The AI reality: Navigating strategic shifts in SEO

Barry:

I mean, it’s the same thing. I mean, traditional SEO is required, for the most part, to get Google to cite your articles in Google Search as a blue link; in Google AI Mode as a link that nobody will click on; and in Google AI Overviews, another link that nobody will click on. Without that, Google will not find that content.

Thomas:

Go through, and we look at it, and it’s like—just automatically changed all of their SEO title tags, and they’ve, like, experienced a big drop-off for a whole bunch of pages, and I’m unsure why.

Janna:

I still don’t think it’s quite there to do the analyzing step for me, but at least then it can scrape competitors, come up with underlying suggestions about new topics that they’re writing about. Doing that by hand would take weeks, days, multiple people, and it takes forever to find the right adjudication, anchor text—all those kinds of things.

Halyna:

Hello, hello, everyone! We have already three experts that we are having in our panel today—Barry Schwartz, Thomas Phillips, and Janna Pugh—here to discuss the AI reality. Hey, guys, could you please introduce yourselves briefly, and then we will move forward with the questions?

Thomas:

Absolutely. Do you want me to start? I’ll start.

Halyna:

Great!

Thomas:

My name’s Thomas Phillips, co-founder at DTC SEO Agency. We’ve been doing SEO for 18 or so years now at this point, primarily on our own brands and now with our clients as well. Obviously, AI is becoming such a big thing lately—a huge thing that we’re diving in on—so I’m super interested to hear and understand more from the two other panelists on here as well and learn as much as possible.

Janna:

Hi, everybody. My name is Janna Pugh. I have been an SEO for about six years now, so not quite as senior as my other panelists here, but I have been able to do quite a lot in my career so far. Right now I work full-time for a tech company in the Bay Area, and this panel has been—like, all day today—this summit has been really awesome, really great information, and I’m really excited to dive in.

Barry:

And I’m Barry Schwartz, pretty well known for creating a lot of content that people don’t like to read but they have to. Yeah, I’ve written over 40,000 pieces of content—some call it spam, spammy content—over the past 20-plus years.

Halyna:

Great! Thank you. Thank you very much, guys!

Let’s move forward to, actually, the questions that we want to cover today and discuss.

So, the first one is: How are generative-AI tools like OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude changing day-to-day SEO tasks? For instance, are you using them for keyword research, content ideation, metadata generation, or SERP analysis? And what benefits or pitfalls have you observed so far in these use cases?

Thomas:

Oof, who’s going first?

Janna:

You did—you talked first.

Thomas:

Alright, I’ll go first.

How are we using it? I mean, we played about with it from a content point of view, but we’re still very much on the side of, like, I didn’t want to, I don’t want to risk it, and our writers and, like, expertise seem to have a lot more in-depth knowledge than what, you know, ChatGPT, for example, can come up with.

Now I’m sure there’s other people doing much more in-depth writing than we might be doing with ChatGPT and things like that. But, right now, like, we’re still very much on the edge of seeing what is going to happen, so to speak.

In terms of meta descriptions and meta titles, again, from my experience with it, I just can’t trust it. It seems to, like, you know, hallucinate things and just not be anywhere near as accurate as, like, someone who has expertise to really deep-dive in and do the keyword research first and actually build out something that is going to rank, based off of experience, you know, with things. But I’m very interested to hear how other people are using it as well.

Janna:

Yeah, I would say for me, primarily what we’ve been doing, my company is working mostly on the background work, like the data collection. For me, I’ve been using it for a lot of competitor research and analysis, the things that used to take a lot of manual hours to scrape, and then you would have to go through the process of analyzing.

So I still don’t think it’s quite there to do the analyzing step for me, but at least then it can scrape competitors, come up with underlying suggestions about new topics that they’re writing about, if they’ve updated their navigations or processes or things that we need to also jump in on the line with, as well as just coming up with, again, some of the boring manual collection of interlinking strategy building. I have a couple of, like, NotebookLM—more closed circuits—but able to program them specifically to get a little bit more interlinking opportunities, whereas doing that by hand would take weeks, days, multiple people, and it takes forever to find the right adjudication, anchor text, all those kinds of things. So that’s primarily what we’re using it for.

Barry:

Yeah, and just to be clear, I’m not an SEO, so I don’t really do SEO tests, but I write a lot about SEO, so I talk the talk, but I don’t walk the walk. But that being said, if you run a website, you pretty much do SEO. And I agree with pretty much what everybody said here. I use Midjourney to create a lot of images for my website because there’s nothing better than creating a robot eating paper to convey Googlebot indexing stuff. It’s just you can’t do that—you can’t find that in stock imagery—so it’s really imaginative and stuff like that. And it’s clearly AI, so I don’t really need to, like, pretend it’s not.

I tried using it for my content here and there, and it just doesn’t read like the way I write. So I get to the point; ChatGPT and AI do not get to the point. They really just give you a lot of fluff. You said it hallucinates. Just recently I’ve seen people, like, talk about how AI Mode is, like, is it, you know, 2025? And, like, it’s no, it’s not 2025—and then it is 2025. And Danny Goodwin wrote an article on Search Engine Land: Is it Tuesday? And, like, no, it’s not Tuesday. It’s Wednesday—well, it’s really Thursday. So it’s kind of funny, and you gotta be careful when you use it.

Like you all said, when you read a piece of content, you could quickly see if it’s written by AI, especially if you write a lot. You know, sometimes your writers have personalities that the users expect, and you should really think about, like, the whole Google thing that makes people want to vomit. But yes, you want to write for your users. And if people read your content and you’re, like, this doesn’t sound like the writer—of course, it could be, like, some ghost writer, so you don’t know.

But if people read my content and say, wait, this is not written by Barry—this is looking like somebody else wrote it—and that would be an issue. So I do use Midjourney for image creation. I don’t use it to write my content. I have used it in specific cases for figuring out how to create a good intro for some of my YouTubey videos, the ones that get me the play-button stuff—those ones that are really YouTubey. But, outside of that, I try not to use it much outside of all that.

Halyna:

It’s interesting how we all speak about not being fully able to trust the AI engines, but still it’s there, and everybody is trying to leverage it and put it in the workflows. So the trust is on this level.

Barry:

Yeah, just one more point. I make my money doing software development. So I have a whole team of software developers that work on pretty much with me in New York.

And they’ve been trying to use it for software development, and it’s like they post examples left and right, saying, “If I would have listened to this AI for this purpose, it would have deleted the whole entire website or would have done this drastic thing and would have broken the software, or, like, 4 billion emails would have been sent out with credit card information.”

It’s just wild what this stuff could do. You have to validate it. They could be useful tools. It just helps with documentation and so forth, but you got to be very careful when you use this stuff, and it could save you time, but at the same time, it could cost you a lot of time and frustration.

Halyna:

Definitely. Do you think it is a temporary thing, or will it start developing and something will change there, just hypothetically? Or do you think it’s still something that we have to deal with?

Barry:

It’s definitely not. This is the future. AI is changing everything. It’s not going away.

It will just get better, and eventually it will replace me and all of us. We’re all doomed—no, it’s the future.

It might not replace us, but our jobs will change, just like when the railroads changed things and horses went away and, you know, things change. You don’t have Xerox machines anymore; you don’t have fax machines. Things change, and will adapt, but it’s not going away. I don’t think it’s not like, you know, some other things we had that are fad. This is not going away.

Halyna:

But while we are here, we can still continue our discussion, not replaced by AI.

So, Barry, you’ve noted that it ultimately matters whether content helps the user, not how it is produced; and Thomas cautions that AI-generated content might seem like a shortcut, but it’s a fast track to lower rankings. So, in light of these, how should SEO professionals balance the efficiency of AI-generated content with the need to maintain quality, uniqueness, E-E-A-T, so that content doesn’t, like, tank your site despite being created faster?

Barry:

Again, I don’t personally use AI for my content, so I really can’t comment. I mean, we’ve seen some examples of websites over the past couple years using AI to basically cheat Google, and then, a few months later, those websites completely tanked in the search results. You could always trick Google in the short term; it might be a few months, it might be even a couple years, but ultimately, if you’re trying to cheat the search engine, you’re ultimately cheating the user as well.

When you think about it, it doesn’t even matter. I mean, with AI Mode and AI Overviews, you’re not getting clicks anyway, even if you rank well, so I don’t know if it makes a difference. In the long run, taking shortcuts is never the good way to do it. People say, you know, “work smarter, not harder.” I mean, I don’t know—either you’re smarter or not; taking shortcuts that you think are smart isn’t always the best way to get there. I could be wrong, but I don’t know.

Thomas:

Yeah, it’s very interesting. I was just thinking, as Barry was talking there, how many clients come to us and they’re like, “Oh, yeah, we just created our whole blog with AI—just straight out of ChatGPT,” and we’re like, “Okay, we’re going to have to—” you know, we turned down clients because of that exact fact. I’m like, “All right, I don’t want to risk it,” and, if your mindset is around the quick, take-a-shortcut way of doing things, then SEO is definitely not the play for you.

Janna:

Yeah, I would say I’m all for using it for an outline or blog ideas—even a content calendar—but I am a big proponent of only having it do something that you could do yourself; it just takes the time for you. I can create the copy myself, and I think users are—we’re all pretty smart in general. We’re all a little dumb in some things, but we’re very smart; we can all tell when copy is AI-generated. There’s, like, nine pieces of copy that all kind of sound the same and they’re all pushed within the same month. Like Barry was saying, they can do really well because they contain that fuzzy, fluffy content that bots and LLMs are looking for, but, at the end of the day, people don’t really want to read that—especially if there’s no authentic copy. If it was just created from AI, then humans could get that somewhere else.

So I think it should go back on the companies—the employers—to actually write the words that go on the pages, but I’m all for having it expedite the actual topics and outlines.

Probably continuing the topic of automation: Which parts of SEO do you see being successfully automated by AI right now? What works for you, and where do you draw the line between AI-driven automations working well and introducing some errors and disruptions in the workflow?

Thomas:

Yeah, I think, as I mentioned at the start, I just can’t trust it. Aside from helping me reformat an email slightly, that’s about as far as it goes in terms of where we’re using things internally. Actually, yesterday, I did use ChatGPT to help categorize a hundred products that a client sent over, in terms of, “What do you think the collections or categories should be for this?” and that is literally as far as it goes in terms of where I draw the line in terms of trust from an SEO point of view.

Janna:

Yeah, my issue is when it comes to the user-facing copy—so the meta titles, meta descriptions, titles, body copy, the things that humans would read. That is where I would have the most issue with AI-generated copy.

When it comes to the back of house, I think that’s where the expediting and the streamlined processes can come in, because then you still have to go through your QA and deployment approvals, sending to the client, and getting your schedules out on time, but it’ll speed up the time on your part, getting the infrastructure laid out and getting these processes.

If you were doing a technical audit in Screaming Frog, you could sort the issues by priority and have it automate a couple of those mid-line steps. Now, I would still be the one doing the QA and deployment—I wouldn’t automate; I think it’s a little misnomer. I don’t know if I really trust it to fully automate anything, but I think it could definitely streamline and expedite a lot of things.

Barry:

And the previous speakers went through—Dale had examples, Judith as well—so the previous speakers on this panel, on the webinar earlier, went through a lot of good examples there. The best place of automating SEO is where Google just replaces us all and shows AI Overviews, and Google is going to change the search results with AI to, again, use our content and not send us the traffic. That’s the biggest challenge for SEO: we’re not going to get the traffic, and that’s a big concern for people. I don’t see any way around it; I know Sundar Pichai was just interviewed by The Verge at Google I/O.

He really seems not to, at all, convey to the users—to us—that there is a real issue here. It doesn’t seem like Google cares. I mean, they care about the searchers, and that’s the most important thing to them, at the cost of the publisher, and there’s no way around it.

So, I mean, the way that AI is going to help SEOs is by basically having them move into new careers—I’m sure it’s a bit too much.

No, I mean, there’s still going to be—the only way for the AI answers to get the answers is for SEOs to make sure that content is indexed. Google has confirmed that AI Mode and AI Overviews need to be from pages that are indexed by Google. Of course, just because you’re indexed doesn’t mean you’re gonna click, and we’ve seen that over the past few years with Rand Fishkin’s zero-click stuff, and now it’s just gonna get all worse. So we have to figure out ways to generate clicks from these new AI experiences.

And that’s gonna change over the years; we have to be on top of it. I don’t know if AI is gonna really help us with that, but we’ll see—we’ll see what type of A/B testing we could do and how AI could help us improve the clue-through rates from these AI experiences.

Halyna:

Yeah, definitely. Well, speaking about Google’s AI Overviews—which are now roughly on 30% of US search queries—what strategic adjustment do you think SEOs should make? And do traditional tactics for rankings still apply, or we’re definitely losing this part of traffic with AI results being on top of search?

Barry:

I mean, it’s the same thing. I mean, traditional SEO is required, for the most part, to get Google to cite your articles in Google Search as a blue link; in Google AI Mode as a link that nobody will click on; and in Google AI Overviews, another link that nobody will click on. Without that, Google will not find that content. So traditional SEO—in terms of technical SEO, building content, link-building, and all that type of stuff, everything that Google uses to figure out what content to use in their AI experiences—is traditional SEO.

How you could go ahead and get Google to actually send you traffic is a whole other story. And with these AI experiences—AI Mode, AI Overviews—we’ve seen Google test different ways of linking within blue links, dotted lines, changing the thumbnails, the favicons. That’s not really in our control.

All we could do as SEOs is create content that is found in these experiences and hope that Google gives publishers something, or website owners something, that will drive traffic to their websites. There’s nothing much we could do in terms of—maybe you guys disagree with me—but I don’t think there’s anything we could do to improve traffic, because we’ve seen that click-through rates—I think SE Ranking have a study on that—click-through rates have drastically dropped, significantly, from these experiences. And hopefully Google—I mean, I’ve spoken to the highest, highest levels of Google about these click-through studies, and they’re like, “What studies?” I’m like, “Kidding me? What studies? There’s, like, several studies on this already.” “Oh, we have our own internal data.”

So share that own internal data. They won’t share the data with us. “Oh, well, we are tracking it.” I’m like, “I don’t see it,” and we don’t see it in Google Search Console. “You could give it to us.” “No, we can’t do that; it’s not good for you.” “What you mean it’s not good for us?” “No, well, it’s not useful for you.” I’m like, “Show us the data in an aggregate or on our own level, or just stop saying we’re getting more diverse, happy clicks that are engaging and driving conversions.” We’re not seeing that from these experiences. So I’ve been very loud; I think the community’s been very, very loud about this. And hopefully Google will take it to heart. But, so far, in the past, you know, two years since Google first launched SGE (Search Generative Experience) and then converted to AI Overviews and then AI Mode, we haven’t seen Google act too much on that, outside of limited tests that try to improve those links. But again, I’m not sure, so I don’t know if anybody has anything to add on that.

Thomas:

Yeah, I mean, it’s only really going one way out, by the looks of things. But, I mean, this is why we’re really trying to focus a lot more on, like, commercial and transactional-intent keywords where the AI Overview hasn’t necessarily caught up yet, or may not, you know, kind of take over. So that’s really our biggest focus—still creating all of that content that drives the topical-authority side (you know, potentially doesn’t get as many clicks anymore), but I think changing the goal, like, in terms of reframing it to a brand or a client or whatever it might be—“This isn’t necessarily going to get you clicks and traffic.” I mean, it didn’t necessarily convert in the first place to a sale, necessarily. But, yeah, really driving that conversion, transactional-intent keywords is our biggest focus anyway.

Janna:

Yeah, I was going to say, I know that we can take kind of a nihilistic approach to it sometimes—that, you know, AI is here, SEO is dead; you know, we can say all these things, which it’s not true. But I think, in a greater sense, our customers and our traffic are still there; they’re just experiencing the content that we have to offer in a different way than I think we were maybe prepared for or anticipating.

For my site, for example, we’ve been really developing FAQs, because we work in call-center software—it’s kind of a niche industry. And so a lot of our customers that are heavy and ready to convert already know the keywords and buzzwords and topics that they’re looking for, and so they still come to the website. And so our conversion rate is still doing fine because we are setting up the website for a more conversion-oriented process.

But, as a result, our FAQs, our sessions, our impressions are all not doing so hot, because we’re able to see through, like, Ahrefs, SE Ranking, things like that, that we’re getting a massive increase in AI Overviews. Like, from March to May, we, I think, quadrupled the quantity of AI Overviews we were ranking for. So I don’t think it’s a matter of pivoting the strategy, because, like Barry was saying, SEO fundamentals are still going to be in practice. It’s just a matter of changing up the reporting and the metrics that you’re using to set those benchmarks, because the consumers and the traffic are still there—it’s just not coming to your dot-com.

Halyna:

And speaking about dead SEO, let’s continue this topic to the recent rollout of AI Mode and introduced features like Deep Search reports and AI-driven shopping recommendations. What is your take on that? What implications does this have for SEO?

Barry:

So, this Deep Search isn’t live yet, nor are the shopping experiences. They’re very cool.

Deep Search—honestly, Bing came up with it, and they launched it like a year plus ago, and then they changed it to Copilot search with Bing, or in Bing. Anyway, it’s basically, instead of looking at, like, the top five results to give you an AI Mode answer, they’re looking at, like, the top 1,000 results and giving you a much deeper answer with a lot more content. It’s basically a research report with tons and tons of links from tons of different websites—which, again, nobody’s going to click on—but it’s very cool.

And then the whole shopping experiences, that’s where you really kind of, like, “Oh, this is maybe a way that we, as SEOs and website developers, can actually go ahead and get conversions for our customers.” With the AI agent—they call it agentic experiences—if you do a search for “I want to get a flight,” “I want to go on vacation to,” I don’t know, SE Ranking’s in Ukraine, right? So if you want to vacation in Ukraine (which I don’t know if people do that these days—hopefully you guys are safe), but if you want to do it, like, “Is it safe to go to…?” They say, “Yeah, these are the safe places to go. Let me go ahead and book you a ticket and find you a hotel,” and it’ll give you some answers and let you actually check out with Google Checkout to buy those right there.

And if you could integrate your website with these agentic experiences—making agents so that Google would actually integrate directly into your backend (which I don’t know how that’s going to happen exactly; right now they’re testing it—Google doesn’t even know)—that might be the new traffic or conversion that SEOs have to optimize for: building these agents into their websites so that these AI experiences could actually go ahead and convert directly without having to go to the website. So that might be the new conversion, which is super exciting.

Deep Search—I’m sure people don’t want to read. It’s like Google has said that. I don’t get it, really, because everybody’s like—Google’s like, “We want to give these little snapshots, these little clips of the answer,” with Featured Snippets and even AI Overviews. But nobody’s really reading past the first paragraph, especially these days.

So I’m not sure how Deep Search would work. It’s a cool experiment for maybe college people, you know, people who really want to dive deep into a topic. But then again, if you read it, do you believe it? I mean, that’s a whole thing. If you do all this research and go through thousands of links, who’s doing that? I don’t know. But the injective stuff is really cool.

Janna:

Yeah, I was just about to say that generative AI is here—that’s the present; that’s the concern that we have. But I think agentic AI is where the future plans are going to be, just reading through. I mean, very much early stages now, but I think that’s why we have these conversations: to be proactive as to what the future of AI search is going to be. And it’s definitely agentic AI in terms of the one-click conversion opportunity. I mean, that’s what Google has been dreaming of for years.

So they’re on the verge of keeping the user from start to finish on Google in an expedited, streamlined process that will still provide value to us—because we’ll still get the conversion, we’ll still get a user; maybe there’s some sort of educational journey on its way as they convert in the process—but the future is keeping users in the most streamlined, effective process for them, and I think that’s very much where it’s headed.

Thomas:

Awesome. I haven’t really got much else to say on that point other than I’m learning a lot, which is fantastic. So that integrating directly—shopping with Google—very, very interesting. I’m certainly going to start…

Janna:

I think it’s gonna be cool. It’s gonna be kind of rough when it first rolls out. A lot of people, I think, are gonna be scared, surprised, concerned, but I think it’s definitely the future, and I think it will be pretty cool once they roll it out and, you know, all the bugs and things are worked out.

Thomas:

100%. Yeah.

Halyna:

Yeah, and let’s talk about the whole different way that people are now searching for new things—no matter if it is shopping or information.

So, AI is just changing the way we search, but there are also some other platforms—for example, Bing, or the TikTok factor, which is rising in that direction. Should SEOs start optimizing for query types beyond text, like voice, images, videos, just to make them more AI-readable and optimized for this discovery part of the journey?

Thomas:

100%. I think, you know, if you can repurpose a bit of content that you’re doing into multiple different ways of using it—you know, like I’m sure you guys have seen, like, podcasts, and then they transcribe the whole thing, and now, all of a sudden, the whole thing’s indexable, or, you know, uploading a video with the transcript so Google can actually go through and crawl it.

I think it’s all extremely important and, you know, great ways of brands using one piece of content and using it across multiple different areas across the site, and potentially getting more clicks and more traffic coming across.

Barry:

Yeah, I do agree with that. Yeah, and also the whole agentic stuff—you know, we should be looking at that as well, like we discussed before. But, yeah, it was funny because when Sundar Puchai did an interview with The Verge, the head of The Verge, the Editor-in-chief, is like, “Honestly, if I started The Verge today, I wouldn’t probably make a website; I’d probably, you know, make a TikTok account or Instagram account and go that route.”

So people are thinking in a website-last approach as opposed to, you know, a website-first approach; and then repurposing your website is maybe—it’s like repurposing your other content to put on your website. In fact, when I just made a video recently that’s getting a lot of views, I’m like, “You know what, maybe I’ll make a piece of content on that video after I made the video.”

So it’s really, sometimes I’m thinking the other way around when it comes to building content, because, like we said, clicks are—you know—traffic to publishers is way down compared to what it was, you know, a couple of years ago, especially five years ago. So I think it’s thinking about it the other way around: What—where should I put my content first? And then how else should I put that content based on that?

Janna:

Yeah, I think it’s really cool that this is—I know it’s scary in a lot of ways, but I think it’s really cool that, for the first time in many years, if ever, we’re finally in a multimodal, omni-channel industry where Google’s organic search isn’t the only way that users can experience your brand in a real fashion. Like, Bing can actually be a presence that we can optimize for and pay attention to now because of the, you know, the change-ups that have been made in recent months and years.

I think it’s a really awesome way to add videos, add images, to have a specific content calendar where all of your media across all your plans and social media—they’re all pushing on, you know, this one specific thing for a single month. So you can see how universal and omni-channel your brand is for a specific thing. You build up your experience and your semantic understanding of this one specific area, but you do it in a variety of different ways.

If you have a video that you’re pushing on YouTube, you can also splice and put it up on Facebook and TikTok; or, if you’re a local business and you have a couple of products that you’re launching, you throw it up on Google Business Profile, you throw it up on all the other channels—not just on your website anymore.

I think it’s going to take more time for people to figure out which new avenues are best for their specific brand, for their industry, but, at the same time, I think it’s a really cool opportunity to be less domain-focused.

Halyna:

We have several questions in the chat from our guests, so I guess we can shift to that discussion, because those are really good questions too.

First one is: In your view, what’s the biggest AI-driven shift in SEO that pros need to adapt to right now, apart from everything that we’ve already discussed?

Janna:

Whatever is the most time-consuming.

If you’re not using AI and you’re manually doing stuff that takes 15 to 20 hours, that needs to be automated. I think spending your brain-power is now used for strategic connections, analysis, deployment.

So, if you’re still manually pulling and you’re using your VLOOKUPs, I think there’s more efficient ways to spend your time—not necessarily in, like, creation, but more so in efficiency.

Barry:

Yeah, if you look at what Reddit did recently with their translated pages, they even touted it: they’re getting so much traffic from other Google Search properties outside of the U.S.—like other specific regions—to their translated, AI-translated pages.

They were so excited about it that they touted on their earnings call how Google even helped them use Gemini to go ahead and create these AI-translated pages with their Google stamp of approval—whatever you want to call it—to make that happen, which is sending a tremendous amount of traffic to Reddit.

You can look for opportunities like that. That sounds very scary to me. The whole Reddit thing is very scary. I mean, one day—you see this with Google all the time—Google ranked Yahoo Answers very well in the early days and ODP, and then we had Wikipedia ranking very well in Google. Now it’s Reddit’s turn, and then, in a matter of a second, Google changes their algorithm and Reddit traffic goes away like that. It’s a big concern for Reddit, and I’m surprised people invest in that stock.

So, I’m not a stock— I don’t know. I mean, I’d invest in it because I don’t, I guess, know how to invest money, but it just seems scary to me that people are using AI for shortcuts. Even though they’re useful, and Google’s distinction in their documentation around creating content is “You could use AI to help you with content and so forth.” It’s not about if it’s AI or humans or whatever, as long as it’s providing value and it’s helpful and so forth, and that’s a very subjective thing that an algorithm has to figure out.

You could be—you can get in trouble, even if you think… A lot of people think their content is great. A lot of people think what they’re creating is for users, and search engines might have a different approach or feeling about what you’re creating versus what you think you’re creating.

So, look for ways you can utilize this technology to make something that’s really valuable to users and maybe have a third party look at that—someone not part of your website—to say, “Is this really something that’s valuable or not? Are they getting a value-add, or do they think this is kind of garbage?” If it’s something that most people think is actually useful, and you can use AI for that—great, do it.

I mean, we’ve seen amazing movies coming out of, like, Veo 3, Google’s latest AI-generation tool for videos. It’s absolutely amazing; it’s getting tons of attention on social media. You create a cool video using Veo for your website, you put it on your website—it could generate a tremendous amount of brand recognition, links to your website, and so forth. That’s a value-add.

So, just do it in a smart way and get a third party that doesn’t have any connection to your website to really vouch and say, “Yeah, that’s actually something that I find valuable,” as opposed to, “This junk that I’m creating that I think is useful.”

Thomas:

I mean, it really depends on, I guess, what angle you’re coming in with that question.

I instantly think, from a client point of view, the biggest thing that we’re seeing—as I said—is a bunch of clients coming in with AI-written content across the board. Something that SEOs really need to think about is vetting the client site beforehand: How much of it is actually just ChatGPT garbage, and how much of it is from the actual in-depth expert who built their product in the first place and actually wrote about it on their product pages?

So, that’s the way I think about that question specifically, but—yeah, hope that’s helpful.

Halyna:

Great, thank you, guys.

Then, moving forward with the next one—also interesting and a bit painful, hard to think and hard to talk about—but with AI automating tasks, how can SEOs stay essential? Barry has already started talking on that, completely removing the SEO professionals from the landscape, but maybe, Barry, it’s not that pessimistic for us.

Barry:

Yeah, maybe not. Again, these AI engines need content to be indexed. If you want to show up there, think about how maybe that content is showing the AI answers with your brand there—like, somebody reads, “Oh, this brand keeps coming up for the same types of queries.” People will trust that brand over time and maybe go to you.

That’s a really hard thing to do, but SEOs need to think more about branding and the engine’s experience to become relevant. SEOs have changed many, many times in the past 20-plus years—from the days of, you know, Excite, AltaVista, Ask Jeeves, to Google overtaking everybody—and now we’re in this new phase with AI search engines.

So, I think the best SEOs have survived over the past 20 years, and you just gotta keep trying these experiences, seeing what you could do to give your clients some type of edge, and experiment, test—experiment, test—and keep doing that over and over again.

Janna:

Yeah, and to Judith’s point earlier today, I know some people can say that, you know, “AI can do everything,” and I think, on paper, maybe that’s true. But when it comes to the logical reasoning, the conclusions, the fact-finding, the reporting, the analysis—that all needs the human element.

So, there are a lot of things that can be done with AI. Not everything should be done with AI, but it still takes you, with your expertise and your understanding, to make the connections between the analysis and the implementation dots. SEOers, as humans—as employers, as people working these jobs in this industry—will still be around for a long time. It’s just a matter of working with the AI and not necessarily against it.

Thomas:

Yeah, I think that’s a really good point: working with it, not against it. There’s a tool that we keep seeing popping up—I think it’s called Search Atlas or SEO Atlas, something like that. A bunch of clients coming in using it already, and then we go through and we look at it. It’s, like, just automatically changed all of their SEO title tags, and they’ve experienced a big drop-off for a whole bunch of pages and are unsure why.

I’m like, well, you had a pretty optimized title tag before, and now it’s just pulled in whatever other ChatGPT garbage you had on your page beforehand, and now it’s thought that it should be about something different.

So, I guess, in terms of automating tasks, like, 100%, we’re still essential when it comes to that—as opposed to, you know, just trusting everything and clicking all the nice green buttons and it all looks good. You know, the tool says it’s optimized—and, yeah, it might be, but I’m pessimistic still, you can tell.

Halyna:

Okay, thank you! Thank you for covering this topic.

And probably the last one for today—just because we are short on time—but if there are any AI SEO trends, like content generators or AI search or automated optimizations, that you really are skeptical about?

Thomas:

The whole thing. *laughter*

Janna:

Yeah, kind of a variety of the above. Anything that humans are reading on your copy side, I am always skeptical about, because people say it’s super easy—you know, you train NotebookLM on all of your 15-month recent blogs and it can just make 6 million new blogs for you, and it’s super, super easy. But I think it still needs to be human-generated; it still needs to be reviewed.

My biggest pet peeve is people thinking that AI-written content is somehow new and unique when AI can literally only create what it already can index. By sheer nature of AI-created content, you are not providing anything that’s truly, authentically new to the users and/or LLMs, Google Search, anything above. So I am always skeptical when people have, like, easy solutions about just AI-generated copy.

Barry:

Yeah, on that point, I do think, at one point, AI will be able to generate completely new, unique content. I mean, what do reporters do? They get sources, they get press releases, they put their own spin on it, and they might, like, quote some people. With social media, they could just go find those quotes automatically from people who are reacting to this new announcement; they get the announcement from the source, and they can write this stuff up.

So I don’t see why—that’s my biggest thing. People are saying, “There’s no way; they need our content, and so forth, for them to produce content.” I really do somewhat disagree with that. I’m, like, the only one, I think—or one of the few—that disagree with that, which is a whole other realm.

But, yeah, I don’t know. My biggest, I guess, skeptical area is how Google’s going to handle—Google’s like, “AI content’s fine as long as it’s valuable and useful and so forth.” I have to imagine, at some point, Google’s going to have to be like, “You know, this is where we draw the line,” and be more clear about that. But they’ve been very, very quiet about it over the past several months.

I’ve been pressuring Google to give us more information about this, and they think they’ve given us enough. It’s going to be a big issue for Google. I can’t see how they’re going to continue to index as much as they’re indexing and try to rank as much as they’re ranking. Eventually it’s just going to implode by itself and it’s going to be a big issue, but I don’t know—maybe I’m wrong.

Janna:

No, that’s fair. I respect that.

Thomas:

So we’re actually now tracking revenue that’s coming from ChatGPT, because we’re seeing, for some clients, it’s, like, in the thousands per month. So I definitely don’t think that, like, AI—if you want to call it AI search, whatever you want to call it—things like that are going to slow down.

So many people—I introduced my parents to ChatGPT and my dad’s mind was just blown, you know? So it was just like, okay, I can only see this starting to increase now. So I definitely don’t think AI search is going anywhere, but that’s really where we’re starting to track a lot more and try to get more of a handle on that data and work out how we can appear in those curated lists.

Thankfully, it all comes down to fundamental SEO that we’ve already been doing—getting links on big authoritative websites and talking about the unique selling points of that particular product and brand, and, you know, have ChatGPT link it back and pull that into its curated lists for different things.

So, yeah, it’s very, very exciting. I definitely don’t think that that is going anywhere, but, yeah, content generators and automated optimizations are probably just going to get better and better with things, but I’m still skeptical.

Halyna:

Yeah, thank you. Thank you for sharing and being honest about all of these trends and all of these changes.

And probably the bottom line would be: AI is not going anywhere, but SEO experts are not going anywhere too—at least for now. So we have to just make sure that AI is helping us be more efficient with some tasks and some things that we can leverage it for.

So, thanks a lot for this discussion. That was really great!