What Google's AI Mode Data Reveals About How People Search Now

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

User behavior in AI search has moved, and Google's own data confirms it. This week, Cassie breaks down what that shift actually looks like, why Bing's newest Webmaster Tools features were built to measure exactly this, and what both mean for your visibility strategy.

We also cover Semrush's expanded 2026 AI Visibility Index (126 million prompts — the biggest AI search dataset we've seen), Google's June spam update, and a quick note on Anthropic's Fable 5 and what it signals for the frontier model landscape.

In this episode:

• Semrush 2026 AI Visibility Index — 126 million prompts analyzed across ChatGPT, Gemini, Google AI Mode, and AI Overviews. What the citation gap between "mentioned" and "cited" means for your brand, and why brands with integrated SEO + AI visibility workflows are outperforming everyone else by 2x.
• Google June 2026 Spam Update — Rolling out globally as of June 24th. No new policies, but worth flagging in your reporting.
• Anthropic Fable 5 — Export controls lifted July 1st. Why this matters beyond the headline, and what it tells us about where frontier model governance is headed.
• Google AI Mode data + Bing Webmaster Tools — The anchor segment. Google released a full year of AI Mode usage data showing the average query is now triple the length of a traditional search query, with follow-up queries growing 40%+ month over month. Then Bing drops four new AI Performance features — Intents, Topics, Citation Share, and Compare — that are essentially built to measure exactly how users are searching now. Here's how to use both together.

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'm Cassie Clark, an AI search visibility consultant. Yes, that is a new positioning title. It more accurately describes what I do. We'll see how long it takes for the AI engines to pick it up and start telling people that's what I'm doing. I'm also the host of the show where we break down what's actually happening in AI search so you don't have to spend your Thursday morning doom scrolling through five different industry blogs. I did that work for you. Today is Thursday, July 2nd, and we have an actually kind of big news week. The kind where I kept opening new tabs and going, okay, this is going into the episode, and this is going into the episode, and this, and then I had to think, nope, that's not. We've got four things to cover today. First, Simrush just released their 2026 AI Visibility Index, which is the biggest data set we've seen yet on how brands actually show up in AI Answers. The Google Spam update rolled out, and it's housekeeping, but it's worth knowing about. Anthropic has some news with Fable 5 that's relevant if you're thinking about where Frontier AI is headed. And then the most important thing that we're covering today, Google released a full year of AMO data that shows exactly how user search behavior has shifted. And Bing, because I love them. Has also dropped four new features and webmaster tools that are essentially built to match that shift. Those last two connect in a way that I really don't want to skip over. I'm gonna spend some time on it. So let's just get into it. Let's start with Sumrush because they just published the expanded version of their AI Visibility Index, and the scale of this thing is massive. They scaled from an initial 2500 prompts to 126 million US AI search prompts analyzed from January through April 2026 across ChatGPT, Gemini, Google AI Mode, and Google AI overviews. Now, this data set again is massive, but it tells us some things that I think were gonna feel kind of obvious in retrospect, but are actually useful to have with conferred numbers. First, being mentioned in an AI-generated answer doesn't mean your website is being cited as the source. On Gemini, that overlap between mentioned brands and cited domains can be as low as 30%. That's a stat that I think we all need to keep in our head at all times. It can be talked about without being trusted, and those are two different problems with two different solutions. Second, the four platforms show significantly different citation patterns. ChatGPT cites an average of 15 sources per response. Gemini, on the other hand, cites an average of three. That matters for how you think about which platforms to prioritize in your visibility strategy, and it means you can't just optimize for one and then assume that it always carries over to the other. Sometimes it does, sometimes it doesn't, but that's a good thing to keep in mind and not make that assumption. Third, and this is the one that really closely aligns with the FSA framework that I talk about, which is freshness, structure, and authority. Brands with integrated SEO and AI visibility workflows were more than twice as likely to report increased traffic or leads from AI platforms compared to other brands managing the two areas separately. It's 81% versus 36%. AI visibility is not a separate channel that you just bolt on, and I've been saying this forever. It runs through absolutely everything that touches your organization. That's your content, your PR, your social media, your third-party presence, customer service, and customer support. It touches absolutely everything. The brands running in this index aren't doing just AI stuff alongside their SEO, and they're not relying just on SEO for this. They've built a system with SEO as a foundation and they have expanded it to everything else. Okay, quick housekeeping. Google began rolling out the June 2026 FAM update on June 24th, applying it globally across all languages. It's the second spam update of the year. There are no new spam policies announced alongside of it, so this isn't a signal that the rules have changed. It's more just about enforcement tightening up. If you see your ranking or traffic shifts over the next few days, this rollout might be a candidate for that cause. So we're flagging June 24th in your reporting so that you can separate this from anything that comes after it. There's nothing dramatic here, just note, keep an eye on your Search Console data and then just move on. Up next we have Anthropic and Fable 5. This one's a little bit different from the coverage of the rest of this episode, but I do want to touch on it because I did mention it when Fable 5 was released. Anthropic's Fable 5 model was briefly subject to export controls that started on June 12th after the US government became aware of a report in which Amazon researchers found a method of bypassing Fable 5 safeguards around cybersecurity tasks. Access was suspended for all users during that time. It was kind of odd, like it was here one minute and then gone the next. As of June 30th, those export controls were lifted, and Fable 5 became available again starting on July 1st, which was yesterday. Why does this matter to us? There are two reasons. One, this is a preview of a dynamic that's going to become more common. Anthropi is proposing a shared industry framework for assessing the severity of AI jailbreaks, and that was developed with Amazon, Microsoft, and Google. And it is essentially a common standard for how the industry talks about and responds to model vulnerabilities. That's infrastructure building. I've said this once and I'll say it a hundred times more. When infrastructure gets built, the behavior follows. We have seen that pattern with every phase of this AI search space. Two, the Fable 5 situation is a reminder that the Frontier model landscape is absolutely more volatile than it looks from the outside. If you're building a strategy that depends on a specific model's capabilities, you absolutely need contingency thinking baked in. These models change with the wind. What happens one day might be something totally different than the next. You have to stay on your toes if you're using these models in your day-to-day workflows. Okay, let's turn our attention to the two news updates that I have been geeking out about this morning and trying to connect together. Because when you look at it, it really does feel like an inflection point. Let's start with Google's data. It's a little concrete and it's a little bit alarming if your content strategy hasn't evolved. That's a good word. Has it evolved in the last 12 months? According to Google's own report on how people are using AI mode in the US, covered through May 2025 through April 2026, the average AI mode query is now triple the length of a traditional search query. We had Charlie Graham on the show eons ago, and through his research, he found that search terms have moved to up to at least 12 words in a search query. Google is now finding the same thing, it's triple. That single number invalidates a lot of 2025 content strategy that focused on these short little keywords. That has been a thing for the last forever, but now it's changing. Google also found follow-up queries in AI mode have grown more than 40% per month on average, meaning that users aren't landing on one answer and leaving. They're staying in the conversation and they're going deeper. And the top keywords that AI mode searches information, identify, find, explain, summarize, those were paired with opening words like what, how, and I. And that tells you something important. People are narrating personal context into the search bar. Not something generic like running shoes for flat feet. Something closer to, I have flat feet and my knees hurt. Can you help me find a running shoe that will not make it worse? That's not a keyword, that is a whole conversation. And if you're lucky, it does show up into your Google Search Console report. Here's the thing that a lot of content teams are circling around but maybe haven't connected to yet. Most teams are still writing for that shorter keyword. Optimizing page titles, meta descriptions, H2 structures for the three to four word keyword target that represent a shrinking share of how people actually arrive at answers. I know I sound like a total nerd, but this is a behavior shift we have been watching happen in real time. It is fascinating, and now we have Google's own data confirming it. Which brings me to being Webmaster Tools. Being just released four new features in their AI performance dashboard that are essentially built to measure exactly what Google's data describes. They have new capabilities, and those are Intense, Topics, Citation Share, and Compare. Intense classifies the grounding queries that triggered your citation shares into categories like informational, commercial, navigational, learn and solve, research, creation, local, and a whole slew more. So you can understand not just where you showed up, but why the AI system serviced you. Topic goes a layer deeper. Instead of analyzing visibility one query at a time, publishers can understand which larger subject areas are driving citation activity. So queries like solar panels or solar energy efficiency or something like residential solar installation, those get grouped into a cluster like solar energy. That's how AI systems actually reason across themes, it's not isolated by keywords. And now you can see your brand through that lens. Citation share shows what percentage of citations attributed to your site out of all citations shown across all sites for that same grounding query. It's not whether you were cited, but more how much of that space you held. So you know, share a voice. And compare lets you overlay time periods. So something like compare the current 30-day period against the prior 30 days, or you can select custom date ranges to better understand evolving citation trends. Here's where I've been nerding out. Why do these two stories connect? Well, Bing just built a reporting layer that mirrors how users are actually searching. Google's data tells us users are operating in themes, context, and conversations. Bing's tools now let you see your citations through intent categories and topic clusters. We're not just looking at those keyword strings. That is a loop that is closing and I'm really excited about it. User behavior has moved well past keywords and the tools are now catching up. The question is whether your content has. Now, if you're looking for a place to start with this, pull your grounding queries and Bing Webmaster Tools and then look at which intent categories are showing up in. Then take a look at topics. Are you showing up where your audience is actually asking questions? Or are you showing up for the three-word version of a problem that users are now describing in full sentences? That is the work that we need to do. It's an easy fix. It just takes some time to go back and edit all that content, but it's well worth it. That's it for this week. It was a lot of news, and honestly, most of it is pointing in the same direction. User behavior has moved. The infrastructure is building around it, and the brands that understand that connection now are the ones that are going to show up in AI generated answers. If you want to dig into what any of this means for your specific visibility strategy, you know where to find me. Head over to CassieClark Marketing.com. There'll be more information in the show notes. I'll be back next week. Until then, stay visible.