Latest / The Search Signal / How To Future-Proof Your Search Strategy for AI Mode
Transcript
- michael: AI Overviews disrupt the clicks from even like the top most search positions in last year in April, so 2025, Ahrefs had found that the top ranking pages lost about 35% of the clicks you would normally expect from that spot when an AI overview was triggered on the page. And then by February 2026, 10 months after that, that loss had grown to almost 60%. Hey, welcome back to the search signal. I am Michael Transon. I'm the founder and CEO of a search marketing agency called Victorious. And here at the Search Signal, our goal is to help marketers understand and take action on what is happening in the world of search marketing. So, as a team, we work on hundreds and hundreds of websites every year for some of the biggest and fastest growing brands out there. And we bring what we learn from those experiences plus our own research to help you with your own work. So before we get into today's topic for the episode, let's start by talking about what has happened this week in search marketing news. Google is testing a couple of pretty significant potential changes to their traditional search results this week, including a sign-in requirement to review and view search results past the first few pages. So, according to Search Engine Land, some searchers are seeing a prompt to verify that they're a human once they click through the first few pages of Search results. Now, Google has not confirmed this publicly or said how far it plans to roll this out. And this test appears to be aimed at bot traffic. But more importantly, it would just as easily catch the rank tracking tools that you might be using, things like Semrush or Ahrefs or Profound, since most of them pull that data by paging through some of the results the same way that a bot or a scraper would. So this is absolutely something to keep an eye on because if it does block bot traffic, it could also seriously impede some of these third-party tools that you might be using to track your search performance. And the other thing happened this week is Barry Schwartz also reported that Google is testing some new buttons on its homepage. So some of the users have been reporting seeing things like create images and ask about files. And also an interesting one called brainstorm that are buttons that lead to generative AI experiences below the regular search box. So again, no word yet on this either from Google about this test. But those are the two big search news things that are happening this week. We'll see how those evolve. But with that being said, let's get into today's episode. Every week we get together and talk about some aspect on this podcast of how search is changing. But today I want to pick up on one of those threads and do a deep dive specifically on Google's AI Mode and also how I think it fits into the future of search marketing. And more importantly, I also want to talk to you about what you can be doing to prepare your business for the search changes in AI Mode and also Google as a whole that I think are ultimately going to be coming down the road. So, first, we're going to quantify what is actually changing in Google's search experience, particularly over the past year. Then after that, I want to discuss how Google's AI rollout has accelerated and the momentum behind those changes. And then I want to talk about how all of these different search surfaces are working together. We need to understand how buyers are able to move between traditional search and Gemini and AI Mode and AI Overviews to help us navigate these changes. And by the end of this conversation, I want you to have a very clear sense of what you can start doing right now to mostly benefit from where Google Search is heading. Instead of trying to play catch up after these things start to happen. So let's dive in. What exactly is changing right now in the Google search experience? There are, I would say, two big pieces to this. There is AI Overviews and AI Mode. And I want to talk through and walk through both of these, starting with you know what we know for a fact and also what we have been able to measure as Google has been evolving in the last few years. So let's start first with AI Overviews. So if you are not aware or familiar, AI Overviews are the snippet of content that shows up at the top of a search result that typically will give like a consolidated answer or an insight to a query or a question that you submit in the search bar. And you might be familiar with AI Overviews yourself. They started showing up around mid 2024. And at that time, they would only really surface on a few specific results, but their presence has been growing and it's been growing monthly. There's a study that's linked in the show notes that tracked AI Overviews and they did it daily from February 2025 to February 2026. And it found that AI overview presence grew from roughly 30% of queries to now almost 50%. 50%. And the same study also found that AI Overviews are taking up more room in the search results. So the typical AI overview grew from about 1,000, or 1,050 pixels tall to about 1200 over that same year. And some of them are actually stretching more than 1300 pixels. And for context, a standard desktop screen only shows about 900 pixels before you have to scroll. So On those results, the AI overview alone is taller than what most desktop screens can fit. And it also pushes both the sponsored and the organic results further down the page. So as of today, you think about it, AI Overviews are taking up the majority of the real estate on over 50% of all searches that are happening in Google. And then at the same time this is happening, the share of zero-click searches is just keeping on going up. And if you're not familiar, a zero-click search is it's essentially exactly what it sounds like, right? Someone runs a search, they get what they need right there on the results page, and then they just never click through to any website. There was a published study this past June that measured how often Google search ends in a click. And between January and April of this year, almost 70% of Google searches ended without a click. And that is up from almost 60% in 2024 and roughly tw 45% back in 2016. So right now, only about 30% of all the searches are actually resulting in a click to the open web. And you know, the question becomes why is this happening? And it's AI Overviews, right? They're designed to quickly and very simply answer just common search queries for users without needing to click on a website in the search results to get to that answer. So as a result, these searches are not yielding clicks like they have in the past, despite Google search volume actually has been increasing year over year. And now AI Overviews have. you know, historically surfaced, I would say more commonly, in informational queries. And those are questions that are not necessarily commercially minded. They're more like ones where a user is looking for more general information to a question. And and as a result of that, marketers have, you know, generally accepted over the last few years that AI Overviews are to be expected on informational searches. But that is also changing. So SEMrush, for example, they tracked almost 600,000 US keywords. They did it from November of 2025 to April of 2026, and they found that AI overview visibility existed on commercial keywords. And, you know, those are when users are actively searching for something to buy. And the commercial searches grew by more than two-thirds in just like six months. Additionally, on top of that, you know, AI Overviews are now showing up alongside paid ads. And they're showing up roughly twice as often as they were just like a year ago. So personally, I feel like I get an AI overview on pretty much every search that I run now. And the data from both of these studies shows that that might be also true for other users like you, right? Overviews are just appearing a lot more often, and they've moved also exactly to the commercial queries that typically matter for most of our pipelines. So I would say taking a step back rather than feeling like this is a dire situation, I think that there is an interesting opportunity in the trend, right? People who do click through are going to be better educated prospects. So when someone lands on your site after reading an AI answer, most of their evaluation has already happened. And by the time that they click, they've likely narrowed down their choices. Like this is also reinforced by some research that we saw from Microsoft's Clarity team. And they published this last November. They found that AI referred traffic converts at roughly three times the rate of other channels. And I saw and read a white paper recently that came out of Stanford's business school where they called this the "perfect click." And all of this also means that AI Overviews are affecting the value of your traditional ranking position, right? A number one organic ranking used to guarantee that a searcher would, you know, see your listing near the top of the page. And, you know, historically moving up from position three to maybe position one produced a really predictable jump in clicks to your website. But AI Overviews disrupt the clicks from even like the top most search positions in last year in April, so 2025, Ahrefs had found that the top ranking pages lost about 35% of the clicks you would normally expect from that spot when an AI overview was triggered on the page. And then by February 2026, 10 months after that, that loss had grown to almost 60%. So as a result of all of these updates, showing up in AI Overviews is becoming a much, much more important component of both SEO and search marketing than it ever has before. And it's something that all marketers need to be aware of as they are optimizing for attracting highly qualified clicks from search to their own website. So that covers from a high level AI Overviews. But the second big change that is happening is something called AI Mode. And as you'll see, since this episode is really built around AI Mode, I want to make sure that we are all working from the same picture of what it is. So AI Mode is a full conversational search experience that lives inside of Google search itself. So you'll see it as its own tab on the results page, typically sitting, you know, right next to where you would normally click over to things like images or news. So when you run a search there, you do not get a list of results. Similar to other tools, you're going to get a complete AI written answer, and it's going to have links folded into it, and there's going to be a box underneath where you can keep the conversation going with follow-up questions. So if you've used like Gemini or ChatGPT or Claude, the experience of AI Mode feels a lot closer to those than to a traditional Google search, except it's sitting inside of the search product that you are already using every single day. So the simplest way to keep the two straight between AI overview and AI Mode is AI overview is a summary added to the top of a normal results page. And then AI Mode just replaces the result page completely. Now, on top of this, some advancements that have happened lately between the two is we are seeing now the two are also starting to feed into each other. So now at the bottom of an AI overview, Google is showing a button And an ask anything bar that's inviting you to keep going in the conversation. And the moment that you type a follow-up question, you are suddenly now in AI Mode. And you didn't pick a tab or opt into anything. You just asked a second question, and Google moved you from one experience right into the other. And that handoff is live on mobile, pretty much everywhere both of those features exist. And Google also started testing a merge of the two late last year. So every AI overview, and remember, now these show up on roughly half of all searches right now, is now a doorway into AI Mode. And that replacement is going to change how people behave. SEMrush had analyzed like tens of millions of US Google search sessions during AI mode's first couple of months. And what they saw was that people were treating it like a conversation, like instead of a typical search box. So queries in AI Mode were running nearly twice as long as traditional searches. And people asked fewer of them per session because each answer was effectively covering more ground for the user. And on top of that, very importantly, the clicks are basically disappearing. So more than nine out of 10 AI Mode sessions ended without a single click to an outside website. So that zero click trend that we had just talked through on the regular results page is even more pronounced in AI Mode. However, with that being said, the upside that we saw with AI Overviews that the clicks that are happening are highly qualified and have a lot of strong intent also holds here too. And I would say probably even more so. So the the few people who do click out of an AI Mode answer have already done most of their evaluation inside the conversation. So the visit that we are going to get is going to be a very qualified one. Now let's talk about AI Mode adoption numbers. So SimilarWeb estimates that AI Mode accounted for about a third of 1% of all Google searches from January through April of this year. So that sounds like a tiny amount. But when you consider that, like there are almost 14 billion searches that are happening a day. Which means that there are almost 45 million searches that include AI Mode. And separately, Google reports that more than one billion monthly active AI mode users exist on the platform. And also says queries have also more than doubled every quarter since launch. So both of these numbers are accurate, but they're measuring very different things. So Google's figure is built to demonstrate I would say reach in momentum. And they haven't published how they define a monthly active AI user and whether that report means you know, repeat usage or or even how that audience could overlap with AI Overviews or even traditional search or Gemini. So point being, someone can potentially count as a monthly user after trying it a couple of times. And still running though almost all of their actual searches through traditional search or AI Overviews. So neither number answers the question that I would say would actually help us build a plan, which would be things like how often does an average person choose AI Mode? Or how much of the growth has moved over from Google's other surfaces instead of just being straight new demand? And also like what share of Google's total search volume does AI Mode actually really represent? We don't know any of that. But here's how I view the bottom line on AI Mode usage numbers, because here's what I think. Because Google controls the default search journey, and because it can insert AI experiences directly into habitual user behavior, Google doesn't need to convince you to pick AI Mode and use it. It just needs to keep putting it in front of you inside of the search experience that you're already used to using. So that follow-up bar on AI Overviews that we talked about earlier is exactly that. So this is where I think it's all heading. And full disclosure, this is this is speculation, but I'm basing this honestly on a forecast of Google's track record and the evidence that I've been seeing in the data. I think Google will eventually start sending some search bar queries directly into AI Mode instead of loading a traditional search result page first. So they haven't said that, you know, anywhere, but in my opinion, it fits the direction of everything that they have shipped this year, especially now that AI Overviews and AI Mode are merging into just one unified experience. And honestly, searchers might already be meeting them halfway there. So SimilarWeb reported about a 5% increase in the average query length since AI Mode launched. So that doesn't prove AI Mode caused that change, but it is consistent with people starting to talk to search the way that they talk to an AI assistant. And remember the two numbers that I gave you just a minute ago. So AI Mode is sitting about a third of 1% of Google's total searches against Google's own claim that over a billion monthly users exist with queries doubling every quarter. So on its own, neither number tells you AI Mode deserves your biggest push right now. But what it tells you is everything that Google is shipping, right? There's an accelerated rollout timeline. They are redesigning the search box. There's merging of search surfaces, and they're testing new ad formats. So if you are waiting for adoption data to catch up to Google's product decisions and waiting to get that information before you adjust your search strategy, I think you're going to find yourself back on your heels. And so my recommendation is that I think marketers need to prioritize showing up in AI Mode, arguably over the traditional result page placement, even before the usage numbers eventually are going to force your hand to do that. Okay, so let's get into what you can do about all this. So, and also a quick note on where this picks up, because last week's episode of The Search Signal and this one, I would say are really two halves of the same argument. So last week I had walked through which platforms deserve a dedicated strategy. And my recommendation was to put Google Surfaces first. And that would be traditional search, AI Overviews, and AI Mode and Gemini also as well. But today really builds on top of that. And the the bet underneath this whole episode is that Google search ultimately is not going away. What's happening is it's adapting. And I really believe that AI Mode is going to become the predominant search experience, the place where like most search activity is going to happen rather than the traditional results page. So I would say this when the same work helps all of these Google surfaces, that's great. That is the ideal outcome. But when it doesn't, AI Mode needs to get the bigger share of your attention. And the clicks that you earn there are going to be fewer, but they're going to be better qualified, which is the trade that we have to make peace with. And, you know like I said earlier, what makes this really manageable is most of a really strong search program is going to carry across every surface. And we've covered these foundations in pretty much every episode of the show so far. So I'm going to move through them quickly as a reminder for you, but then we're going to spend the rest of our time here on what's very specific to AI Mode. So let's start with foundation number one, which is machines have to be able to read your website. And we went deep on this last week. You know, we talked about a 30-second check where you can go in, view your page source and search for a sentence that you can see on the screen. So I'm not gonna rehash it here. Go back and re-listen to that episode. But the same family of work there, that technical work includes things like clear page titles, really specific product and service descriptions, logical internal links, and also structured data, which is basically like labeling your pages so a machine knows what it's looking at. You know, this is a product, this is a price, this is a review. we'll link in the show notes to a guide on how to implement structured data. So you can you can look into that further on your own. That's foundation number one, the technical health of the website. But foundation number two, your website has to be able to tell a very accurate and also substantive story about who you are. And that story has to match with what the rest of the web says about you. So if your website is thin or it doesn't really demonstrate real expertise in your space and the content on it, you are not giving AI systems any really raw material to trust. And that story also has to be able to agree everywhere. we covered this back in our brand reputation episode, where your version of your brand and an independent version of your brand on another website might conflict. The independent third party sources tend to win. So as a result of that, we have to be able to audit our third party mentions. And we need to be able to confirm that on other websites, they accurately reflect what you do and what you've earned as a brand. And then if they don't, we need to be able to contact those sites to request updates so that they don't. create an issue with something we call brand disambiguation. So that's foundation number two. Foundation number three, we've talked about a lot, which is third party mentions. I think three weeks ago, we talked about this in our Q2 quarterly search report. And if you missed that episode, go listen to it. And the finding that really matters here is that the two strongest predictors of getting named by AI, referring domains, and third-party mentions. Both live on parts of the web that you do not own. So AI recognized in our study 96% of the brands we studied, but still didn't name 89% of them. So being understood as a brand is like, I would say, the entry ticket. But getting named is actually one offsite. So earning that presence takes outreach, it takes relationship building, and it also takes content good enough for someone else to publish. So you can do this manually. you can build a target list and send a lot of emails. And I will be totally honest, that is really hard. And the return on your time is pretty small. So the better option for most teams is partnering with someone who runs this as a service. And and you can run with a agency or a partner if you've resourced the program budget to be able to do that. So either way, we need to be able to fund third-party mentions deliberately instead of using whatever is left over in paid media. So that is foundation number three. And then foundation number four for all search surfaces is going to be measurement. And I would say this is the cornerstone. Before you do any of this work, you need to know how you are performing right now. And most companies genuinely don't have any idea. And what usually happens at this point is someone in the company runs a search that they think the buyers would run and they see where their brand shows up. And then they just decide right there whether they are good or bad at performing an AI search. And I will say that is not a measurement. That's essentially an AI coin flip because AI answers, we've talked about this, are not deterministic. The oftentimes the same question asked twice gets you two different answers. So what I would do instead of that is build an actual real prompt strategy mapping the questions that your buyers are asking across their entire journey. So when we do this at Victorious for a company, we typically land at about at least a start, about 100 prompts, right? And doing that selection well is actually pretty complicated. It's the kind of thing where, in my opinion, working with someone who builds prompt strategies for a living does pay off. And yes, I mean, Victorious, I run an agency that does this. So you can weigh that recommendation however you want. But once that's done, then you track that set of prompts consistently over a long period of time. And the number that you care about is something called your share of voice. And that means how consistently you show up across the whole set of prompts asked repeatedly over a period of time. And we laid this out, the full tracking setup in some earlier episodes, talking about splitting citations from brand mentions, how we keep platforms separate, and also how we split early versus late buyer questions. So we will link that episode rather than rebuild that and rehash that and spend the time here. So, but with that being said, that is the fourth foundation. And those are those are the main foundation. And I would say those all are important and they are non-negotiables and they all matter very much no matter which surface we are working on. Now, with that said. Let's talk about AI Mode specifically because that platform has its own personality. And the whole point of prioritizing it is understanding how it works. So the first thing to know is that AI Mode typically leans on Google's own properties more intensely than any AI surface leans on anything else. So Google dot com is the second most cited domain inside of AI mode's answers. And this is mostly driven by things like business profile cards and product panels that are showing up in line right inside of the response. So for AI Mode specifically, the Google-owned assets out there are not a side project. So if you are like a local or a service area business, that means a complete business profile and an active review program for your Google business profile. Or if you sell products, for example, it is going to be your product data and your merchant center feed. Or if your business, you know, lends itself more to video, YouTube is going to need to be a core part of your AI Mode strategy and not some nice thing to have. The second thing that we need to focus on with AI Mode is AI Mode reads differently than a search engine ranking. So when someone asks it a question, AI Mode breaks that question down into smaller sub-questions and then retrieves individual passages to answer each one of them. And it's not whole pages. And people also talk to it differently too with these. We talked about this earlier, longer, more conversational questions. So both of those change what winning content looks like in AI Mode. It needs to be built around the decisions that buyers are making. These are like comparisons and trade-offs and pricing options, use cases, and all of the follow-up stuff that someone just naturally asks a AI assistant next. And we have to structure the content so that each section in our content stands alone. Typically, I would recommend with a heading that matches or is close to the question and then a direct answer to that in the same block. So a page can be honestly great from top to bottom and still lose an AI Mode if there aren't single passages answering specific questions on their own. And then the third thing, and this also came out of our Q2 research. AI Mode is much more willing to say brand names out loud than some other AI assistants. So comparing it against, for example, AI Overviews on the same questions, both tools or surfaces cited sources at about the same rate, but AI Mode named brands more than twice as often. So that makes, in my opinion, AI Mode the Google surface where the commercial opportunity is really concentrating. And it tells you also what to watch for in your tracking. We need to be watching for brand mentions on commercial prompts specifically. These are the prompts where you are, I would say the prompts in the commercial set where you are not showing up in a brand mention. These are great ideas and considerations to be included in your upcoming content roadmap, right? You just pick one, pick a prompt. And produce a piece of content that directly answers it. And then watch, see whether you start earning your way in, either as a citation or as a brand mention. And you know, the sites that are being cited in these commercial answers are also the sources that Google trusts to formulate what your buyers see, which also makes that list a really good opportunity for outreach targets, right? These are lists and roundups that are, in my opinion. Worth reaching out to to get included in, right? Sending an email to the website, finding out who the webmaster is, asking for inclusion into that piece of content that's being used as a citation, and be aware that usually if you get a response back, there will be an ask for a cost. So it needs to be incorporated into your budget. And then lastly, specifically for AI Mode measurement, and I also want to get a little bit more specific here because. You know, track your performance advice is useless without naming what is actually open and what we need to actually write down. So there are two different things we need to measure in AI Mode, and they live in two different places. So the first is what AI Mode says about you, and that is your visibility. And then the second is what AI Mode sends you, which is traffic and conversions. These are two different systems. And you know, in a lot of organizations, they're gonna be different owners. So let's talk about each one of them at a time. So let's start with visibility. The visibility side runs on the prompt strategy that we talked about just a few minutes ago in this episode. So you take that, you know, 100 prompt set. These are questions that are mapped across the whole buyer's journey. And I would strongly recommend against you know a couple prompts that someone might have made up in a meeting. And you track that against AI Mode as its own separate performance and data line. And I would not, if we are focusing on AI Mode, see it as part of a blended AI number. and also, you know, to be clear, a cumulative share of voice, right, across all the AI platforms, it's a very important, genuinely useful metric. It's I would say it's a trajectory number, right? The one that tells you you know, I would say at a glance, whether your overall visibility in AI is growing definitely belongs at the top of your reporting, but it just cannot tell you where movement is coming from in specific platforms or what to do about it. So we covered this in our Q2 episode, our report episode about how differently each of the platforms behave. So for the work we're talking about with AI Mode, we need to break it out on its own. So each month, we're recording four things from these prompt retrieval runs, we need to talk about and track our share of voice in AI Mode specifically, meaning the specific this means the percentage of runs where your brand is appearing at all. Then we need to track our citation rate and our brand mention rate. And we need to track these separately because they do tell us different things. And then we also need to split between the early type of prompts in buyer's journey. Do I need this questions? And then later on, you know, which one should I buy questions? So in practice, the repeated runs are the reason this lives in the AI visibility platform. Okay. the tools exist out there. there's are there are a lot of them. There is Profound, there's Scrunch, there's Peec and there's dozens more. These tools I strongly recommend that you get you get into your system because it allows you to ask, you know, your hundred prompts over and over and over and over on a schedule so that you don't have to go out and manually track these prompts. If you are working with an agency, they should be setting up this tool for you. This should be their job and you should be expecting these types of numbers inside of your reporting. Now that is the prompt tracking side. Let's talk about the traffic side. So this is just a very simple, basic baseline pull from your GA4, which is your website analytics, right? So we need to pull monthly organic sessions, conversions from organic, and direct traffic. Ideally, going back 12 months so you can see any seasonality that your business experiences. And this is also something we want to grab from Search Console, some different metrics, monthly clicks and impressions and split into branded queries, which is searches that are containing any variation of your company name, and then all of the rest of the queries. Now the question becomes why this specific set of numbers? It's because Google won't isolate right now, as of right now, the AI effect for you. So you have to set up a system so you can read it as a pattern. GA4 did add an AI assistants channel, and they did that this spring, but the catch on that is it only gets referrals from AI tools on their own domains. So I would like think like something like chatgpt.com. A click out of AI Mode still arrives labeled as Google Organic. And Search Console folds AI feature traffic into its ordinary web results in the same way. So we need to be able to identify the pattern that says your program is working. Okay. Share a voice climbing in your prompt tracking, branded search impressions and direct traffic rising. And that is representative of how people are encountering you in AI answers and then coming to find you in search. And this is also maybe while organic clicks could flatten or even dip. But conversion rate per organic session goes up, right? That last pair right there is that fewer but better qualified trade that we talked about showing up in your own data. So I know this is a lot. We are all still figuring out a lot of this ourselves, right? Victorious, we are researching it every single quarter and we are testing which tactics perform better in AI Mode versus all the other things. The Q2 search report on our site, I would say also for more reading, has industry by industry breakdown. If you want to go any deeper on that. So I would say if you take one thing from today's episode, I would make it this don't wait for AI mode's usage numbers to become undeniable before you act, because everything Google is shipping tells us and tells me where this is headed, right? And the foundations that you are already building in your broader search strategy should be carrying over, right? That is the good news. But what is new about this is the AI Mode work, right? The strengthening of the Google owned assets that AI Mode leans on. And also getting your prompt tracking and baseline performance in place right now. So that when Google makes inevitably its next move, you can see it in your own data while everybody else is out there debating whether or not they should be adopting a new AI Mode search strategy. So that's today's episode as we talk through what's happening in Google search, how to future proof your search strategy as AI Mode is continuing to evolve. And in my opinion, will become the preeminent way in which the search experience happens on Google search today and moving forward. So thanks for listening to this episode of the Search Signal. If you have found this conversation at all helpful, please consider leaving a review for the podcast and hitting that follow button so you can find the next episode when we publish it. That's it for me. I will see you next week.