Episode 3

Does Topical Authority Actually Matter for AI SEO?

November 5, 2025 · 30 minutes

Share this episode

A lot of SEOs talk about "topical authority" — but does it actually move the needle in AI Search?

In this episode of The AI Search Report, Nathan Gotch breaks down what topical authority means in 2025, how it's being measured by Google's AI Mode and ChatGPT, and what you should focus on if you want to stay visible as search evolves.

We'll also dig into the latest AI search trends, including how YouTube and Reddit have become key retrieval sources, why Grokipedia published 885,000 AI-written articles, and what Google's "billions of clicks" claim really means.

Whether you're building content clusters or rethinking your keyword strategy, this episode shows you how to build real authority that earns visibility across both traditional and AI search systems.

You'll Learn

  • What "topical authority" really means in the AI Search era
  • How Google's AI Mode and ChatGPT evaluate authority differently
  • Why YouTube and Reddit now influence AI retrieval results
  • How to structure your site for measurable topical authority
  • The right way to prune or cluster underperforming content
  • What we can learn from Grokipedia's 885k-page AI publishing experiment
  • How to align your SEO strategy around intent, not just keywords
  • Which new AI tools can help track visibility and citations

Key Takeaways

  • AI search engines reward usefulness and coverage, not just keyword density
  • Topical authority is becoming quantifiable — treat it like a metric, not a mystery
  • Fewer, more focused pages outperform scattered topical content
  • Building authority now means mastering both human and machine signals
View Full Transcript

00:10

Dedos and Spark, Toro just launched a state of search report. Just going to show you a couple quick data points here. So number one, when we look at the overall search landscape, just traditional search, we're not talking about AI right now, just traditional search, Google continues to be the dominant force. So as of right now, the main players in search, we know that we have Google, Bing, Yahoo, DuckDuckGo, and Baidu obviously much more prominent in China. But the other one here that's interesting is we also have DeepSeek as well. DeepSeek is actually basically a search engine that's owned by China, but it is relevant in the US as well.

01:08

But over the last, I guess three months here, it's slowly been kind of just reverting right back to Google being that dominant player in traditional search again. So in essence, not a whole lot has changed. Another thing here is the distribution of clicks over time when people are actually using Google when they conduct queries. So there's a bunch of data here, but the main query that you really want to pay attention to is that about actually over a quarter of clicks. Over a quarter of clicks end without a click. So that means people search query on Google and they actually don't even click on anything. So think about AI overviews, think about featured snippets. Think about how many Serp features Google has. That's why this number is so high.

02:08

So that's a lot of people in the grand scheme of things, how many searches are conducted per minute on Google? It's a lot. So over a quarter. And then also the intent of queries that go on, this is specifically desktop search. Only about 1% of all searches have purchase intent on Google, so only about 1%. And then about 20% have some sort of, let's call it investigative intent, where they're actually looking for information to make a purchase. So this might be something like best SaaS, best CRM software, best, whatever, whatever product you can think of. So about 20% have some investigative intent.

03:08

So overall though, we're looking at about only about 2020 to 21% of searches conducted on Google actually have some sort of path to conversion. So that means it was a classic 80 20 rule, and the other 80% is really probably informational in nature. So when people talk about losing traffic to their websites, that has affected pretty much everyone. But a lot of those queries and a lot of that traffic that has been lost, it was informational queries that likely were very top of funnel and probably didn't have a really strong path to revenue, at least direct path to revenue.

So just keep that in mind. That's why you want to focus, of course, we know this. I'm not saying anything revolutionary here, but you want to prioritize those much more commercial keywords and focus on creating commercial-based content because those tend to be much closer to revenue. We want to focus on those much more than just these really, really top funnel queries that are just so far from conversion.

04:01

And then the top domains visited in traditional search. So in Q3, 2024, number one was YouTube. And in Q3 2025, once again, it is YouTube. So the top five here right now, we have YouTube, Reddit, Amazon, Wikipedia, and Facebook. But the biggest change here is actually in Q3, 2024 chat, ChatGPT wasn't even in this top 15. Now ChatGPT is actually number six. So as far as traffic is concerned, that is the trajectory we're going. So chat GBT is being used a lot more when it's all said and done.

04:49

And then the next one here is the share of desktop users using AI in the us. So this is basically showing the next one here specifically about AI platforms, and it's the share of desktop users using AI in the US chat. ChatGPT is hands down the most popular one as of right now. And then Gemini's number two. And then after that, it's basically a toss up between perplexity, Claude, copilot and DeepSeek.

Okay, so if you're thinking about how do we actually drive AI search performance and how do we rank in AI, you absolutely, number one focus should be ChatGPT because it's the most popular one. So if you can control and understand how to do things to increase your odds of appearing in ChatGPT, that should be your number one priority. Okay? So that should be at least your number one focus. And in general, like I said, if you do well on ChatGPT, that will definitely bleed over into the other platforms as well.

05:38

Okay, share of desktop users using AI in the us. So as of right now, according to this data, we're looking at about 37% for ChatGPT, and about 9% for Gemini. Okay, so we're getting close to about 50% of people are actually using AI in the us, which is pretty crazy. That's moving very, very quickly. And so the last one I'll point out on this one is just once again shows the share of desktop users on content platforms in the us.

06:30

So what are the most popular content platforms? Now, this is very, very important for AI search because once again, we see that YouTube is hands down the most popular content platform followed by Reddit, Pinterest, LinkedIn, TikTok, X and Facebook, all the usual suspects. But when you're looking at trying to drive AI search performance or AI visibility in ChatGPT two or the AI platforms, we know that YouTube and Reddit are very big sources of retrieval.

So when you run a prompt and you ask a question on ChatGPT, for example, a lot of times you're going to find a lot of citations from YouTube and Reddit. And you think about why. Well, obviously the numbers check out because there's a lot of people using both of those platforms. So once again, it's another kind of area where obviously the best time to start building on YouTube was 10 years ago. But the second best time is today. So you might as well get going right now.

07:19

And there's two ways you can attack YouTube. Number one is obviously building your own channel. This is the hardest thing to do, is building your own channel. It's going to take time, it's going to take effort, it's going to take consistency, and you're going to have to learn how to do that. Okay? So right now we're on the rank ability channel here. I have another channel, my personal brand channel got SEO. We built that channel up to about 60,000 subscribers last year. And then this year we basically, I was running essentially our GEO course and just was running live sessions every single week. So I was basically streaming live, live streaming on YouTube every single week.

So that's path number one is building up your presence, which I think you should begin. You should have begun yesterday. And the second best technique is actually just to identify influencers on YouTube. Get influencers. You can actually pay them to talk about your brand as an advertisement, and that will actually be used in retrieval because the way that the AI platforms work is not that they're actually going to YouTube and watching your videos or watching someone else's videos. What they're doing is they're, they're sifting through the transcripts on the videos. They're basically using those YouTube transcripts as sources of retrieval.

So if someone does a video about you and at the end of it they say, by the way, this is sponsored by X, Y, and Z, that is more than likely going to be mentioned in the transcript of the video. Therefore it can be used in sources of retrieval. And I've seen it happen. I've also done it on our own channel with our own sponsors that we've worked with. That's the second best path is that. So you have two different ways to attack YouTube. The second one is much faster and easier to do, of course, because you're just basically, you can get someone to produce something if you paid them. It's not free though, and it's much easier.

09:18

Alright, so now moving on to a post from Sean Anderson. He said that topical authority is no longer a theory. The Google leak reveals the site focus score, which algorithmically measures your site's niche focus stray too far from your core topic site radius, and you dilute your authority. Specialization isn't just a strategy, it's a score.

Okay, so I wanted to highlight this because this is obviously a highly controversial topic and you can always find examples that disprove this. However, I think just in general, you want to follow this as a principle, which is you want to really focus on your core topics and you don't want to meander too far from that core topic because you will dilute your authority. I think at a certain level, if you have a very strong website, let's say you have a 90 DR, you probably can dilute your authority a little bit more. But I have audited hundreds and hundreds and hundreds of websites. I've done probably the over course of my 16 years or 17 years now doing SEO, I've audited, I don't know, probably over 2000 sites. And I've seen this phenomenon where a site will have a lot of content that's off topic or wildly off topic.

10:21

I'll give you kind of a small example here. When the pandemic was going on, I was doing audits on a lot of personal injury lawyer websites and they would be talking about, I won't say it because YouTube might flag it, but how to avoid getting blank on an elevator. So this would be a personal injury lawyer talking about this topic, but they're not a subject matter expert and that's not their field of expertise. So what happened is they had all of this content on their site about this topic that had nothing to do with personal injury law.

And that content was ranking really, really well, but once we pruned it, we deleted the content. That's actually no longer relevant. All their commercial keywords actually went up. So I have done many, many, many audits where this is a major factor. And so I'm not an alarmist and I don't follow a lot of the trends per se, but this is one area that I think you should have a really core competency on a very core area, core subject. And when you go off topic, don't go so far off topic that you're talking about something that's not related to your niche. Like I will talk about AI search. I won't go wildly into traditional link building just because we're mostly focused on AI search right now. So that's kind of the parameters.

11:15

I'm not going wildly off base here, okay, you're still really close to your core topic. If all of a sudden I start talking about organic Facebook marketing, technically that's still in the digital marketing space. So it's still somewhat there, but it's still too far off from that core competency. So focus on going a mile deep on your topics and there's so much opportunity to expand that people greatly underestimate how much opportunity there is. So that's just more of a just, it's a rule. In general, I would follow this rule. I would pay attention to this site radius, focus score. Obviously, like I said, there's always going to be counter examples to this where you can find, there's so many variables in algorithms that you can find someone that breaks the rules and still does well, but it doesn't mean that this is not an important rule and an important variable to pay attention to.

12:12

Alright, and then another one that's really, really fascinating just as an SEO case study is Grokipedia. So we know that X just launched Grokipedia, which is essentially a Wikipedia alternative, and as of right now they have published 885,000 articles on this platform. Okay? So I'll tell you how my theory as to how they're actually doing this, but the reason why this is really fascinating is because like the ultimate SEO case study, because we're going to see what happens when you have basically AI generated content, but it's supported by what is already a very strong domain.

So X is already a very powerful domain. They have a lot of links. They have a lot of credibility and trust associated with that domain. So we're going to see will that will overall domain strength outweigh this kind of AI generated content. So when we look at the data here and those are listening, I'm looking at an Ahrefs screenshot here, and you can see as of right now, they already have 1400 referring domains. I think their domain rating is already up to 50 and it's been like a week. And as of right now they're showing about 56 organic pages, which it's already more than this. When you do a site colon search in Google, you're going to see already more than 56 pages. So it's clearly growing every single day. And once again, it's probably like 885,000 URLs right now.

13:30

So obviously Google is still working through it. But once again, my prediction is that it's going to do really, really well because it has all the links it's going to need. They're already getting links, they're already getting links from major publishers. It will be absolutely impossible for them to not gain a significant amount of organic traffic from traditional search. I don't know how long it will last, but I would say it's probably going to be pretty impossible for them to not see results given how many links they're already getting.

14:05

Alright, a couple interesting things about this because obviously I can't help myself is an SEO just looking at this. So number one, when you look at one of the entries on Grokipedia, I'm looking at the Elon Musk entry here, and it's more than likely that they're using DeepSeek. It's certainly possible they're using Grok as well, but in general I think it's probably DeepSeek because DeepSeek is a very cheap way to do this at scale. So in general, this is probably an automated process. They probably have scrape data from Wikipedia and they're using DeepSeek to rewrite these articles or entries on Grokipedia.

And then one of the things I noticed is that when you click on the citations, and this is obviously like a part of the functionality of the site, they have citations. So when you click on the citation itself, it should take you down to the footer, but instead what it does is it just takes you to the external link immediately.

15:00

What they should be doing is if you click on a citation, it should just go down to the footer and show all this. So it should be a jump link. It shouldn't be an external link going out. So I'm not sure who made that decision. It's not exactly the best strategy if you're trying to keep people on the site. And then the other thing that really stood out to me is because when I think about, just from my SEO mind starts to get to think about this is like, okay, 800,000 pages. So we have 800,000 pages. Before I would publish that many pages on the website. The thing I'd be thinking about is like, okay, how are we going to be able to optimize this site for crawling and indexing? And so one of the best ways to optimize for crawling and indexing is through internal links.

15:44

So based on what I'm seeing here, I'm looking at the entry number one, I don't see any breadcrumb navigation, which would be absolutely critical in a scenario like this. When you have potentially in the next couple of months, there's going to be millions of pages. So the breadcrumbs is one small little mechanism you can use to make sure that the crawlers can access all these pages in an efficient way. Okay? So that's number one. Number two is within the entry itself, there's no internal links. So over time, as they build new entries, they need to have some sort of programmatic process to inject those internal links to other entries. So that way that you don't have any pages that are more than three clicks deep in the architecture. So just looking at this entry here, we see, okay, we're on the Elon Musk one, and we go down here and I see certain entities.

16:33

We have PayPal, we have SpaceX, we have x.com, we have the boring company. All of those are entities that should have their own dedicated page. And when you mention those, they should have an internal link going to those entries and then you repeat that over and over at scale. That would help the pages get proper internal link coverage and allow every page on the site to get crawled and indexed appropriately. So internal links are the key to basically solving a lot of this, which by the way, I'm not saying is easy, and this is a huge, huge operation, but that is the way that you would handle this. And if you go and look at Wikipedia, I'll show you here, this is Elon Musk Wikipedia page specifically, and you can see how they leverage internal links really, really well. Okay. You can see they're linking to Tesla and SpaceX and Twitter. So they have entries for each of these entities, which allows them to have just incredible crawling and indexing capabilities. Plus they have a lot of links, which obviously helps with that as well.

17:46

And then I wanted to show you this part because this is more than likely happening as far as how they're able to pull this off. Of course, my mind was just thinking about how are they able to do nearly a million articles or a million entries at this point? And so more than likely, the simplest solution is usually the right one. So more than likely what they're doing is they're just scraping Wikipedia. So they're extracting all that content, they're running it through their LLM, which is grok, and then grok is likely, for a lack of a better word, regurgitating that content. In other words, rewriting it, and then they maybe run it through a little bit of QA and then they publish it on the site. So there's probably some automation layers going on here, but in general, that's at a very, very high level.

18:35

That's probably what's happening. So extracting from Wikipedia, run it through grok to rewrite it, publish it on the site, and I don't think it's any more complicated at a high level than that. I don't think anything highly sophisticated is going on here, but that's probably what's happening. Now, is that ethical? That's not for me to decide, but I just think obviously it's quite the experiment for us to at least track and pay attention to.

19:00

And then the biggest news here is Alphabet. Google of course, announces their third quarter, 2025 results, and they had a really, really good quarter, one of their best ever. In fact, it was the first time that they've ever had a hundred billion quarter, which is just mind blowing as far as the numbers. And what's really fascinating, of course, is most of their revenue comes from search and specifically Google ads.

19:32

Okay, so we're going to talk about this here in a second. So couple kind of highlights from this. Gemini is really has a lot of usage. They said they claim they're up to 650 million active users. What does that really mean? Okay, does that mean that someone went on there and did one query? I don't know how they're defining active user, and obviously there's different levels of usage, so it's always a fascinating number one, they use that, I would take it with the grain of salt because they don't really ever define what an active user is, but going through this, and also they have 7 billion tokens per minute on Gemini. So it's a lot of usage no matter how you, no matter how you think about it. Okay?

Alright. So a couple quotes here. This growth rate increased in Q3, largely driven by our AI investments in search, most notably AI overviews and AI mode.

20:32

And this is a continuation, which is it now has over 75 million daily active users. So these products are being used a lot, which we were talking about earlier in the statistics that over a quarter of searches end without a click, you can think AI overviews, and now ai, the emerging AI mode here, that's going to be even worse, more than likely. And we're going to talk here in a second about some of the things they didn't talk about in their shareholder reports here.

Another one here from them is search. And its AI experiences are built to highlight the web, sending billions of clicks to sites every day. Okay? Just remember that quote. Okay? This is what they said. They said, sending billions of clicks to sites every day. They never say how many clicks. They never say if those clicks are growing, okay? Just keep that in mind as we go through this.

21:29

Another one here, our investments in new AI experiences such as AI overview and AI mode, continue to drive growth in overall queries, including commercial queries, creating more opportunities for monetization. Keep that in your mind as well. So just remember billions of clicks to site every day. So they claim they're sending billions of clicks to sites every day and that overall queries are growing. Keep in mind, queries are not clicks. So the reason why queries would grow in this scenario is if you're using AI mode and you search your initial query, then you ask 3, 4, 5, 25 follow up questions, each of those gets counted as query. Okay?

So of course queries are going to grow when you use chat platforms. ChatGPT, same thing, right? Same exact situation. Gemini, same thing. Claude Perplexity, they're all the same. And those queries will grow because every single time that you prompt an AI chat platform, that's a query. So of course that's going to grow.

22:41

And so now to the conclusion to this, which is what's conveniently missing from the transcript and the quarterly report. Okay, so a couple of things. You didn't see anything about organic clicks, you won't find anything about organic CTR, you won't find anything about organic traffic, and you definitely did not find anything in these reports about AI effects on publishers. So why would they just omit this information if they claim that organic clicks are so healthy and that they're sending billions of organic clicks to websites? Just something to think about. And the truth is, if it was a favorable number, they would report it.

So the reason why they're not reporting is probably because it's not super favorable and it is because of these AI products. And these AI products are killing clicks. Let's just face it. They are killing clicks, which is good because in some ways, because it's forced a lot of us to change our strategies, we can't just rely directly on Google anymore.

23:45

And as a result, it has allowed us to explore much broader strategies going after the AI platforms or thinking more about YouTube or whatever it may be. I think it has kind of forced a lot of SEOs to think much more broader and become more multifaceted in their skillset. So in some ways there is a silver lining of this, which I think SEO professionals are the most equipped to handle this new environment. But at the same time, Google has made this very, very challenging for publishers specifically, especially any website that is relying on ad revenue or affiliate links. It's tough out there. I mean, I would be very, very nervous if it were me in that scenario.

Now, of course, businesses in general that sell products and services, they can adapt much better because ad they're going to lose a lot of that top funnel traffic, but at least they can still focus on the commercial queries and still be able to drive revenue.

24:48

So the effects on revenue still remain to be seen. As far as ai, I think it's more than likely that it won't affect businesses that much. I think who's going to be affected the most unfortunately in this scenario with Google is obviously pure publishers, content businesses, so that rely on organic search traffic from Google specifically. So just keep that in mind when you're looking at this. Google is, Google is a very, very smart company, one of the smartest companies ever in the history of humanity. And they don't just conveniently leave out critical details about their search engine. That's not an accident. It's not an accident that they just didn't talk about clicks or organic CTR organic traffic. I'm just telling you that is not an accident. That is by design, and that's because it's not super favorable and don't expect to see anything about clicks going forward.

25:43

I would imagine you're not going to see that really ever. They're going to talk a lot about queries, they're going to talk a lot about revenue growth, but it's, remember Google's business model is dependent on getting advertisers to use their product. So when there's not as many clicks for organic, you are left as a business to just pay for ads. This has always been the case. And slowly this has been happening over many, many years with SERP features and various things. Google just slowly keeps chipping away at clicks, organic clicks specifically. And when AI came around, it kind of gave them even more power to attack this. So I don't suspect it's going to get any better as far as CTR. So you need to refine your strategies.

26:39

Okay, last thing is I've got some really, really good free tools for you that you can use. So if you go to rank ability.com/tools, we built some really good tools recently. So I'll show you a couple of these. And if you're listening in, once again, just go to rank ability.com/tools and you'll be able to find these. First one I'll show you is actually an AI search index ability checker. So you can just take any page like this. And the reason why this is important, because first of all, index ability and crawlability is kind of weird terms when it comes to ai, so just bear with me. It's just what people are used to. There's not exactly an index that ChatGPT has, right?

ChatGPT is not like Google traditional search where they have an index of pages. It's not exactly how it works. They do cache pages, we know that, but no one really knows how long something is cached, but they don't have an index, okay?

27:31

So we're just using SEO lingo here, but in general, this tool, you run the URL through and you're going to be able to see is the page are AI crawlers actually able to access that page? So I would say accessibility maybe would be a better word. Do the core search engines, is the page actually indexable? So that's going to be Google and Bing, which we know are used predominantly for retrieval for the AI platforms. And then also the uptime of the page and the response time. Response time is important because these AI platforms are really, really intolerant of slow loading pages. So ultimately you want to just measure your page on this. So page by page basis, you can do a quick little analysis here. You can also download a little shareable link here if you want to share that with someone. So that's a one free tool.

28:20

Another one you can use as well is this AI search query generator. Once again, another free tool, but this is really important too, if you're trying to do AI search tracking and analytics. So you want to be tracking to see if your brand is showing for commercial queries. So I show some good prompts and bad prompts on the page, but basically you can go in here and just search something like baseball cleats and what this, if you enter a seed keyword and then it will produce some synthetic queries for you. And we've tested this quite comprehensively. We've used queries directly from a Google search console. We've used synthetic queries. The AI is really good at producing synthetic queries, okay? They're pretty realistic. So you can take one of these queries, run it through ChatGPT, grok, whatever AI platform you want, and then see if your brand is popping up in that AI generator response and then repeat that across and many prompts, okay?

29:16

You want to have a lot of prompts, a nice prompt bucket, 25 to 50 prompts around one particular product or service to get a good surface area about your visibility. Okay? So this is once again a totally free tool you can use. And then the last tool I'll show you is another tool we just built yesterday is that people also ask Explorer. So this is free, once again, you just go in here, you can enter your query. So let's do baseball cleats. Run this through. And this is actually going to pull live data from Google SERPs. So you'll pull this and we've got a nice two different views for you. So you've got a list view here where you can just go and every time you click on a query, it will open up more queries, and then you can also look at the tree view so you can kind of visualize what this looks like if that's your thing. Okay?

So these are all free to use, so check 'em out. Once again, rank ability.com/tools, and I'll see you next week. Thanks.