The AI Search’s Impact on Publishers - #SEOFOMO TL;DR June 16, 2025

The latest in SEO and AI search is brought to you today by the SEO Formula to Land in Read. And today, to go through the top news from the last week, I have Barry Adams, who's an SEO specialist consultant focused in the publishing world. Before we start reviewing the top news from the latest week, I want to thank Similarweb. Similarweb is our today's sponsor. Thank you very much for your sponsorship. One of today's top news that have received more clicks are aligned on the same topic, which is why I wanted to invite you today, Barry, regarding cool AI mode, the first thoughts and survival strategies in particular for publishers, right? This is also very aligned with another top resource of the week, which is the interview that Gianluca Fiorelli did to Lily Ray, pretty much highlighting this particular topic, how publishers can survive. If you have a transactional or commercially focused websites that sell a product or a service, at the end of the day, will end up in some ways suffering far, far less because the actual product or service is provided through a conversion and the goal is the revenue is connected to a conversion. Unlike publishers, publishers rely in impressions, in traffic, raw traffic for their business model. They sell ads. So Barry, can you please explain a little bit the thoughts that you share on this piece regarding first what you are seeing across major publishers and then the survival strategies, the tips that you share here in this piece and any others that you think that might be good to take into account for SEOs working in the publishing world?

Yeah, I think it's really interesting to note that AI overviews, what we currently have live all over the world, tends not to show for news topics. If it's a very recent news event and there's a top stories carousel on the Google search results, it tends not to show an AI overview, which means news-focused websites haven't felt quite as much pain just yet, even though they're still feeling quite a lot of pain because they also write things like backgrounders and explainers and other related articles around the news event. And AI overviews do tend to show up for those. So I've spoken with some publishers who are seeing traffic declines on those sorts of background stories or evergreen content stories. Of like up to 40% less traffic from Google. But the news content is still driving a lot of traffic because AI overviews do not serve for news. And that's a key difference with AI mode. AI mode will give you a summary of the latest news. It will respond with detailed explanations of current news events if you ask it for it. And I think as a result of that, if AI mode rolls out more widely and it gets more user adoption, I expect it to have a very dramatic impact on traffic to news publishers all over the world. I suspect even though AI mode will try to cite sources and will try to show source links, you and I both know it'll not drive as many clicks as you currently have in the top stories box or in a regular Google search result. So best case scenario, you, you hang onto about half of your Google traffic if AI mode becomes the default.

Yeah. In one of the latest news, they actually said that they're aiming for that little by little because eventually I guess that here the criteria will be if they are able to test it out in the next few months while keeping their ads revenue or even finding ways to increase it even further. And depending on that performance, they will push it further or not, I imagine.

Yeah, absolutely. I mean, I think monetization is the only thing holding Google back because I think in terms of user feedback, AI mode is a success for Google. And I have to admit this myself, I'm an AI skeptic, as you probably know. I have used AI mode. I like it. I like using AI mode. I like that I can just dive deeper with follow-up questions and get more specific results of what I'm looking for directly in the search results without having to click through to a website, having to click through cookie consent banners, email popups, getting bombarded with advertising. I can just stay in AI mode and do all the research that I need to do. And if that becomes the default, then, you know, traffic to information-based websites, including news publishers, is going to take a massive, massive hit. And that is something that publishers need to be prepared for. I think you can't afford to wait until it happens. I think you have to be proactive and start looking at ways to diversify your traffic acquisition, diversify your revenue stream, but also prepare for a landscape where there just will be fewer clicks. There will be less traffic to your website, no matter how successful you are, no matter how broad a net you cast in terms of the channels you focus on. We are going full steam into a future where there will just be fewer clicks to websites and you need to be ready for that. So far, only 1% of users on average are using AI mode as a separate tab, but Google is going to make it more prominent. So this is going to increase. Google Discover is going to pick up some of the slack because it's going to roll out to desktop. I already know some publishers who are very heavily invested in Discover as their primary traffic source. Unfortunately, Discover algorithms lend itself very well To sensationalism and clickbait. So I don't think if you're wanting to stick with quality journalism, that is the correct channel for you to focus on. And what's currently on the screen, direct brand visits, that's something not even Google can take away from you. And I've said this many times before, and I'll probably keep repeating it until I'm blue in the face. If you're a publisher who doesn't do anything different, that doesn't do anything unique, if you're just another reporter of the same news that everybody else is also reporting, regardless of your niche or your industry. You're probably going to disappear in the next 5 years. You need to have something unique, something that makes you different, something that makes you better perhaps than your competitors, and something that, that your audience is wanting to actively seek out, actively go directly to your website to get that type of news, which only you can provide to them. And if you don't have that USP, why are you in business at all? This writing has been on the wall for a very long time, by the way. I just think that The introduction of AI inserts is accelerating the process. And I actually do believe that we're going to go through a fairly rough transition period in the next 3 to 5 years, but we'll probably end up with hopefully a healthier publishing ecosystem at the end of it, where the chase after clicks is disincentivized, that it's just not worth chasing after clicks. And therefore publishers are forced to focus on what makes them different, what makes them better, what makes them unique. And we'll end up with fewer publishers, but hopefully healthier publishers with engaged, loyal audiences and are willing to interact and even pay for their news in different ways, perhaps through paywalls, micropayments, whatever it ends up being, so that they will actively support the publications that they want to support. The biggest challenge I still think publishers will have is feeding that particular mechanism with new traffic, new subscriptions all the time. Because currently publishers, even the hardest paywall publishers like the Financial Times, for example, still rely on a constant source of new traffic to keep new subscriptions coming and, and supply the churn, the drop-off of expired subscriptions. And that traffic channel is going to mostly disappear. So we'll need to find other ways to keep feeding that machine and make sure that you have a powerful publishing brand that even new readers are willing to try out and engage to become loyal readers.

If you want to highlight potentially a couple of things to move with the next topic for publishers or SEOs working with publishers.

The biggest thing a publisher can do is be honest with themselves and understand what makes them unique, what makes them different, what makes them better, and try to nail that down. And that's always a bit risky because I don't think a lot of publishers have enough self-reflection to do that. And they might need to rely on external sources. They might need to rely on independent external people, maybe do some customer research, some user research to really ask, why do you visit our website? What do you want from us? And I think that's a necessary step for them to take the next step and try to build that, that brand loyalty and that, that long-term engagement that makes them more or less immune from, from what Google and the other technology companies are up to. So I think that's the number one most important step that a publisher needs to understand. What makes them them, and then try and find a way to capitalize on that and build a brand around that.

That's a really, really good answer. I'm not going to name, but yeah, it's like, it comes to my mind, like one that started in the financial or business-related sector, and then they were writing about shoes and gadgets, right? So specialized publishers ended up diversifying and became generalists. So I can definitely see how it might be. A little bit trickier, but yes, this is about brand positioning at the end of the day and establish that very well and develop on that.

Classic marketing, Aleyda. It's classic marketing. And a lot of publishers don't like marketing as a word. They think it's a dirty word, but they're going to have to learn to love it.

100%. To move on with the next topic, I think that it connects very well with what we have just talked and is about optimizing content for AI mode and LLMs in general. Pieces covering that were actually featured in the last edition of the SEO FOMO newsletter. In one case is a post from Gianluca Fiorelli. In the others is this post from Jan Willem Boving in LinkedIn. And I love them because they are very actionable, explained in easy-to-follow steps. I have seen that a lot of pieces have talk about the importance of chunks in this case, but it's this very high-level concept that is difficult to grasp. So I love how specific and actionable the pieces from Jan Willem and Gianluca are here. And Gianluca also with a very specific example for this search, why this chunk is highlighted and this other is not. And he starts analyzing it. It's like, oh, how it is organized, how it is structured, how easy is for the AI system to extract information from one versus the others and use it, especially you who works with people who have dedicated their life to write content, how, what you would like to highlight from this?

I liked both of those pieces a lot. And one of the reasons I liked them both is that they're very light on technical jargon. They explain it very clearly and very directly in ways that pretty much everybody can understand. There's always been something that stuck with me that someone once said, I don't know exactly who it was, but it basically amounts to, if you cannot explain it clearly, you don't understand it well yourself. And I always find it that people use a lot of heavy-handed jargon and information retrieval terminology, are maybe not aware of the fact that their audience doesn't necessarily understand.

Well, good luck explaining vector embeddings to a copywriter or a journalist or content marketer. You know, at the end of the day, our work as an SEO, and I'm sure that you feel this heavily also, especially working with enterprise-level companies, the more we work in this field, our work relies more on that good communication, good alignment, actionability, and this is it. So everybody in the team is able to understand it.

Exactly. I mean, you and I may, when we hear vector embeddings, you think, yeah, multidimensional vector space, or every word has a vector and the closer they're aligned, the more similar the meanings are, those sorts of things. You can't explain that to a journalist and not expect them to glaze over and zone out. Also, don't ever tell a journalist how to write content. They know how to write content. They really do. The only thing that you need to do when you talk to a journalist, when it comes to explaining things like this, is tell them to maybe emphasize things a little bit different. Like you use subheadings, big to the top a little bit, but don't ever tell them how to do that job, because that's just going to get them into an antagonistic mode. And also, we're not journalists. I'm not. I come from a technology background. So I have no right to tell a journalist how to do that job. And I think if you come at it with, with that perspective, as in just, it's not about telling someone how to do that job. But telling them how, what little things they can improve to maximize the chances of achieving whatever goal they want to achieve, be that a top stories ranking, a Discover ranking, or a citation in a generative AI answer, you have much higher chances of success. And that's what I like about both of those posts from both Jan Willem and from Gianluca, that they just, they strip away all that lingo and that technical language. And they just say, well, if you do this, you're going to have a better chance of the right outcomes. And that's exactly the kind of communication we need to have to be taken seriously in these conversations with these companies when they ask us, so how do we optimize for generative AI?

100%. So hopefully this will give more references and to facilitate all this. In fact, I have created a new section in learningseo.io, which is dedicated to optimize for AI search to pretty much aggregate on one hand tips, frequently asked questions. And then resources that have been already published about how to optimize for AI search across different areas, subsections. All of these have been chosen directly by me. So I have only really tried to feature here information that is reliable, that come from reliable sources, actionable resources, also an aggregation of all of the tools in the market that have some sort of AI-related optimization functionality so we can start assessing them. So hopefully this also facilitates it all. I want to highlight a couple of other pieces that got a lot of engagement. I believe it's because of all of the interest regarding AI mode, but what about ChatGPT? What is the consistency between the visibility and the mentions and the sources that are taken into account between the different LLMs? And yes, in the last study from Patrick Stocks from Ahrefs, it showed that they are not shared across all of them. Only 14% of the top 50 mentioned sources were shared across all of them. So I know that people get nervous with this because it's extra work. They need to find out more what works or gets cited in one versus the others. I understand that, but we have been also complaining during Yours about Google's monopolies and how all of the eggs were in a single bag. And at the end of the day, it's about thinking what is the platform that is more used by our audiences and allocate our efforts to identify the sources, the patterns accordingly, and focus on that.

I like the research. I like how companies like Ahrefs are doing this sort of research and sharing this data freely and then helping bring the industry forward into this AI era. And it's interesting you mentioned, you know, such a low percentage of traffic. You're talking 1% of traffic to most websites coming from the AI bot. And it can be very tempting to then just not optimize for it. And to a certain extent, I agree with that because you're not really optimizing for traffic anymore if traffic is your key differentiator. But as we just talked about, traffic is going to decline anyway. And the stuff that people like Rand Fishkin and Amanda Nadi Vidat have been saying About zero-click marketing, it's very clear here that you're not optimizing for the traffic. You're not optimizing for the click. You're optimizing for the visibility as part of the customer journey through this. Um, and that point's really important to understand how different AI systems generate these answers inside their sources and inside their brands as well. Um, I think it's very clear that some systems rely very heavily on user-generated content and classic authority signals. Mm-hmm. Others cast a bit of a broader net and look at different types of sources or maybe rewarding volume more than quality signal. But we haven't really nailed it yet. It's a lot of second-guessing and reverse engineering, which is why I think it's important to keep doing this research as well. And of course, the AI tools themselves are constantly evolving with newer models, different systems, different guardrail prompts built in to generate certain types of results, but we haven't really found mechanism yet other than, you know, be a powerful brand, but not everybody can be a powerful brand. Not everybody is going to be the next Wikipedia or YouTube or Reddit. So, you know, you're going to have to find a way to build a certain niche, carve out a certain niche in whatever it is that you're offering and make sure that you're ticking the right boxes for whatever AI assistant you want to optimize for.

Let's see what happens in the future, but we need more data. We need more insights in order to be able to focus the efforts accordingly, 100%. To start wrapping up, last resource here. It took my attention because I have covered many times excellent guides, especially there are quite a few from Dana Di Tommaso regarding how to track AI and LLM traffic. Even if we have highlighted how is this not necessarily the ultimate in this case goal, but one of the many along with visibility, but there's definitely interest on this and there's this free Looker Studio template to analyze traffic from AI chat with the integration of GA4. So here's an additional resource that you can take into account with a step-by-step that hopefully enables more visibility and understanding on the engagement that LLMs generate to websites.

I love it when people do stuff like that and give it away for free. The generosity and willingness to share information in our industry is something that I really, really appreciate. And I hope we will keep doing that. And I've checked out this Looker Studio template myself as well. It's really good. It shows you exactly which pages are being mentioned or being clicked through to from ChatGPT and other sources. And I think that can help inform certain parts of your content strategy to see which types of pages are driving the links and the clicks in ChatGPT and other chatbots to see if you can replicate similar structures in other web pages.

Yeah, that is something that I appreciated in particular too. To end with today's edition, I want to thank again to Similarweb, who's our today's sponsor. In fact, I would like to highlight a couple of new reports that they have about AI traffic and AI brand visibility too for you to check out. I believe that there are many features out there, but they are able to provide with a very impressive level of accuracy, not only the information of AI traffic, but I believe that they are Probably one of the very few right now in the market able to showcase the top prompts per page. And if you don't want to wait until the next week for the latest news, remember that you always have the SEO FOMO News Aggregator where the community curates the latest information with submissions and comments about the latest news. You can go and take a look, submit your own resources, guides, the top ones are also included in the SEO Pharma Newsletter. Thank you very much again, Barry. I really appreciated your inputs and insights and very valuable tips today. Thank you. Thank you very much. Any last thought that you would like to finish with?

Thank you for having me. I really enjoyed this conversation and talking about this stuff. Like you, I'm passionate about this industry and about our constant evolution. And I'm very excited to see what the next chapter of SEO is going to ring, I would say it's still SEO. I think AI SEO is probably the right terminology because we have always optimized for different verticals, news SEO, image SEO, video SEO. So why do we need to invent a new acronym for AI SEO? We really don't. We just need to adapt our tactics, evolve as the technology evolves. But in the end, the foundational skill sets of SEO remains and is still underpinning everything that we do for generative AI.

I love, I love that take. I think that potentially, and that is something that I mentioned the other day, the only reason why it is reasonable to rebrand will be just for marketing purposes, just to package what we now do as yet another thing and to sell it differently. But yes, besides that, 100% agree on that. So yes. Well, thank you. Thank you again for watching. If you liked today's episode, remember to like the video, subscribe to the channel, and until the next week, bye-bye. Bye-bye.