:00Speaker 00

  Finding customer information shouldn't mean digging through emails, call notes, and spreadsheets. HubSpot's Starter Plan brings all of those details together, so it's easier to track what's happening with every deal, and what to do next. It pulls information from calls and emails. Then it gets to work, logging and summarizing them. Right after your call ends, HubSpot pulls next steps from your call notes and drafts suggested follow-ups. And it stays up to date for you, so you can build demand, win more deals, and spend less time on repetitive work. Your records, insights, and next steps are ready for you to act on. And you don't have to miss a beat searching for any of it. Try the Starter Plan for $10 a month. Learn more at hubspot.com.

• 0:34Speaker 01

  LLMs are citing listicles more than any other content type. So, Everton AI ended up releasing a chart. And the chart breaks down the 6,000 most cited URLs per model on each platform for both March and April 2026. You wanna guess what percentage of the time Gemini shows listicles? Or cites them? 75%. 65, close. Google AI mode is going down from here.

• 1:04Speaker 01

  Gemini was the highest, 65.1.

• 1:06Speaker 00

  Okay, now we're talking about AI mode?

• 1:07Speaker 01

  Uh huh.

• 1:09Speaker 00

  45%?

• 1:10Speaker 01

  62.3.

• 1:11Speaker 00

  Okay.

• 1:11Speaker 01

  Perplexity is at 60.4%. Google AI overview is at 55.2%. ChatGPT is at 50.1%. Copilot is at 39.8%.

• 1:23Speaker 00

  Okay.

• 1:23Speaker 01

  Right? So, it just goes to show, if you want to get cited by one of these LLM platforms, the listicles are your best approach. But, Google's actually been making changes to listicles. Have you even seen they've been playing whack-a-mole to listicles?

• 1:40Speaker 00

  No.

• 1:41Speaker 01

  Yeah, so, they're specifically going after self-published listicles, which is very different. You and I have talked about this, like, dude, people are going out there. We interviewed a marketer from a competing agency who ran their marketing, and I won't mention the marketing agency. They're like, "Oh, we've crushed it for them. We've created, uh, I think it was almost 200 listicle articles on their own website." So they're ranking for a lot of agency terms. And we didn't end up hiring the guy because although it was impressive on the results, we didn't like the strategy and the tactic. We believed it was like a

• 2:15Speaker 00

  Short term.

• 2:16Speaker 01

  Short term, one trick pony in which if your whole playbook is I'm just gonna go create hundreds and hundreds of listicles on your own website. Well, of course, your own business is gonna list your own company, your own products, your own services as number one compared to all your other competitors. You agree you would be biased. Any business owner would be biased. Even if you felt you weren't the best, I don't know any business owner who wouldn't put themselves as number one.

• 2:41Speaker 00

  I mean, look at Amazon. They put their first product number one. Why would you not do that? You have the distribution.

• 2:46Speaker 01

  Exactly. And what it's causing is Google is just be like, whoa, whoa, whoa, this is just manipulated results. And they're just playing whack-a-mole at this point, but we're not seeing ChatGPT going after them as hard or anywhere near. It's more so Google is going after, and the key here is self-publish listicle articles.

• 3:07Speaker 00

  You know what's interesting here? So the fact that Google is going after listicles, the self-published listicles, it's no different than how they played the game before. They, they already have the search spam team, so they know exactly what to go after, what's being manipulated. And so when you and I saw this a couple of years ago when it was happening, it's the same game because you know why? We've been whacked before. We played this a very long time ago. This is 15 plus years ago, 20, 15, 20 years ago. We had played this game and it was harder back then for Google's search spam team to catch this, but they have to preserve their product. So they have that laid out. The thing is anthropic and OpenAI eventually will catch up too. And I think they'll have their own versions of search spam teams, but this has been no different, which is why we've been saying this at the beginning.

• 3:48Speaker 01

  Yeah, these are terrible strategies. In marketing, I'm sorry, I'm meeting some yogis that Eric got me, which I sadly got him addicted to.

• 3:56Speaker 00

  No, yogis are great.

• 3:59Speaker 01

  It's just yogurt with fruit in the inside.

• 4:01Speaker 00

  You get on Amazon.

• 4:02Speaker 01

  Affiliated, but they're great. So any marketing tactic that produces high short term ROI. That's really easy to do, and that's a caveat, that's really easy to do. Usually nine out of 10 times won't last long. And as a marketer or a business owner, you should avoid those tactics. And you can look at them one by one, there may be some edge cases in which you should do them. But in almost all cases, you should avoid those tactics. The reason being is. You're spending energy, time, money on these short term tactics that will go away. And then what you'll find is two, three years later. All the effort and the money and the time you spent from previous tactics won't help you.

• 4:44Speaker 01

  Versus if you took the long term approach, spent the time and money. On strategies that don't produce a media ROI and are not that easy. But then three years from now, if you keep doing those kinds of strategies. The results start compounding and your ROI is much higher. So we've all, well, I've always said this, that you want to do black hat things in a white hat way. So the way you would do this black hat thing in a white hat way is you would collaborate with other websites. Okay. And then maybe they would list you on their listicle. But the whole idea is you don't want to do the self-serving make because it, by the way, think about it this way. If you're an algorithm and you're like, whoa, people are really gaming this thing where it's like listicles are. If you're self-serving and everyone's doing it right now, it's pretty easy to identify if you're an algorithm.

• 5:27Speaker 01

  So you're better off thinking about collaborating with other people, building long-term relationships and then doing that type of strategy. Maybe it's less volume, but you do it with these, these, you know, high authority collaborations. And the thing is, it's not even the link that you're going to get from it. And the listicle at the end of the day, it's a relationship that you're going to get. You're going to be able to get a lot more from that partnership. Yes. Yep. So by the way, on, on this note, because we're talking about G E O S E O right now. You know what's interesting to me? I spoke, I spoke to a couple of people recently. One person, these are both content creation tools. Okay. One dropped 80% in revenue. Okay. 80%. I'm not going to name names right now, but just. AI content creation tool? 80%. Okay. And I spoke to someone today, so some, a, a, a G E O tool was introduced to me today. Okay.

• 6:11Speaker 01

  And let me talk about something else that pisses me off. Last time I started the episode, someone pissed me off. Today, someone pissed me off again. You know what?

• 6:17Speaker 00

  I can only tell you this afterwards, after we do. So, but, but here's what happened. So, you know, one of my friends introduced me to this G E O. company, okay? And basically, I spoke with them and they're like, I'm like, we start to call. It's like, oh yeah, you know, we don't want to ask you for anything yet. I'm like, okay, well, tell me about your product. So I don't really get much about their product, okay? And then they're like, oh, uh, they tell, they don't tell me much about the product. I'm like, okay, so how's your traction right now? You guys are looking to raise another round because they're looking to partner with me, right? They wouldn't tell me their traction. They wouldn't tell me about the product. I'm like, so what do you guys want, right? And then they're like, well, you know, we don't want to ask for anything. We want to ask what you want first. I'm like, dude, you're like pissing me off. Like you're like, I get on this call because my friend asked me to get on this call.

• 7:00Speaker 00

  And then, and then it's like, you don't tell me what you guys really do. You don't give me enough info. And then like, you don't tell me about your traction. So how the F. am I supposed to help you, right? And then, and then you do this whole thing, like this old weird positioning thing where you're like, well, I want to help you first. It's, I really want to cuss right now. I'm just like, and I was, I was like, guys, you're like, you're really annoying me right now. And I said that on the call. And I'm just like, I'm, I'm giving you feedback right now. If you're going to take these calls in the future, like you need to be specific about your ask. Cause I have no idea how to help you right now. And, but it's one of those content creation tools again. So then, and, but the thing is, what my, my point of saying this is that these content creation tools seem to be going on a downward trend right now. And it's interesting because one person is, is like, they're done with that company. The, the like, um, they're like, they don't want to work on it as much.

• 7:44Speaker 00

  And then the other, the other ones, like they're just starting out. So my point of saying this is, by the way, if, if that company is still listening right now, the, the one that got on the call with me. I don't, I don't know if this is the space to get into, so.

• 7:55Speaker 01

  Dude, did you see Lily Ray's study on A I. content creation? And the websites that are using these tools and what's happened to their traffic?

• 8:02Speaker 00

  No, what happened? And she did an interview with Ross Hudgens, didn't she?

• 8:05Speaker 01

  I don't know who the interview was.

• 8:06Speaker 00

  Ross Hudgens was the first employee of Single Rain. I wasn't here at the time, but he was the first employee of Single Rain.

• 8:10Speaker 01

  I didn't know that. So, uh, pretty much the title of it is like, A I. content creates a rapid growth before a steep decline. So there's 220 websites that she was analyzing using tools to see what happens when you create tons of content using A I. What ended up happening after a while, right? Keep in mind, content's been published, so they saw traffic increases from organic perspective. But when you look at the data. You know, from their peak, 54% lost 30% or more of their peak organic traffic. 39% lost 50% or more, 22% lost 75% or more. Within those declines, a reoccurring trajectory appears.

• 8:54Speaker 01

  A rapid growth in organic pages over six to 12 months. Organic traffic peak within roughly three to six months of the content peak. And then a steep decline in traffic that erases most of the gain and frequently drops below the prior baseline within the following year. So what's happening is, is companies are using AI to create a lot of content at volume. It's causing traffic growth. But what they're missing out is, from an SEO perspective, you're not just evaluated on a page by page basis. You're evaluated on your whole site, right? Think of it as an entity. And if you have a lot of junk on your website, it hurts the good pages that were written manually by humans.

• 9:36Speaker 01

  And overall, we're seeing, or what she's seeing, and we're seeing this at our agency as well. Traffic starts dropping more and more and more. And you have this big, messy website with tons of pages. And you have to do a massive cleanup and it's very hard to recover the traffic.

• 9:53Speaker 00

  It's interesting to me. I mean, I would say this, if you're going to produce something naturally, there's like a half-life where you start to, that traffic is going to decline over time. But I'm looking at the graphs that she has on this, this page over here. And we should be able to pull up, pull this up for the, um, the clip. But, um, you know, I don't need to show the camera, but maybe I will show the camera. But you, you see these, like that, these are pretty steep drop-offs. It basically goes nothing in, in the, the orange line. And you also see the number of organic pages disappearing. At least this is from like an Ahrefs index. So, um, all that to say is, you know, it's been the same since the beginning. Like Neil and I have been doing SEO for a while. You publish poo poo, you, you end up with poo poo traffic.

• 10:28Speaker 01

  Definitely.

• 10:29Speaker 00

  Yep. Just a quick break. If you want to run LinkedIn personalized ads and landing pages, you can't do it on LinkedIn ads right now. But you can do it with Karr. It's www.karr.k-a-r-r-o-t.ai. Again, K-A-R-R-O-T dot A-I. And we'll see you on the other side. So you have a prediction here. You get three in a row here, Neil. So you have a prediction here.

• 10:48Speaker 01

  This would be four in a row.

• 10:50Speaker 00

  Is it four in a row?

• 10:50Speaker 01

  Yeah.

• 10:51Speaker 00

  Okay, four in a row.

• 10:51Speaker 01

  I messed the numbers up.

• 10:52Speaker 00

  Neil has a streak. Neil, just for months and months and months, added no topics at all. And then today, he had a random ballot of inspiration.

• 10:59Speaker 01

  Yeah. So my prediction. Is Google will look more like ChatGPT and ChatGPT will look more like Google in the future.

• 11:07Speaker 00

  Frictionless Growth Marketing hosted by Sonya Thompson is brought to you by the HubSpot Podcast Network, the audio destination for business professionals. Frictionless Growth Marketing hosted by Sonya Thompson shows how modern brands grow by removing the friction in their marketing and customer experience that pushes customers away every single day, often without leaders even realizing it. Each episode delivers practical insights, real brand examples, and conversations with marketing leaders, showing brands how to win with the new rules of growth, identity-driven marketing, cultural relevance, and frictionless customer experiences. Listen to Frictionless Growth Marketing wherever you get your podcasts. What do you mean by that?

• 11:43Speaker 01

  So, when you think about Google, people associate Google with traditional search. We know from a, I'm not talking about it from an engineering perspective, but if you're looking at just LLMs and usage from people using it purely for search, ChatGPT is the winner by far. It has the majority of of of usage. I'm assuming you agree with that statement if you're excluding AI overviews. The second winner is Google. They're in the number two place with AI overviews, Gemini, etc.

• 12:10Speaker 00

  AI mode, yeah.

• 12:11Speaker 01

  AI mode, etc. But most people associate Google search with just traditional search, in which, hey, you're gonna get some blue links and some videos and maybe an AI overview. You can go to AI mode to see more, or you can go to Gemini. You can do Lens on your phone, but not that many people use it. There are roughly 20 billion searches a month right now for Google Lens, which I know seems like a lot, but to put it in perspective, Google's getting around 13.7 billion searches a day. Just, you know, people just typing in some queries. So, the 20 is not bad, but it's not a big chunk percentage-wise. And that, again, is monthly. And Google just made some new announcements.

• 12:56Speaker 01

  They changed their search box and made it an intelligent search box. It's the biggest change to search, to the search box in 25 years. So I don't know if you saw what they just released today.

• 13:07Speaker 00

  What? No, I didn't see the product releases today.

• 13:09Speaker 01

  So, it's actually really cool, and if you see my screen, I'll play the video, because I know Noah can see the camera there.

• 13:15Speaker 00

  Yeah, so just for the editors, by the way, when we're showing this stuff, we have a link column for this. We should show these in the videos.

• 13:21Speaker 01

  Yeah, so in this example, for anyone who's just listening, In Google's search box, imagine a plus before you type in the queries. So in the search box, it's pretty much elongated oval. And on the right side, you have the search button. Now imagine now on the left side, you have a plus button.

• 13:41Speaker 01

  So you can click the plus and you can click on anything. AI mode, create images, or gallery. Where is it? Let me see the options actually again. Yeah, AI mode, create images, deep search, canvas, gallery, camera, files. So for example, you can click on gallery and you can select two or three items that you like. Let's say it's a dress and that's what Google's using in this example. So the searcher could select two dresses. And then it highlights those two images and then you add in text. You know, the example here is help me find a similar dress, but in this color scheme instead. So they showed one dress image, one color scheme image, click search. Google showed all the similar dresses, but in that color.

• 14:26Speaker 01

  And that's the example of what Google's doing with search and AI. And that's just skimming the surface. They're getting really end up, even with their new search feature. You can actually build an app just typing in a prompt into their search feature. Yeah, it's called seamless Google search to AI mode. Uh huh. Yeah. So this is something that launched to many back in January, but it's now fully live. Okay, cool. So yeah, I mean, all that to say is this, this kind of reinforces, this is not investment advice. By the way, my dad was asking if you had to pick one stock from the magnificent seven, what would you pick right now? I said, Google, because they're, they're on a tear and you know, they're, they're clearly profitable and they've clearly been continuing to innovate. So by the way, this, so you have this other piece over here.

• 15:09Speaker 01

  Google lets you build your own app within Google search using a Gentic coding. Is there something else you wanted to go deeper on there? No, I thought it's, it, it just shows and it goes more with the narrative that Google is really going after OpenAI and Claude and they're integrating a lot of these features. And I was talking to my brother-in-law the other day. We were talking about businesses. I was like, you know, the business that I believe is massive right now, that not too many people. I haven't really seen anyone build it yet is a consulting company that just focuses not on AI integration because there's already too many of those that just focuses on token optimization and fee reduction. And then you just take a percentage of the savings that you save. So think

• 15:52Speaker 01

  of like AWS and cloud, right? During back in the day with AWS and a lot of these cloud computing companies, there were companies that said, we would help you optimize your cloud architecture. So instead of spending a hundred thousand a month, you now spend fifty thousand dollars a month, and they would take a percentage of the savings. I think there should be a similar business for tokens in which someone helps you optimize your token expenditure because there's a lot of companies that are spending millions and millions, some even tens of millions. And I bet you there's many that are spending hundreds of millions because the supposed rumor is Anthropic's now around forty five billion in ARR. So if someone just helped you with that reduction and their fee was a

• 16:34Speaker 01

  percentage of the savings, I think that's a win-win model and that business would blow up even though it's boring and ugly.

• 16:41Speaker 00

  Here's what I'll say about that. So two things tied to Chamath. I have two things on Chamath today. This is a Chamath episode. So I spoke to a guy who basically is a leader at a $2 billion a year company revenue, okay, $2 billion revenue. So pretty sizable, right? And he actually worked with Chamath's 8090. So initially Chamath's 8090 company was a company that said, hey, we're gonna help eliminate all of your costs on like Salesforce and all these things right now. And bear in mind, this company is spending like twenty five million dollars a year on software. So they're like, okay, maybe we can chop five to seven million dollars a year. That's a good amount, right? And so they work with Chamath's company for a year. And apparently, it didn't it didn't work out because it was too complex to make those. And I was like, okay, so how would you have paid them?

• 17:25Speaker 00

  Well, they were gonna take a take a percentage of the savings, right? And this isn't token savings exactly, but this is cutting on software costs. And that didn't work out initially. So here's what I will say, though, because he has found a model that does work for him. That model didn't work initially. So maybe this token optimization model might work. I think it's still more of like a could be an add-on service. I don't know if it's gonna be kind of the primary service. Could work, right? But what I really wanted to get into here was actually on Chamath saying, if you're running a consulting business and you're deploying Anthropic or OpenAI directly into your organization, I'm looking at you, PWC and

• 17:59Speaker 01

  Accenture, you are letting the fox into the hen house. OpenAI and Anthropic are openly funding and starting competitors to you while also using your usage to drive more success for them. This is not a failure on their part, but a failure on your part. So consulting businesses that understand this are adopting a control plane that allows them to arbitrate where tokens go and who generates tokens for them. So controlling the tokens, Neil, is controlling the spice, like Dune, right? So this was a key pillar of AD90's global partnership with EI. And we're close to another global partnership, but we're taking control of token generation and can direct them to any model provider, which is kind of what you're talking about, actually. So, yeah, a little bit different. I saw his tweet out there. Yeah.

• 18:38Speaker 01

  I think he's more so going after like the Accentures and the McKinseys and, you know, pinpointing how they're helping build OpenAI and Claude with all their customers and all the data and all the spend. I do agree that, again, like, I don't think that's all that AD90 does, but like, to my point, what I just made earlier, I think token optimization and token routing is going to be a core service that companies add. And I actually think you're going to have it. I think I'm going to have it too as a service. I don't know if we're going to have it as a service. I genuinely believe that the larger players will eventually, I'm not saying in the short run, eventually will help you fine tune and reduce the cost for you and do it in the most economical way.

• 19:20Speaker 01

  If you look at AWS, you know, you take your cloud bill, they try to do reserved instances, they have consultants, they try to help you save on the bill. And we've even been pitched by Google, hey, switch to Google Cloud, we'll help you save X, Y, and Z on infrastructure and all of that. And that's why I believe that the LLMs themselves or the AI companies will try to save you as much money as possible in the long run. Because if they don't, a competitor will come by, like a Google, and say, hey, we can do similar things and we can do it for half the token price. And here's how. And then you're going to start seeing a lot of companies switch. I have two thoughts, two more thoughts on tokens. I was talking to a founder today.

• 20:01Speaker 00

  And we're just, it was like 30 minute talk just on AI, right? And we're just like, dude, OpenAI and Claude are so smart. Because you know what? You pay the $200 a month. And it's like, you feel like it's all you can eat buffet. And you're just going to town every day, right? But inevitably, what's gonna happen in the next year or two is they're gonna have you so hooked on it. It's like, you're gonna have to pay more. And I was like, okay, fine, let's pay more, right? Because it's so damn good. And we will end the episode on that. Goodbye.