SEO TV

AI Overviews SEO: Can You Get Your Traffic Back?

AI Overviews SEO is the subject of this episode, and David G Quaid says old posts can win new traffic when their titles and headings match current search phrases. He also says LLMs will not find a brand on their own, so a company has to publish pages that describe its market and where it fits.

Running time
53 min
Released
Aug 21, 2026
Starring
David G Quaid
Channel
David Quaid SEOtv
Watch
Open on YouTube
Rated PG
PageRank Guidance suggested

Three Takeaways From AI Overviews SEO: Can You Get Your Traffic Back?

Three takeaways from AI Overviews SEO: Can You Get Your Traffic Back?, in David's terms.

AI Overview Traffic Loss

AI Overview traffic loss varies widely, with David G Quaid saying he has read figures from 28 to 38% and heard of sites losing 90% of their clicks. He says short definition and FAQ queries took the hit, along with publishers and affiliate sites. He also cites a Shopify research post saying LLMs added traffic on top of stable Google traffic. Not every drop is the overview, he adds: his own content arguing that no tool can check EAT lost clicks after a core update, which he calls a content mismatch.

How to Do SEO for AI Overview?

To do SEO for AI Overview losses, David G Quaid says to start with tight data: pick one exact query with a stable baseline, not average position, and see whether the drop lines up with a Google core update. He warns that a slip of one to three positions can push a page out of the click zone, and extra ads above it can make that worse. He then suggests dropping a domain into Semrush, turning on SERP features to see which keywords carry an AI Overview, and publishing for keywords that show none, starting with long-tail ones.

AI Overviews Traffic

On AI Overviews traffic, David G Quaid says old content can earn new traffic when its titles and headings are tied back to current search phrases. He describes sites with a thousand blog posts where as many as 600 get no clicks, and advises checking whether the search phrase has changed, following the new one and republishing, since he says "dates are surface data." He says his team added new headings based on how search changed over the last three months versus two years ago and saw up to a quarter of a million impressions a week on some posts.

How to Show Up in AI Overviews

To show up in AI Overviews, David G Quaid says a company should write pages that define its marketplace from its own point of view, as sites like Capterra and G2 do, and say where it fits. He says query fanouts let people ask about competitors and market analysis in one phrase, so companies covering both appear more often in AI Overviews and LLM content. He also says sites that align internal headings with search phrases pick up more fanouts because there is less competition. He calls the belief that "the LLMs will come and find you" nonsense.

How to Rank on AI Overviews

To rank on AI Overviews and other AI answers, David G Quaid says to find the query fanouts, which he describes as blog posts and articles still to be written. He recommends Perplexity and Claude because they show the search phrase when they search the web, while Gemini and ChatGPT need a look in the browser. Bing calls them grounding queries, and he says they also appear in Search Console and Bing Webmaster Tools. He says existing tracking tools only report the percentage makeup of prompts, so he tests his own prompts and their drift instead.

Transcript: AI Overviews SEO: Can You Get Your Traffic Back?

This transcript of AI Overviews SEO: Can You Get Your Traffic Back? comes from the video's captions, lightly cleaned. Click a time and the set jumps to that moment.

Introduction and sponsors

Welcome back to the David G Quaid SEO show. In this is episode two. In this show, we're going to be discussing Google overviews and how it's affecting SEO in 2026. Got my great partner in crime, David G Quaid, with me, the goat of SEO. It should be a great show. Like I say, we're going to be discussing all things Google overviews. So, David, would you like to quickly introduce yourself to the new listeners and viewers?
Yeah. Hi, I'm David Quaid. I'm the former SEO king or former king of SEO after Kasra temporarily displaced me yesterday. But Gemini is reporting a site change in the royal lineup hierarchy since last night. But yeah, I love SEO and I love chatting about it on X.
Yeah, I I see you as the king coat actually, David. So there we go. Should be a great show. Like I say, a lot of people are really a bit concerned about Google overviews, how it's affecting SEO. David, I think has some interesting views on this. But before we go into the main part of the show, I we need to take a break for the major sponsors of the show. We will be back in a few moments, folks.

WP Tonic is sponsored by Jetpack. Managing WordPress sites shouldn't mean stitching together a pile of disconnected plugins. Jetpack brings security, performance, publishing, growth, and AI tools together in one WordPress native toolkit. Use what you need. Backup, scan, boost, social, video press, stats, and more. All with modernized dashboards built into WP Admin for your site or your client's sites. See what Jetpack can do. Jetpack.com

Impact on search traffic

We're coming back folks. Like I say, should be a great discussion. So, David, you know, how much organic clickth through traffic you think publishers have actually lost since Google AI overviews rode out in a broad sense?
I'm sure it's quite a lot. I've read and seen sort of anything from 28 to 38% I I imagine it's quite a lot of traffic has gone to AI overviews especially for like I guess shorter definition things you know like what is WordPress certainly I think even the media have reported that they've lost quite a lot of traffic I guess people looking at headlines things like that they're able to get a quick spread rather than reading 10 20 different pages. So I would imagine at a board level at a high level it's quite a lot.
Well, you know, the people I follow they're saying that you know from 40 50 60% of traffic reductions. Do you think that's probably correct? Yeah, I think the worst I've heard is like 90% where some people are saying that they've lost 90% of their clicks due to AI overviews since AI overviews have rolled out and gotten worse.

So, what do you think are the type of websites and content that have been hit the most? David,
I think a lot of what is questions, FAQ questions, you know, foundational layer stuff where the answer is pretty short. I think a lot of people I think a lot of AIO's may also lead into deeper questions. And that's very hard to measure because we're not getting clicks. And so the what's what's inescapable is the attribution and the customer journey is getting more and more fab sort of fragmented. And I think that's really difficult for people to tie together. And so even if you are losing AI overviews, you're not going to be able to see how many impressions you're getting and you're not going to be see able to tie that to sort of potential leads that come in. I think some of the sort of like silver lining news that a lot of SEOs talked about at the start would be that you might lose a thousand clicks but you might gain 100% in clients particularly in B2B or e-commerce because people are doing more of their research in AI overviews and then coming to you for the final step.

Obviously for publishers that doesn't help especially if they're on affiliate marketing or AdSense you know advertising revenue that's going to be difficult because they're not getting any impressions for those ads and so a lot of that is moving up to Google who essentially I guess are it's they're sort of crossing that fine line between directory and publisher right they're now publishing news and that's going to be interesting and the same with ChatGPT What was interesting is Shopify put out an some a research post a week ago saying that LLMs haven't reduced Google traffic, they've just increased the amount of traffic. So, they're able to look at all of their shops and they're able to look at all of the search traffic and say Google has remained stable, but AI is sending a different sort of strand of traffic over above that. So, it's expanded search, which is incredibly interesting. And that's an incredibly interesting comment from Shopify because they don't really have a dog in the race, right?
Well, in a way, but yeah, I I wouldn't say that they would publish something that was deceitful. No, I think the data is probably shown that. Obviously I think they do they do they're one of the leading e-commerce platforms so they want people to keep searching for products don't they but I suppose in a way their mo is one of the more defendable because if people I suppose fundamentally it just depends it gets to the stage where people can just buy stuff through like ChatGPT or but if it got to that stage like they were acting like Amazon I think that would be so difficult politically that who knows I'm just speculating there. So

SEO strategy and recovery

Let's move on. So what do you so you know if somebody's facing it this and they've they've seen a decline in traffic is there anything they can do about it really David
I think so we've definitely seen a lot of B2B sites losing top of funnel or adjacent traffic you got to remember as well when you look at Search Console, if you're building out your website, there's a massive difference between impressions and clicks, right? So, even a site like yours with where you're helping nonprofits and community organizations, schools, universities, colleges do things like build community newsletters, build websites, get find WordPress hosting. A lot of those impressions are happening for queries that you haven't even thought about, right? That Google's matching you with or a lot of their indexes are full of sites that are adjacent to a topic but not talking about a topic. And so losing that traffic hasn't hurt B2B sites as much, right? So I think that for both publishers, e-commerce, B2C, B2B, one is you can test to see did your pages fall and lose ranking, right?

Because a number of of Google updates have happened. Number of people have shared their Google search console with me on screen and I can see what I call a write down. And I don't know if that's the right word, but it to me it just looks like a write down where if you start looking at specific pages and specific queries, right? Where you've got a very strong stable baseline, right? Don't look at a at don't look at aggregated information for average position. It doesn't help. But if you pick an exact query like for example for your site, I would pick, you know, WP hosting as an example, right? Rather than all of the other WP associated terms because it might be include plugins, it might include things that aren't relevant. If you see a particular point that lines up with a Google core update, not a spam update, where the traffic is going up and then it just goes down and it just starts to go down. You can see a specific point. For example, the last date I saw was something like July the 14th, for example, where it looks like during the core update, Google looked at a bunch of pages on a site and it's like it was looking at new or sort of looking at user engagement from a different light. So, so basically taking the numbers or looking at new numbers or altering the weight of those numbers in the algorithm and then writing people down two to three positions. And I've seen a lot of people say look our positions are stable. And I want to just sort of like tackle that point of logic first. Stable isn't it could mean a one to three position drop and you say like that's stable but that could push you out of the click zone. Right? If you were if you were averaging 3.9 and now you're averaging 4.9, you could lose 90% of your clicks.

That could be enough in that particular search. It depends how many other things are showing up, right? Or are there YouTube videos? Are there more ads on that page? So, if there's more ads on the page, you're really, let's just say there's four instead of two ads, you're really in position six to seven. And you might, that might even push you to page two, in which case you'll see a 98% drop in clicks potentially. So the first thing is to get really tight on the data and see if that drop is enough to push you out of the click zone. The second thing is to take those keywords and using a tool like Semrush or Moz look at the SERP features report in the public database. So just drop your domain name into the domain name explorer in Semrush and look at how many queries have an AI overview in them. So you basically the graph for Semrush will show all of the positions like I think the top three are in yellow then it shows the top 4 to 10 to 20 and those clusters if you turn all of those off and just turn on SERP features you can see just the keywords with AI overviews and I think it's fairly reliable and what you could do then is start to look at other sites and competitor sites and do more keyword research and find keywords that aren't showing an AI overview and start publishing more content and move away from those queries. Now this I don't want to make it sound like it's easier said than done. I'm pretty sure a lot of people are trying to combat it and I'm not saying look you've got to do your homework, right? I'm just for the people that are just sitting looking at their data and they don't know what where to start. I would start there.

I would start looking at longtail keywords or I'd start looking at keywords that are were different. Also, something I see a lot is that people have like a thousand blog posts and as many as 600 aren't getting any clicks. Go back and refresh that content. Go back and adjust that. Go back and see if those keywords are still ranking or if the search phrases has transformed, right? And if it's transformed, then follow the new search phrase. Republish it as a new post. I'm not I'm I'm never worried about dates. I think dates are surface data. Well, I sit outside of news, right?
Yeah.
And so if you can take the slug, change the slug, edit the pages, look at what people are searching for, look at your titles in your page, are they matching? Right? So a lot of people sometimes write page titles and H1s from you know very clickbait type not clickbait but I guess to encourage click you know sensational headlines or something like how open AI is disrupting the market and that may not match a search phrase right people might be searching for something very specific so that title might work well in X and it might work well in LinkedIn and it might work well if you're getting a lot of traffic but as your traffic is slowing as your authority falls, your semantic broadness falls with it. And so if you start tying those back to current phrases, you can take old content and start getting new traffic for it. And that's my advice. And we've done things where, for example, on some pages, we've added new headings based on a change in search in the last 3 months versus two years ago. And we've seen up to a quarter of a million impressions a week go up on some blog posts, for example. And so you can definitely take that descending line and either add content or edit content you already have, change your internal linking, add more site maps, edit page titles, just start doing SEO across your site and then just start looking at the macro output.

Content and search intent

Well, you've thrown a lot of information there, David. So, [snorts] but don't you think I'm not going I'm trying not to be negative here, but it's in some ways it's good news because what you're saying is by changing the slugs, by changing the title, by changing some of the content, and don't be fixated by the date, you can you can get a retrieval to some extent. But on the other hand, isn't Google just pushing just gradually pushing out more AI overviews or is there types of content that are a bit more resistant to what Google is doing?
I think there's definitely you know, like I said, for my own blog, I haven't lost any traffic. Wow. Now I'm not but I didn't have massive traffic and I'm not targeting adsense. So, you know, rel relative to
What some publishers are seeing, the traffic I've got in the last two years wouldn't make up a single day in their lives, right?
But traffic for SEO hasn't gone down and it's not going to the AI overviews.

And
Well, I'm surprised you so but I think yes, I think I think that it's affected publish and affiliate focused websites the hardest. Those that were writing blog content for affiliate income. I think a friend a friend of ours Edwards Edward Sturm that he's had a few pe a couple guests on his podcast that saw got really hit hard but they've managed to turn it around. But I'm talking about people that have actual services, have actual products. They've seen a massive decline as well. You know, I speak to a lot of entrepreneurs
In the WordPress space and they all report a lot of decline. Now, I'm not talking about WP beginner because I kind of classify that as a publisher and [clears throat] I think that's a very individual scenario.

But I'm talking about actual people that got actual products, actual services. First of all, would you agree with that? And what can that are you saying that they just do what you've outlined a couple minutes ago?
Absolutely. So I'll give you an example of some of the content that got written down on my site. And a lot of people will draw similarities between this and matching search intent. So matching search intent is a very common proxy for understanding logo sticking. So I started trying to rank for each topics, right? So I could put my voice into content. And I was talking to James Dooley about this both on the Edward Show and off. And we were talking about how you can change a single word that changes the meaning of a title. You can you can rank for the opposite of something, right? So my EAT content I couldn't believe stayed ranking for the first half of this year because let's say I'm ranking for EAT checker. My content is essentially saying there is no such thing as a tool to check EAT although there are there are lots of tools that each that check your
Can you explain what you mean by EAT?

Yeah. So the EAT the experience expertise authority trust score that Google had quality raters judge sites on. And a lot of people started reading into that far too literally, thinking that the more claims you make in content equals a higher eat score. And obviously, a lot of web engineers, web developers have come in and built tools that try to give you a score based on how much EAT your page has, which I think is a ridiculous concept.
Yeah. However, you know, if people want something, the market fills in and gives people what they want, unless it becomes illegal, which I don't see happening in this area because it's just not something that lawmakers care about, right? No one's going to get injured. But what happened after the last core last Google update, the core update, is that clearly people reading my content weren't enamored with it, right? Because I'm effectively telling the o telling them the opposite of what they wanted to hear. And while that content was going up since the start of the year because I'm just really effective at understanding the levers in Google putting data in the algorithm that will force my content to go up. That content is actually starting to rank down and it's getting less impressions and less clicks. And that's why I assume Google is looking at a different set of metrics to judge whether the user was happy with the content or not. And that's actually what I expected to happen with that content because I don't see people looking for the antithesis of what I'm saying. Right? So if people start searching for like why is EAT right? I will rank for that and I imagine that the click-through rate and the stickiness or whatever you want to call it would be much higher for that cadre of visitors for that particular search phrase. And so that's why I'm saying that's not an AI overview issue.

That's a my content didn't match. And if I take content that does answer it, that content has continued to go up even in the presence of AI overviews.

Visibility in AI overviews

All right, I'm following you there to some extent because but on the but I don't I want to go on to the next point then. So are they any particular kind of websites that are gaining visib visibility inside AI overviews and is there anything commonality between these websites? I think sites that are very clever at again aligning their internal headings with different search phrases are picking up a lot more of the query fanouts because there's less competition for query fanouts. And if they're constantly testing longtail prompts and finding out what the sub queries are, they could definitely be expanding. We don't we've stopped running AI overview sorry AI reports for most of our clients because we don't necessarily know what the clients are what our end users are searching for but we have seen in some areas a dramatic uptick in the amount of leads coming in and tying that back we've seen that a lot of that information comes from where people are trying to draw something over two different topics. So one of the funny things about how Google has worked and Google's worked for the last 26 years and not a lot of people have questioned is that it's very difficult to get a page to rank for two topics right so people will rank a hotel say a page about hotels in Italy or holidays in Italy or flights to Italy and it's very difficult to rank for all of three of those things in one go but with query fanouts people are able to ask sort of two questions in one right so they're they're maybe able to ask about competitors and market analysis in one phrase. And so companies that expand beyond just either listing their competitors or avoiding their competitors and listing in both say market analysis as well as their own messaging are appearing more and more in AI overviews and in LLM content if that makes sense. In other words, they're massively increasing their footprint.

Well, I didn't follow that. Is any
Let me let me give you an example, right? And let's make it hyper local to WP Tonic. So, traditionally, you followed a very similar content strategy to a lot of other companies in your space, right? So you talk about the services you provide, the types of packages you provide. You talk about some of the plugins that you have optimized for your hosting environment. You talk about common issues that people provide, people run into. Now, one of the things that you that a lot of people don't do is they talk about where they fit in the market. So for example, I'm in the community hosting in market. I'm in the WordPress expert hosting market. And so providing information about how you plug into a market is something that a lot of people don't think about, right? They just think about talking about the services they're doing and matching what people search for. And they put their messaging and market fit in their about us page. But if you were to start writing pages about not just you and your competitors but essentially taking on like a site like Capterra G2 and defining the marketplace from your point of view and how you fit into that you could then expand the amount of visibility you have outside of just hosting but hosting marketplace. So, when people say, for example, who's a competitor with GoDaddy for expert WordPress hosting, GoDaddy's going to appear and maybe WP Engine's going to appear. How do you get yourself to appear, right?

It's a it's a substantial mindset change. It means writing about something that you don't do. It means connecting what people are searching for to what you do, right? Yeah, I think I'm I'm understanding, but obviously I'm not going to WP Tonic isn't going to complete in any shape or form with WP Engine, you know.
No, I understand. But if someone is looking, people have to start with their own knowledge base, right? With their own vocabulary. And so if they start with I want to do WordPress hosting for my college, they might start with WordPress hosting for their college and they're just going to get generic hosting information.
Yeah.
If they get generic hosting information, that probably will include WP Engine,
But they might notice that WP Engine isn't is a specialist for WordPress, but not a specialist for community hosting. Yeah.
And so if they dig deep, if they're digging deeper and they might say, "Well, who competes with WP Engine for community?" If you've if you're ranking for community hosting, but you have nothing WP Engine, there's no way for an AI to include you in that conversation,
Right?

And so part of this goes back to a question. It's one of one of Ed's favorite topics talking about why aren't people appearing in AI? Why aren't why is appearing in a search not necessarily why do you appear in the synthesized output? Something that James Dooley touched on and something I think James Dooley will be coming back with myself and Edward to talk about as well and something that you also talked about like what is the difference between SEO and geo and obviously in the debates that we had before we w I wasn't going to hand this to the guys on the plate because it's something they'd never thought of right and it's something they should have known coming on to the show but essentially you now have to work out how to frame yourself in the context of a comparison between you and WP Engine. It doesn't mean you're competing with WP Engine. What it means is you have to tell the user you're not competing with WP Engine. But if they're looking at WP Engine because they're a nonprofit or a school or a community or a student or a teacher, you provide specialized hosting because you have these services, you have people with the skills, you have the plugins, and you have the apps that help them do things like run a community all built into one hosting package that neither GoDaddy nor WP Engine nor anyone else has. But you have to go and write it. I think the mistake that a lot of the geo tools and geo agencies are making is they're assuming that the LLMs are going from sight to sight figuring it out. They're not.

You actually have to sort of imagineer and invent and in and start somewhere by laying out those bricks.
And I would imagine you would be one of the first people to rank for that. And that not knowing. I think it's kind of scary because you can't see keyword volume and you're never going to know how people are going to find it. But by putting that information out there, you're increasing the chances of you coming up in an expanded query fan out and then being cited by an LLM and feeding it the information that it needs to complete an answer for a user. Just to finish off this. But I thought you explained that really well actually that became much more clearer now. David, thank you.
Thanks.
That's great.
But, before we go for our mid break, David, there's a competitor. They're based in Australia and they have office in the US and their website, they've got a lot of schools. They got a lot of colleges. They've been around a long time and I think they offer their hosting to WordPress professionals, you know, a college that's got internal developer team and they offer but I'm I'm only summising that. But I noticed with their website they've got no blog section at all. They've got no content section.

They've just got some key pages. And I just presume they went to WordPress or educational conferences and they might be doing outreach. But when you see a website that hasn't got a lot of blog content and they're a competitor, should you should that tell you that you shouldn't write content or should that be a signal that you should write content or is it more finite? Is there more variables there?
So I think it's always easier to have leads come to me. I think the hardest thing for me is to cold call up a company and say, "Hey, do you need SEO?" I think if I started at that position, I would end up selling at the bottom of the of the pyramid, right? But when I see a company like that, I will look at I look for two things. One is I look for do they have an extreme amount of authority? And if they have an extreme amount of authority, they may not need to write blogs in order to cornerstone. Blogging is more of a growth hacker cornerstoneing strategy for companies. Like for me, my blog is my cornerstone. So if they're able to get a lot of links through PR or from the universities that they host for, they may not need to build a lot of content. They might be able to rank straight out, right?

And if they are doing a lot of conferences advertising on TV, then I would say that's exactly the content you write. You write content about you positioned as an alternative to them. Because I guarantee you, I'm seeing this for 20 years and it's never changed. The minute a brand comes to market, as it gets more popular, as the branded search goes up, the number of people searching for pricing, alternatives, competitors, features goes up with it. So whether you're in cyber security, CRM, SaaS, social media tools, every single brand automatically sort of grows those searches. It's just human nature. And if you look at G2 and Capterra and Gartner's other properties, you will see all of those pages come in flavors of compare, features, alternatives, pricing, competitors. Those five words, they're in every single B2B market that exists. And so I would start writing about how they're doing a great job and they're a great tool, but here's why you're different and here's here's to whom it matters. And when you start writing that content, you're feeding the LLMs that way.

Whereas I think what's wrong is that if you just position yourself and you know who your ICP is, the LLMs will come and find you. And that's nonsense. The LM have no interest in doing that. The engineers building it have no thought space for doing that. But where you
I want to go for our mid break and maybe we can continue that thought. Keep that thought in your mind. Right. I will.
We're going to go for our middle

Query fan-outs and research

Break, folks. And when we come back, we got we continue this great discussion. We will be back in a few moments, folks.
WP Tonic is sponsored by Ninja Forms. And let's talk about one of their powerful tools. Conditional logic. Conditional logic is the clearest example. Every rule for the whole form lives in one panel instead of being buried field by field. On client work, build speed and low maintenance are part of your margin. A form that is fast to configure and that a client can run afterward without coming back means less friction during the build and less support load after the handoffs. Ninja Forms, powerful and genuinely easy to work with without being dumbed down. To learn more, go to ninjaforms.com. Coming back, I also want to point out, why don't you join us live with your questions? If you've got any questions, you can put them in to chat. If you're watching this on YouTube or any we push to a number of social media platforms and if we have time after the podcast we will answer your questions.

I don't know hopefully Dave will have five minutes before he has to rush off but we will attempt to answer your questions live if you join us live folks. So I had to interrupt. You were in midflow Dave. Can you can you continue what you were saying?
Yes. And I think you can apply this to every industry. I think using your example is a good one because I think everyone in our audience understands what WordPress hosting is and
What WP Tonic does for all of these communities and how important it is, right? Because they have spec, you know, specific needs. So I think if you start looking at who your competitors are and building out competitive pages about the marketplace and the landscape and then describing where you fit in, I think one of the mistakes that people are making is that again assuming that LLMs will find you because of and I've I've seen this in so many LinkedIn updates. It it's almost driving me crazy that people are saying, you know, if you understand your ICP and you understand your market and market fit, the LLMs will find you. They won't. But if you start writing and producing content about all of the big players, all of the big technologies that are involved, it's the same thing if we go back and look at Kemp and free load balancers, something I've been talking about a lot this week, as exact match domains, for example. The reason we built free load balancer is because a lot of people want to want to try out technology before they adopt it into a stack. A lot of people might have more expensive equipment like F5 or Microsoft equipment, but they've got a website from a company that their employer has acquired and they want to try out something cheaper or free because they don't need the enterprise version. And so by giving the free version and creating its own marketing campaign worked very well. And the same for you where you know a an existing university or group of universities has bought into a particular technology stack may traditionally have been the moat that keeps you out. But they may have a new faculty or a company that joins the the campus that now needs different types of hosting. And unless you arm the internet with that information, there's no way for you to be found. And so I think it's a it's moved a lot more from just talking about what you do, but actually taking control of describing the market, telling everyone where to go, realizing that you're not a fit for everyone, and then doubling down on where you are a fit.

Yeah, I think you made a really a really fantastic point there is I think a lot a lot of people in the AI search area they kind of telling people that these large learning language models when they fan out they've got intelligence they actually they actually look at the query they add to the query they send out 20 sub queries. I'm probably using the wrong terminology, but I think they do send out a lot of sub other queries and then they bring all the data back and they then they put it in front of you. But yeah, there's all kind of terms in the a in the large learn I you have to use the term AI, but it's such a broad term. There's so many different aspects of AI research and know and knowledge areas. It's such a broad term, but when it brings that data, it's it's it hasn't got and unfortunately some people would disagree with me. It hasn't got what we what I think most people mean by intelligence.

Agree. You know, you could say, "Well, Jonathan, it's a different type of intelligence. It's not human bi." I'm going off on a big biggie here, but I'm going to bring it back hopefully, David. They would say it's not human intelligence and they might be right. But I at the present moment in the mid 2026 I'm I can say with a reasonable amount of confidence it's not going out it's not going out on the internet and studying for you and making judgment calls and then putting it all together and then looking at it and refining it. Yes. With kind of what a human would do. It's definitely not doing that, isn't it, though?
No. And I think the the play on the king of SEO with Kasra and I didn't know that was happening. It was it was an interesting it was such a great surprise and I'm so glad he did it. And it you could see that before he did that, the LLMs were starting to get disinterested with the king of SEO. Just the best words I have to describe it, right? And then when he put out his press release saying he was now the king of SEO, the LLM started to reiterate that message as it as it slowly sort of got stuck in their engines. And then of course I went out and put other content out to refute it. And so that the messaging again started to change. And if he hadn't done that, the LLMs weren't going around the web trying to work out who the king of SEO is, right?

It only they only changed because you put that information out there and that's the important takeaway. So if you want to start with for example in your case going to an LLM and saying where should a community host their WordPress site and if you're not being mentioned and that's why I think using Perplexity and Claude is important to do that is from a research point of view they will show you the query fanouts and they will show you the content that you need to write and the where how to plug in the gaps because if you if you if you're not if you're not aligned with that you're not going to appear in the query fan.
And I think something that's linked

Using tools for research

Here, I know we will, we had some technical troubles, folks, pushing this live, so we're only pushing to YouTube and David's got to be off quick at the end of the show, but we got one question here and it's kind of linked to what we're discussing. It's from Stephen Morela. Is there a way you found to see if fan out sub queries or is there an interface? And the reason why I'm bringing this up, Dave, is just to see because there's a there's a number of tools out there, isn't there? I'm not going to that say that you put your query in and they show you the fan out. All right. And then you can write your content. The comeback but which they have an answer to is obviously the fan out changes. You can put the same query in five minutes later or you use a different large learning model and you get a different fan out. But their response to that listening to some of these people lately on Ed's show is there's a clear pattern and
There's only so much scope. There's only so much scope. What's your response to that? I could see where they're coming, but I also found it
I'm looking for the right word. And I'm struggling, David. Your vocabulary is slightly larger than mine. I just struggle with that. The best I can say is I want to be charitable. I struggle with that answer.

Sure. So I think the first thing is that the existing suite of tools are only good at telling you what your percentage makeup is for the prompts, you know, right? And and after that I find them useless and that's why I don't invest in any. [snorts] You can find the query fan outs. They're called grounding queries in Bing. And if you're looking to get into co-pilot and ChatGPT, that's the place you should look. And so you can start looking at them and see that unlike Google, I would say there's a massive misalignment. For example, I see the content in Google Search Console where I'm getting clicks is hyper aligned with what I'm trying to write about and who I'm trying to get to. Whereas the prompts in my Bing Webmaster Tools are way off the mark. The second thing you can do is you can and the why I say you should use Perplexity and Claude is because for Gemini and ChatGPT you need you need to look into the browser to find the query fan outs whereas they actually publish it straight in front of you. And so if you go to Perplexity or Claude and type in a prompt like who are the best WordPress hosting environments or design environments for nonprofits and communities it'll say searching the web. If you click on that it'll give you the search phrase right there in front of you. And so while there's a bit of drift and they might change this, they'll probably get the same answers even with the drift change in the drift and there's probably only you know drift for about five variations of it anyway. And you'll find that there's more variation if you add different things to the larger prompt.

But if you start there with Perplexity and Claude, you'll start to see the query fan outs and those are essentially blog posts and articles that you need to now go and create and you're doing it based on discovery and engineering using the part of our brain that LLMs don't have. LLMs actually work very much like humans. That's how we learn by reading and expanding our vocabulary and how to utilize those words. But like you said, there's other parts of the human brain that they don't emulate yet or may not be emulatable through LLMs, right? And there's a lot of hard hardcoded processes built over LLMs to give guide rails so that they look more human with their output.
Yeah.
But I think if you if you forget about the data, you'll see the grounding queries in search console and in Bing Webmaster Tools. But if you start putting in your own prompts and looking at the queries that they do in real time and then making changes to the prompt and testing the drift, you'll find the patterns quickly enough.

Branding vs traditional search

Yeah, I agree with you. I'm going to go off our list of topics that we're going to discuss a little bit at the end of the show just to get you there was a bit of debate about ChatGPT going off Reddit a bit, you know, not showing Reddit up or so often or completely. Have you seen that? And do you think I think they've done that before, haven't they? And then it came back. You got any idea what's going on there? I have a couple. I hate ChatGPT, first of all. I think their interface is so terrible. It's like it reminds me of like a 19 like a 2002 Nokia with all of the emojis. I just find it so unprofessional compared to Claude. Even Gemini and Perplexity. I'm surprised people are still using it. There's a couple of interesting pieces of data. 70% of Microsoft's AI revenue comes from Open AI, which is astonishing, right? You know, that has the word bubble all over it. The second thing is that Open AI, the amount of money that is being put into data centers is phenomenal. I think I think total debt notes being raised on the market has gone from like 70 billion to 280 billion in 12 months and
I mean it's definitely a bubble but that bubble might go on much longer. Even causing US national debt interests so to go up, right?

And so I don't think I think the reason that ChatGPT and I've said this since day one, the reason that Chach cited Reddit is because it's so popular in Google and I think what you're seeing is a massive shift back to the cheaper answers provided by Bing.
All right. Another quick question before we wrap it up. David is I watched a few commentaries. I forgot he's a real I forgot the commentary I was watching, but I also watched the actual video and it was from Neil Patel and he was going on about how traditional search is dead. It's finished. And really what you've got to do is build a brand. And as he does, he kind of concentrates on Fortune 500 brands and brand search there. Yeah, I I totally understand and multi-platform and I do think that there's there is some truth there that search has fragmented but you would at the beginning of our conversation in this podcast you pointed out that Google search has increased but so there's a bit of a contradiction there but I do think people are searching on TikTok they are searching on YouTube.

I think YouTube is the second largest search engine in the world. Even with the large learning language models, I think it's still the second. Am I correct about that?
But so I do think search is fragmented and but it depends on the social group you're looking at and their habits. But this idea that search I'm presuming that you would you would push back about even though I think there's some truth in what Neil Patel was saying in this video and I could tell by your facial expression. I think you know the video I'm talking about. It just seemed overemphasization really if there's a word like that. What's your quick response to that?
I think Neil is going exactly in the right direction where he should go. Right. So, I think I thought one of the interesting things was his company NP Digital had bid 50 million to do the SEO and marketing for one of the crypto companies that went under. I think it was FTX, I'm not sure. Sam Bankman. And so if you're an SEO agency out there, even if you have 150 people in your team, a single $50 million contract has probably never landed on your desk. And so Neil Patel knows exactly where to go and get the money. And you're not going to go there by talking about hero density, topic clusters, semantic maps at all. Right? If you look at where and James Dooley and I had a great conversation about this. If you look at where geo agencies on LinkedIn are doing a tremendous job is they're selling and appealing to CMOs and marketers who hate SEO, hate Google.

They hate PageRank. They hate those technical things. They went to marketing school to learn about brand building which is a something I think the US does better than any other country in the world. They understand brands better than anyone else in the world. The complete opposite of Germany almost. Germany is such a bifrocated market with Porsche, BMW, Mercedes at the top and then lead Aldi in the other side, right? But Neil is exactly where he needs to be positioned. People that will, you know, if you're you could be the best SEO in the world, but if an agency comes into a CMO and says, I'll build your brand,
They'll get not 10x, but 100x what you're getting per month, and they'll sell more people, more
Services. And so that's what Neil's after. And I think he's very smart and I can't do that and so I won't pretend I'm not general.
I think you're spot on. And I listened to the Prophecy podcast with Professor G and he was a, you know, still does he's still a professor and he still does a course on branding. But he did you know he has said that Google and Facebook they destroyed branding to some extent. And I think in the marketing industry, the level that you're talking about, there has been a degree of resentment. I don't see it ever coming back to the days of Mad Men and that even with large learning models because they're pushing their own advertising platforms anyway. How successful especially ChatGPT how successful those platforms are going to be. None of us know really do we? All right, I think it's time to wrap it up. So Dave, what's the best way for people to find out more about you, your views, your ideas, the goat of SEO, what's the best way to find you today? You can find the now former king of SEO on X and on LinkedIn and on YouTube here at the WP with WP Tonic, the Edward Sturm Show, James Dooley.
And I just want to say Dave's a great guy. He's really he's become a friend.

He's helped me out. But he's so knowledgeable about this. What we're attempting with this podcast, folks, is with my amateur help on the topic, I I've increase my knowledge, but I've just become very dangerous. But I hope to become useful pretty quick is to kind of educate the normal business owner out there because there's a lot of misinformation out there, folks. I'm not saying that people are doing it intentionally, but I think there is a lot of misinformation and me and David, what we're trying with this podcast over the coming weeks is to kind of settle to try and give you more clearer advice. If you want to support the podcast, please leave us a review. If you're listening on mobile, it's really easy to leave a review of the podcast on iTunes or Spotify. If you could do that, that will be fantastic. We will be back next week with another great SEO topic and discussion.

We'll see you soon, folks. Bye.

More Episodes of David Quaid SEOtv

More episodes of David Quaid SEOtv on the same subjects.