Actually, is SEO dead? An interview AI search expert, Meg Clarke

AI search: What is it? How do we thrive in this paradigm shift in visibility? Should we throw old-school SEO playbooks out the window? Should we be scared or overjoyed?

I asked AI visibility expert Meg Clarke to be my inaugural interview for the revival of my interview series. Frankly, because… I had a ton of questions for her myself.

Currently, there is a dearth of data about how AI search actually works (and certainly no tools helping you “optimize for AI” the way there are for Google)… because it’s all so new and changing fast.

That’s why hearing from people in the trenches is virtually the only way to get information and adapt. People like Meg, who came from the world of traditional SEO and was one of the earliest adopters, experimenters, and documenters of AI search.

Tldr; you should read this interview.

Krista: How does AI, as far as we know, actually recommend companies to people using those tools?

Meg: Great question. I think it’s best described if we start out by comparing it to Google. So, Google’s job was to find the best answer for a search query, and Google determined what the best answer was based on 200 different factors, but the majority of those factors were around engagement.

So, from Google’s perspective, if somebody searched and they clicked on your website and they did one of three things: spent longer than 10 seconds, clicked on the link, or scrolled, Google assumed that was a good answer to the query and they would put that website higher and higher on the search results page.

Now, if we move over to AI, AI is very similar to SEO. Like, dude, Google’s been doing this for decades, they know a thing or two about what a good website is. So AI certainly looks at engagement… but instead of it being 70 to 80% of what they use to determine the best website, they probably weigh engagement only 20 to 30%.

The rest of the things that AI looks at and recommends is third party mentions and how those align with each other. So are other websites talking about you, have you been on podcasts, YouTube, do you have a LinkedIn presence?

So what AI is doing is looking at what you say on your website and how you show up around the internet. So, in order to get recommended by AI, you need to show that you are an expert and you need to show up multiple times around the internet.

Krista: As a consumer, I think AI search is world’s better than search-engine-searching ever was.

For example, I was looking for a new therapist this summer. I used GPT because Googling (or even searching a directory) in Los Angeles felt so overwhelming; there are a million therapists. And with Google, I had to reduce everything into something like ‘therapist for anxiety’, which didn’t reflect what I was actually looking for.

With ChatGPT, I could ask:

‘I’m looking for a therapist who will relate to me like this, I’m dealing with these things, I prefer this communication style…’

It let me search with nuance, which felt so much more aligned with how humans actually think when they’re seeking help.

Meg: Yeah. And even for the business (or therapist in this case), it’s very difficult with Google.

I work with a lot of therapists, and I have a client in Manhattan, and they started to rank #1 for anxiety therapist, and even with that #1 ranking, in Google search results, they’re at the bottom of the page because there’s so many ads and questions asked, and AI overviews and YouTube videos. Even if you get a good organic result, it’s at the bottom. It’s very difficult to navigate through for users and not that great for businesses, either.

AI search is actually just better than Google, for everyone.

Krista: In traditional SEO, the concept of high-quality backlinks mattered a lot. Is it the same thing with AI, or is there some other nuance?

Meg: Well, there is a little bit more of a nuance to it. With Google, they really wanted that actual link. They wanted New York Times to link to your website, and if New York Times linked to your website, then you’re going to rise because New York Times has a lot of authority.

With AI, you don’t even need the actual link. You just need to be mentioned.
AI is looking for mentions all around a page, and the mention doesn’t necessarily have to be “linked” back to your website per se. You can just be mentioned.

So it is similar, but it is different. And the other nuance about this is that with Google backlinks, the thing linking to you needed to be another website, and ideally, a high authority site like New York Times.

But with AI, those mentions can even come from platforms that you own, and I mean like a YouTube channel, LinkedIn, Instagram, a Substack that you write. You just need to show up in multiple places around the internet being consistent and authentic.

So if you say on your website you’re this person, then if you show up on LinkedIn and Instagram and YouTube as that person, that is going to provide that authority and expertise AI is looking for to recommend you.

Krista: So are you saying that LLMs are basically reading what people say about you and making correlations across the internet? Like, does what being said about you here match what is being said about you there?

Meg: Absolutely. And I think it is more authentic way of recommending you. With Google… you could put whatever you wanted on your website, you can make yourself sound like a complete expert in a lot of cases. Google rewarded that. So it was kind of like the one who won the game in Google was the one who had the best optimized website, not necessarily the best expertise.

So what AI is trying to do, they’re like, oh look, we created the tool that makes it really easy to sound like you’re an expert in anything. Now, we need to figure out, are you really who you say you are? So what is on your website is very important, but we need to make sure that it matches across the internet.

Krista: I hear people say “SEO is dead” or “SEO is not dead because AI rewards the same SEO efforts.” Which feels more true?

Meg: I think it’s somewhere in the middle. I think if you look at search, people are always going to be searching for products and services that they want. No, we’re never going to arrive to a place where we won’t need search, like whether we use it via AI or Google.

But old school SEO is dying. The internet’s going to provide less organic traffic because we used to Google something, get 10 links, click on all of them and do our research. Now with AI, they’re recommending a couple of websites. We’re getting a ton of information on the chat, on the response, and not all of our searches are going to lead us to a website. In general, organic traffic is going to go down.

Search is never really going to die, but it is going to look different because AI is just so much more helpful to the consumer than Google searches.

Krista: What old-school SEO tactics are dying?

Meg: I think old school Google search is highly optimized web pages, multiple pages, like articles that would be a comprehensive guide to a topic, and then you would create four or five articles that related and linked to it. So it was kind of like a hub and spoke content marketing plan. I think that strategy is going to be dying.

That kind of broad, very generic content that happens in those ultimate guides… I think those answers we’re going to get on AI.

Krista: As a user, I’m thrilled to hear that — because this type of SEO content marketing never made sense to me. Like let’s say you’re researching, how do I write the best welcome email sequence? You Google it and get an article from Kit or another platform and the first four paragraphs are stuff like:

“What is email marketing?”
“Do you need email marketing?”

You don’t get to the thing you actually asked until halfway down the page. It’s so obvious the content exists for SEO purposes, not for the reader. It was such an inefficient, bad way of delivering information.

Meg: That’s right, self-serving to the business — totally self-serving to the business, very optimized for Google search but not conversions. So it was just putting words out there that you could rank for.

Krista: But does blogging still matter at all in the age of AI search?

Meg: ​​Yes, your content is still what provides context to these tools. So they need to understand who you are, who you serve, the impact that you make, and the expertise that you have. And so typically, case studies and blogs are where they read the most to do that.

Krista: So if AI is looking at what other people online are saying about you, is now the time to invest in PR or start learning how to do it?

Meg: Yes, this is a time for PR because if you can get press and positive mentions, it will build up your expertise and authority across the AI bot LLMs.

Krista: But is it even possible for small businesses or local businesses to get enough people talking about them online for AI to notice?

Meg: You still need to make the effort. One group of clients I spend a ton of time with are local businesses. They’re brick and mortar, whether they are a doctor, or an orthodontist, or a physical therapist.

And if you’re a local business, you can have a Google Business Profile that has your phone number, your address, and your hours. That’s a really great, very trusted third party.

And so for a lot of those clients, I’m like, guys, if you have a LinkedIn, you have a Google Business Profile, and a good website — and we keep the three of those things kind of talking together, you’re in really good shape because those are three highly trustworthy, authoritative sites.

So, I don’t think it’s as scary and overwhelming as it seems.

My recommendation is, hey, blog on another website, share somebody else’s LinkedIn, get on somebody else’s LinkedIn, get on somebody’s podcast, try to be on somebody’s YouTube. Maybe send a few requests out every month and try to get your name out there.

Add in your LinkedIn, or your Instagram, or your Substack, or your Google My Business, and you’ve really built up quite a group of other sites talking about you.

Krista: So if people are talking about you online, does that automatically translate into AI visibility?

Meg: I don’t have a good answer for you yet. I think it’s working. I am getting found on AI just because other people are talking about me, but I can’t link to a specific person or a specific post that’s like, okay, this person said this, which led to clients or led to me being more visible.

I think it’s a combination. It’s all of these little ingredients, and AI is looking at the whole internet going, okay, we’re hearing this person and this person, so we’ll give them more traffic.

Krista: I’ve heard people say that engaging in Reddit is a good idea for AI visibility. Do you agree?

Meg: Yes, I love actual humans recommending and talking about actual businesses because it feels more authentic and more real. So it is a really high quality idea for AI marketing.

However, because now all these people are paying money to have other people “do reddit marketing” for them, it’s not easy to game. There is also a huge surge in AI bots trying to be on Reddit and give answers. So it’s kind of like a battle of robots, like recommending something and then talking about it, trying to figure out, trying to dissect out or tease out actual humans. It is getting trickier and trickier.

All that to say, it is a highly recommended technique for AI, but it is hard and tricky and very time-consuming because you actually need humans to go in and comment and build rapport and build karma on Reddit. So yes, but it’s not easy.

Krista: What about FAQs? I’ve heard we should be writing FAQ sections specifically for AI to read. What are your thoughts?

Meg: This is definitely something that we’re seeing, and we’re recommending.
So what AI does is read the internet and then synthesize data and give you a complete answer. Well, that takes a ton of computing power, and as more and more pages get added every second, they have a hard time keeping up.

So if you can ask really specific questions and have a short, sweet answer on your website or blog, if the LLM sees that, they may not synthesize the rest of the internet and just serve up your answer.

So we are seeing that FAQs are appearing in chats in ChatGPT much more frequently, so we recommend having an FAQ page and then at the bottom of all of your blog posts, do three to five questions that are already answered in the blog post, but just make it easy for the AIs to understand what the article is about.

Krista: How do you know what FAQs people are asking LLMs?

Meg: Great question. With old school Google, we would have all these tools that tell us the search volume of keywords, so you know how often a specific word gets asked. So we would, of course, try to rank for words that had a lot of search volume behind them.

There is not a tool out there that is accurate that gives us how often these questions are getting asked to AI, so I don’t have a good answer for you.

Krista: What are the universal recommendations you give in your AI visibility audits?

Meg: We’re seeing these things work really well for AI for most people:

• At the top of your blog articles, write a one or two-sentence summary that summarizes the whole page. So if you can give Google overviews what the summary is, it tends to show up in the AI overviews.

• Write content that’s easy to understand. Don’t be clever. If you were writing an article about email sequences, be super clear about that because the first thing that reads your site is probably a robot, and they need it to be in simple, clean language. So write really specifically about what it is and try not to be too cute.

---

Follow Meg on LinkedIn. Or jump in to work with her with her $299 AI Visibility Audit to see where you stand in our changing search paradigm + what to do to improve. (No kickbacks for me if you buy — I just know a good offer filling a gap in the market when I see one.)

Xoxo, your favorite website freak,

Krista