57% Bot Traffic, One Year Early. And Google Picked This Week to Say "Check With Us."

📬 You like this podcast? You’ll love the newsletter.
Join the weekly 3-2-1 on AI search + marketing: subscribe

This week on Found in AI, Cassie is going solo today. There are a few things that happened in the last week that make a lot more sense for your AI search visibility strategy together than apart.

On June 3rd, Cloudflare's CEO posted that bot traffic has officially crossed 57% of all web traffic — the first time in internet history that automated systems outnumber humans online. He'd predicted it wouldn't happen until 2027. It arrived more than a year early, driven almost entirely by agentic AI.

Two days later, Google published new guidance asserting itself as the authoritative source on SEO, AEO, and GEO — telling businesses to evaluate all third-party advice against their official documentation.

And in January and April, Microsoft had already published a GEO playbook and a framework for the agentic web... with matching tooling.

Cassie breaks down what each of those moves means, why the timing matters, and where Google's GEO guidance specifically diverges from what independent practitioner testing has shown works.

In This Episode:

• What the Cloudflare bot traffic crossover actually means for content — and why the five vs. five thousand ratio explains everything
• Why bot traffic hitting 57% arrived more than a year ahead of the Cloudflare CEO's own prediction
• Microsoft's "three eras of the web" framework — and why all three are live right now, not sequential
• Why Microsoft's January GEO guide said "SEO got you here, but it won't get you there" four months before Google weighed in
• What Google's new guidance actually says — and the specific line in their SEO hiring doc that Cassie takes issue with
• Where Google's GEO advice breaks down when applied outside of Google's own systems
• The chunking disagreement: what Google says to avoid, what Cassie's December Perplexity experiment showed, and why the distinction matters
• Three things you can do this week to optimize for the multi-engine reality

If you're listening to this and thinking I need someone to lead this for me, that's what I do.

I'm an AI search visibility consultant and a fractional content strategist for startups and enterprise brands. If that sounds like the kind of help you're looking for, email me at cassie@cassieclarkmarketing.com. 

Or request your 7-Day AI Search Visibility Audit: https://cassieclarkmarketing.com/ai-search-visibility-audit/

Let’s connect:

LinkedIn → Cassie Clark | AI Search Visibility Consultant
Website → https://cassieclarkmarketing.com

Transcript

SPEAKER_00 0:00

Hey, welcome back to Found in AI. I am Cassie Clark, a fractional content strategist, an AI search optimization expert, and your host of the show where we track what's actually happening in AI Search so you don't get left behind. Today is Tuesday and I'm going solo today. I know I usually share guest interviews today where we get together and we talk about what we're actually seeing in AI Search and SEO and all the things. But instead of rushing into the interview that I have that I could share today, you're getting something that is a little bit more useful right now. Because there are a few things that happened in the last week, and I want to dive a bit deeper into and share my thoughts on those. If you follow me on LinkedIn, you already know where this is headed. And if we're not connected on LinkedIn, I would love to connect with you over there. But I wanted to record today's episode because there's so much chatter about this particular topic, and I just kind of want to give you some advice on it all. Here's the short version. The web crossed a threshold that nobody expected this soon. There are three major companies that are responding to this threshold in three very different ways. And I think what each of them choose to say and choose to do about this kind of tells us where we're headed in the future. We're talking about Microsoft, Cloudfair, and Google. Let's get into it. So I want to give you the full picture because this episode is really about context, about understanding all of these things together instead of just one at a time. I think it's more useful when we put them all together and really look at it. First, we have Cloudfair. Cloudfair is one of the largest internet infrastructure companies in the entire world. The drops in data last week, it was last week, showing that for the first time in the history of the internet, bont traffic now makes up the majority of web traffic. We are talking 57% automated and 43% human for the first time ever. Second, Google published new guidance that basically asserts itself as the authoritative source on SEO and specifically AEO and GEO. As of June 5th, they now have a page that essentially says evaluate all third-party advice against our documentation. Okay. Third, we have Microsoft. They had already published a GEO guide in January and a framework for AI era web strategy in April. They have not published anything new this week as far as I can tell, and they don't need to. When you hear all three things together, you get a much clearer picture of where AI Search actually is right now. And, if we're honest, who's been paying attention? Okay, let's start with that CloudFair news because it's kind of really the foundation that every bit of this sits on. On June 3rd, CloudFair CEO Matthew Prince posted on X, and I'm quoting this directly, he wrote, Well, that happened faster than I predicted. He's talking about that bot traffic. Bot traffic past human traffic online. Now, for the first time in internet history, the majority of requests moving across the web are not coming from people. They're coming from these automated systems, specifically identical. Programs that browse and retrieve pages on behalf of assistants like ChatGPT and Gemini. Now, here's what makes this really interesting. In March, three months before you posted that over on X, Fritz was at SXSW. I'm slow this morning. He was over there telling audiences that the crossover wouldn't happen until 2027. He was off by more than a year. Now, this is the CEO of a company that literally processes a significant portion of the internet's traffic. He had access to that data in real time and he still did not see it coming that fast. That's not me trying to knock on him. Not at all. This is why the math tips so fast. Every time someone asks ChatGPT a question that requires web retrieval, it doesn't visit just one page. It fans out across dozens or hundreds of sources. Multiply that across millions of queries a day, and you see how bot traffic goes from a nice little chunk to the majority, faster than anyone predicted. And for those of us working in content and AI search optimization, this has a very direct implication. Your content is now being read primarily by machines. Not exclusively, humans still matter, but the majority of let me try again, the majority of requests hidden in your pages right now are from those systems that are evaluating your content for retrieval, not the humans that are just casually scrolling it. This is what the web is now. And the question is, are you optimizing for that? Let's talk about Microsoft for a second. I have been pretty openly a fan of how they've handled this, and I think the contrast between them and Google is actually really interesting. Microsoft published their GEO guide in January. I covered it on the podcast in February. They dropped it January 6th to be precise. They opened it with this line SEO got you here, but it won't get you there. And I really, I really love how that is phrased. They named GEO as something new, something distinct, something that requires a different approach. They didn't frame it as a subset of search, but they called it a separate discipline. And they said, here's a playbook. Use it. Then in April, they published a piece called When Across All Three Eras of the Web. And the framework they laid out, I think is worth knowing. They describe three modes of how people are using the web right now. First, we have the human web, which is help me find it. Essentially, someone searches, they browse, they click. This is most of the traffic that we're used to seeing. Then we have the LLM web, which is basically help me choose. So someone uses an AI engine, it synthesizes the options, it surfaces recommendations, and the person doing the querying still makes the final call. Then we have the identic web, which is do it for me. An agent goes out, it finds something, it evaluates it, it completes a transaction on someone's behalf, the human doesn't visit your website at all. Now, what I love about this framework, and what I think makes it useful instead of just conceptual, is that Microsoft explicitly says these three eras are not sequential. They're not a roadmap where we leave one behind to enter the next. These three things are all happening right now in parallel across every single customer base. The Cloud for data kind of proves that point. The Agentic Web isn't coming. This isn't some future thing. It is already the majority of traffic. It is here or right now. So Microsoft also dropped something else in April that I also want to mention because it's very relevant to what Google did two days after the Cloud for announcement. Microsoft watched AI visibility and Microsoft Clarity. This is free. It shows publishers which of their pages are being cited in AI-driven answers, how often, and where competitors are being cited instead of them. This is where I get to say what I'm actually thinking. Microsoft has been doing this work for months. They named GEO RealTink back in January. They've built the tooling for it in February. They published a framework for the Agentic Web in April. Do you see where I'm going? They didn't wait for this dust to settle, they just moved. And they've been trying to help their publishers figure it out as they go. Which brings us to Google. On June 5th, like two days after CloudFare announced that bots now outnumber humans on the internet, Google published new guidance on Search Central. The title of the Search Engine Journal coverage of it says it well. This was my favorite headline so far. Google's new guidance claims authority over SEO tools and AEO and GEO. Okay, okay, Search Engine Journal, we see what you're doing. Here's what the guidance does in plain language. It tells businesses to think critically about third-party SEO tools and services. It distances Google's from vendors that claim their tools are approved by Google. It says Search Console is the tool to use, and it asserts Google's own documentation as, I'm going to quote it directly, the standard against which any other advice should be evaluated. Specifically, it extends that to AEO and GEO. There's a line in the SEO hiring guide, the page where they give advice on how to vet an SEO professional that asks, if your SEO has been has advice on optimizing for AI experiences, is their advice aligned with Google Search's official guidance on optimizing for generative AI features. Now, I want to be careful here because I don't think that guidance is wrong exactly. Some of it is actually useful. The reminder that third-party tool data is not Google's internal ranking data. Okay, fair point. And the reminder to think critically about before you implement any advice you found on the internet. Also, fair point. I have learned my lesson about this a hundred million times, and I'm not just talking about things related to SEO. But the framing, I think, is what really burns my biscuit. Now, here's why. Google is positioning themselves as the referee of a discipline at the practitioner community. People like you and me and the guests that I've had on this podcast have spent the better part of the last 18 months building almost entirely without them. Google was not in this conversation. We've had the frameworks, we've had the tests, the experiments, the share results. That knowledge got built in communities, in newsletters, in LinkedIn threads, in Discord servers, in podcasts like this one, in Reddit commentary, not in official Google documentation. And then in June of 2026, four months after Microsoft said, SEO got you here, but it won't get you there. Google arrives and they say, actually just check with us first. There's also a specific disagreement I have with Google's Geo Guidance that I want to talk about because I think this one matters. Google Guidance is not to chunk your content. I hate the word chunk. They're framing this essentially: do not reduce your pages to flashcard style formatting. Listen, I get the spirit of it. They're trying to prevent people from publishing junk content that looks like it was formatted for a robot or looks like a bad PowerPoint that third graders are putting out for their presentations. That concern is valid, I truly understand it. I was a teacher, I have seen those presentations. But here's what I know from my own testing. Structured, extractable content works. Now, if you remember back in December, I ran a 24-hour experiment specifically looking at how perplexity receives retrieves and cites content. Pages that were clearly broken into sections, definitions, and step-by-step structure, those pages got cited faster and more consistently than dense pros. The experiment is now about six months old, and I'm due to update that page. But because, as we know, freshness plays a role in what content gets cited, and the recommendation is to update your content every three to six months. We are like right at that end of the recommendation, and I need to go do that. But the underlying logic tracks with what Bing has told us in their AEO and Geo guidance. Structure is machine legibility. Google's advice is written for Google Systems, not for per not for perplexities retrieval model, not for ChatGPT, not for Claude, not for the broader ecosystem where the majority of your AI search visibility actually lives. So when Google says align your GEO strategy with our guidance, what they really mean is optimize for Google AI overviews, which is fine. That's one piece of the pie, but it is not the entire pie. And treating it as the definitive standard is exactly the problem I'm pointing at. So here is what I want you to walk away with this morning or today or whenever you listen to this podcast. I say this morning because I am recording this morning and I usually do this a day ahead of time. I'm just off of my schedule. One day I'll be back to it. But anyway, keep in mind now that the web is not the same as it was a year ago. More than half of all traffic is not automated. That happened almost a year, 18 months-ish ahead of schedule. If you're still primarily optimizing for human readers, or for Google specifically, you're optimizing for the minority of what's actually visiting your content. Now, this does not mean that we stop considering our human readers. They still matter, they always matter. But it does mean that the systems doing the majority of the reading need to be able to do something with what you've written. They need to be able to extract it, cite it, use it in an answer, or do some kind of action with it. And the strategies that help them do that are not identical to what Google is recommending for search. There are three things that you can do right now. One, check your content structure. I know I talk about this all the time, but do you have clear headings, definitions, and sections that can be pulled as a standalone answer? If not, that is the absolute highest leverage thing that you can fix, not for Google, but for the broader AI ecosystem. Two, if you haven't looked at Microsoft Clarity's AI visibility feature, go look at it. It is free. I'm looking at it this morning. It shows you which pages are getting cited and which aren't. That's actual data. And it's a feedback loop that we can use. Now remember, Google has released a generative search report in Google Search Console, but as of June 9th, it's not yet available here in the US. Some websites in the UK have access to it, but as we know, it's impressions only. It's not actually going to tell you which query someone typed to get your brand recommended. So look at these other tools that's giving you that data. It's my advice here. Now, three, my final piece of advice this morning, or something that you can do. Take Google's GEO guidance as just one input, not the input. It's not the gospel. Do go read it and understand it. Apply what's relevant to your Google AI overview strategy, but don't let it replace what the practitioner community has learned through independent testing. Those two things can coexist at the same time. And this is not me telling you, hey, go out, do spammy tactics, those things will get you in some hot water, but I'm just saying, like, take what everyone else has learned and apply it to your strategy in the best way that you can. The brands that show up consistently in AI generated answers in 2026 and beyond are the ones who understood this early. This is a multi-engine game. It's not just Google yet. It's a visibility game across AI engines. Hey, I'm Cassie Clark. If you want to know where your brand actually stands in AI Search right now, which engines are citing you, which topics are missing, where competitors are showing up instead of you, head to CassieClarkMarkety.com. The AI Search Visibility Audit covers all of that. This episode was helpful. A rating wherever you listen would be really nice. It takes 10 seconds and it helps new people find the show. I would love you forever if you left a rating. Big shout out to those of you who already done that. Seriously, thank you. Okay, that's it for today. Stay visible. I will see you on Thursday.