Accelerate Growth with Answer Engine Optimization (AEO)

Catch the key takeaways

In our latest webinar, Amsive’s SEO leaders explored the rapid rise of AI search and answer engine optimization (AEO)—and what it means for the future of digital visibility.

As search behavior evolves and large language models (LLMs) reshape how users find information, brands are navigating a new era where visibility depends on more than just rankings. For marketers, this shift presents both a challenge and a huge opportunity: adapt to the AI-driven search experience, or risk losing visibility.

Here are the top five takeaways from the conversation:

1. AEO expands on SEO

SEO still matters. In fact, many core SEO practices like technical health, structured data, and content clarity are foundational to AEO success. But answer engine optimization introduces new requirements: content must be optimized for how AI systems interpret, retrieve, and summarize answers.

That means shorter, context-rich snippets, conversational formatting, and entity-driven markup. SEO is a starting point, and AEO is an expansion of those core foundational principles.

2. LLM visibility hinges on your reputation across the web

AI engines are drawing from far more than your site. They’re citing Reddit threads, YouTube videos, news publishers, review sites, and social platforms like LinkedIn and Medium. That means your brand’s digital footprint must extend across the broader web.

Community engagement, digital PR, user-generated content, and social presence can now all influence and impact how discoverable your content is to AI search engines.

3. Technical SEO foundations matter more than ever

Unlike search engines, most LLMs can’t handle JavaScript-heavy pages or complex client-side rendering. For AEO success, content must be crawlable in raw HTML, structured with semantic markup, and easily digestible in snippet-sized chunks.

Speed, schema, server-side rendering, internal linking, and content clarity are essential for answer engine citations.

4. Tracking AI visibility requires new metrics

Legacy SEO metrics like rankings can’t measure your visibility in a zero-click, answer-first world. Tools like our partner Profound track brand visibility, share of answers, and sentiment across multiple LLMs.

These insights help brands understand where they’re showing up in AI answers—and where they’re missing opportunities. Forward-looking marketers are already redefining KPIs and investing in tools built for the AI era.

5. Prioritize scalable, high-impact tactics to being improving LLM visibility.

Start with what you can control. Focus on scalable, high-impact actions can create meaningful momentum. Begin by strengthening your site’s technical foundation—ensure key content is accessible in raw HTML, reduce reliance on JavaScript, and use semantic structure and schema markup to help AI engines interpret your pages.

Repurpose existing content across high-visibility platforms like LinkedIn, Reddit, and YouTube, where large language models frequently source information. Use GA4 to track referral traffic from AI engines like ChatGPT or Perplexity, monitor engagement, and identify where your brand already appears. Even small, strategic steps can compound into long-term visibility gains as AI search evolves.

FAQs: Answer Engine Optimization

What is Answer Engine Optimization (AEO)?

Answer Engine Optimization (AEO) is optimizing content so that AI-powered search engines like ChatGPT, Gemini, and Perplexity can retrieve, interpret, and summarize your content as a direct answer to user queries. It focuses on snippet-friendly formatting, structured data, and broad digital visibility beyond your website.

Is AEO replacing SEO?

No. AEO is an expansion of SEO, not a replacement. SEO best practices like technical health, internal linking, and structured data are foundational to AEO success. AEO builds on these by ensuring your content is accessible, conversational, and optimized for how large language models process information.

How is AI search changing user behavior?

AI search engines like ChatGPT and Google’s AI Overviews are changing how users access information by delivering instant answers, often without a click-through on an organic search result. For that reason, brands should optimize content to increase the likelihood of appearing in these AI-generated responses, not just traditional search rankings.

Where do AI engines get their information?

LLMs like ChatGPT and Perplexity pull information from:

• Pre-trained datasets (e.g., Common Crawl, Wikipedia, Reddit)
• Search engines (Google, Bing)
• External sources via retrieval-augmented generation (RAG)
• Social and user-generated content platforms (LinkedIn, YouTube, Reddit)

What technical elements are important for AEO?

To be AEO-ready:

• Ensure key content is rendered in raw HTML
• Avoid over-reliance on JavaScript
• Use semantic HTML5 structure (e.g., <header>, <main>)
• Implement detailed schema markup
• Use fast page speed, server-side rendering, and clean internal linking

What is the ideal content structure for AI visibility?

Content should:

• Be organized into short, context-rich snippets (~160 characters)
• Include clear subject-verb-object sentences
• Use conversational formatting
• Front-load key entities, data, and takeaways
• Leverage FAQs, tables, and bulleted summaries

How do I track my brand’s visibility in AI engines?

Tools like Profound can help track:

• Brand mentions across LLMs
• Sentiment analysis
• Citation sources (e.g., Reddit, YouTube, Forbes)
• Share of voice across ChatGPT, Gemini, Perplexity, etc.

In GA4, use custom channel grouping to track referral traffic from AI sources like Perplexity or ChatGPT.

What platforms matter most for AI visibility?

Beyond your website, a few key resources that AI engines frequently cite are:

• Reddit
• LinkedIn
• YouTube
• Medium

Optimizing your presence across these platforms is crucial for AEO.

Dive into the transcript

Introduction

Lauren Welles Medley (00:00:19):

Alright. Welcome everybody. Thanks for joining our webinar today. We’re going to be discussing the new AI search landscape, new evolutions and changes there, and answer engine optimization. I’m Lauren Welles Medley. I’m the Senior Channel Director here at Amsive, and I’m joined today by my esteemed colleagues and two of my favorite people on the planet, Lily Ray and Romain Damery. Say hi, guys.

Romain Damery (00:00:47):

Hi everyone.

Lily Ray (00:00:49):

Hi, everyone. Excited to be here.

Lauren Welles Medley (00:00:52):

Alright, so let’s address the elephant in the room, which is there’s a ton of noise happening in the SEO space around AI search. It’s sort of flooding all platforms. It’s what you’re living in every single day, Lily, and the rise of AEO or GEO or LLMO, whatever you want to call it, acronym of your choice. So we’re here today to really level set on what’s happening in the landscape and how we’re approaching that evolution here at Amsive.

For those of you who don’t know, we started as a search-first agency and have about 20 years of SEO expertise here at Amsive. Collectively, the three of us have been here for about more than half that time, which is pretty wild. So we’re approaching this search evolution very thoughtfully as we have for all of the major SEO and search changes over the last two decades, which there have been very, very many.

We’ll get into some real data and what we know about AI search and where we’re going, we are sure that there will be lots and lots of questions, so feel free to ask them in the Q and A and we’ll do our best to answer as many as possible. At the end, we’re going to try to say about 10 to 15 minutes, you’ll also get a recording of this webinar and other materials as a follow-up. And of course, you can always reach us@amsive.com to learn more or connect with us directly.

So let’s jump in. What we want you to take away today, some key learnings about what’s changing, some realistic guidance around how to integrate AEO into your existing strategies and some exclusive industry data from our partnership with Profound, which we’re really excited about. So we’re going to start with you, Lily, are AI engines taking over?

The AI search engines like ChatGPT and Gemini are changing search

Lily Ray (00:02:53):

For sure. Thanks, Lauren for the intro. Yeah, so I know there’s a lot of questions right now in the digital marketing world about the rise of AI and large language models and how this impacts SEO and search and everything like that. So we’re going to dive deep today and answer some of these questions. And of course, it’s true that a lot of people are beginning to start their searches on large language models like ChatGPT, people are also using large language models for a lot of different purposes, other than just searching.

But as far as it relates to the growth of usage of different large language models over time, you can see that ChatGPT is clearly outpacing the other large language models. We had a little uptick when Deep Seek was announced in January, and we also see that Gemini is kind of slowly but steadily gaining market share over time.

But ChatGPT continues to be the biggest market share leader compared to the other large language models. But if you go to the next slide, Lauren, the big kind of main takeaway that we always want to talk about in the SEO space is that the elephant in the room is still Google. Google is still driving approximately 14 billion searches per day, and that number’s increasing at a massive rate over time. So if you look at this chart that was pulled, I believe in late 2024, you can see the number of searches that happened per day on Google compared to GPT.

It’s still kind of a drop in the bucket for ChatGPT. Now, it is important to note that ChatGPT is growing rapidly. So I think that its usage has doubled since these numbers were pulled, but compared to Google and compared to the other search engines, it’s still small. It’s definitely nowhere near Google, and Google is growing just as quickly. And it’s also introducing its own large language models and AI products that are also according to Google, really increasing usage of their products as well.

Lauren Welles Medley (00:04:42):

Yeah, that was exactly what I was going to ask next, is there’s some changes directly on Google that we’re paying really close attention to and are already sort of feeling more of the impacts of already. So you want to talk a little bit about how AI Mode and AI Overviews fits into all of this?

Lily Ray (00:05:03):

Yeah, absolutely. So for anybody who’s unfamiliar, Google’s AI Overviews and AI Mode are kind of their new large language model products that they’re building directly into the search interface. So as it relates to AI Overviews that’s been live on Google for about a year now. They’ve rolled it out across many countries around the world. And basically what it’s doing is it’s using Google’s AI technology powered by Gemini 2.5 to answer questions directly on the search results, kind of summarize what the top search results say about that query.

And then with AI Mode, this is a newer product. It’s actually only been publicly available in the US for about one or two months now, and it just became public in India yesterday. They’re really starting to roll this out around the world. It’s a new tab in Google search. So when you go directly to Google, if you’re in the us, you can click on AI Mode on the far left. And this is also a new language model powered by Gemini that answers questions directly in the search results. And it’s much more conversational. It has much more robust, kind of thorough answers to questions.

And it also kind of pulls in answers and data from different Google products like Google Maps, Google Shopping, to give the user directly what they need within AI Mode

Lauren Welles Medley (00:06:16):

And now ads, right?

Lily Ray (00:06:18):

Yes. So that’s been a new announcement from Google as well. This is still kind of rolling out. So with AI Mode, it’s still experimental. It’s not entirely live yet, but we are starting to see Google place ads in AI Overviews, which, of course, makes sense. This is the main way that Google makes money, and we spoke to our paid search team a little bit about this just to understand the requirements and they say it requires Performance Max or AI Max to be eligible to appear in AI Overview ads, and you must be using broad match keywords to appear there.

Lauren Welles Medley (00:06:48):

Our paid search team is going nuts over this, and that’s actually going to be our next upcoming webinar, I believe, on July 23rd. Is that right?

Lily Ray (00:07:02):

Yeah, exactly.

How AI Overviews are affecting SEO click-through rates

Lily Ray (00:07:14):

So yes, this is definitely one of the biggest questions in this space right now. How are these new AI products affecting click-through rates within organic search within Google specifically? And we’ve done a study at Amsive by our SEO teammate, Will Guevara, who looked specifically at some first-party data to understand click-through rates by industry. And it’s true that we are seeing click-through rates decline across many categories when AI Overviews are present. And this is something that’s been echoed across many different studies throughout the SEO industry, particularly as it relates to what we call informational query leak answered by AI.

Google’s generally going to show an AI Overview for that query, and AI Overviews do tend to result in fewer clicks, as users naturally are getting the answers to their questions directly on the search results. And this is another study on the right here that we pulled from Spark Toro from Rand Fishkin and what they showed. I think the main thing that we want you all to focus on here is that lower dark blue bar, how that kind of paces over time, because that’s the percentage of searches that end with no clicks on Google. So I think the main takeaway to look at here is that that blue bar really jumps between February and March, and that’s because Google’s showing more AI Overviews.

So, more people are just kind of staying within Google’s environment as opposed to clicking on external links when they have those AI Overviews present.

Romain Damery (00:08:42):

And the end goal is kind of for AI Mode to become the default, right? Google have hinted at that maybe a few years from now, that’s probably what’s going to happen.

Lily Ray (00:08:52):

So that’s right. Google has been on record over the past couple of months literally saying AI Mode is the future of search. They said that at the Google I/O keynote, and they’ve also said that the things that users like the most about AI Mode, while it’s kind of in this testing phase, are the things that they’ll plan to kind of incorporate into basically the default version of Google search going forward.

Lauren Welles Medley (00:09:17):

So I think that that’s really why it’s so important for brands to be paying attention to this and evolving their strategies and their approaches to SEO to incorporate AEO and pay attention to their visibility here because it’s not going anywhere. That’s pretty obvious. It’s only continuing to climb, and we are seeing that adoption increase pretty rapidly across AI engines, but really where we’re seeing that immediate effect now is directly within Google itself. So shifting how we’re thinking about things and tracking things is really, really important. And this is a new space for everyone.

So, early adopters who are starting to pay attention and track their visibility across these various different platforms and are willing to test and learn from new strategies and tactics are going to be the ones that win early and remain sticky. Would you guys agree with that?

Romain Damery (00:10:25):

Yeah, for sure. And they’ll have a first mover advantage, right? It’s just going to just get more competitive. So now is a great time to really get serious about optimizing for AI chat bots for sure.

AEO vs. SEO: What’s the difference?

Lauren Welles Medley (00:10:38):

But AEO isn’t replacing SEO, is it?

Lily Ray (00:10:43):

No, it’s not replacing SEO. AEO is an evolution of SEO. It’s a new way of thinking about SEO. And one thing we really want to convey today is that a lot of the work that we do in the SEO world is consistent with the work that we’re going to be doing to appear in large language models.

Lauren Welles Medley (00:11:04):

So I think we want to look at, yeah, I think we want to look at what is fundamentally those areas of overlap versus where we see key differences.

Lily Ray (00:11:16):

For sure. And it’s kind of fluid, it’s kind of evolving over time as we start to understand a little bit more about how large language models work and surface content and websites and everything. But we did want to hammer home that, again, there’s a lot of work that we’re doing on the SEO front, like focusing on keywords, focusing on crawling and indexing, page speed, core web vitals, technical considerations like XML site maps and crawl budgets.

And on the far right, on the AEO side, things have really shifted where we’re much more concerned with conversational search. How are people prompting these large language models? What snippets and conversational kind of responses to questions are these large language models most likely to retrieve from our websites and what are they saying about our brands?

And even if people aren’t necessarily clicking on some of the results, it’s really important to be part of that answer to make sure that your brand and your entities are kind of mentioned as part of that answer. We also have fan-out queries, which we’ll talk about, but go ahead, Romain.

Romain Damery (00:12:15):

Yeah, I was going to say some of them are not necessarily really exclusive, absolutely exclusive either. We’re saying core web vitals, that’s typically SEO, like Google said they’re using that as an official ranking factor. It doesn’t mean you can have a slow website for a o probability and rendering the rendering piece as well. That’s common to both SEO and AEO, but we’ve seen that in a lot of cases, not all element chatbots are able to execute JavaScript.

So it’s even more important with Google. It’s important, especially if you’re like a news website, or publisher websites, or even an eCommerce website, but if you want to play in AEO, you have to make sure that you have some pre-rendering and just don’t rely heavily on JavaScript and the client site rendering. And that’s true for even some of the LLMS that claim that they do execute JavaScript. It doesn’t always work.

Sometimes they can, they want to, and sometimes they just refuse to. So we’ve done some test,s and even within the same LLM themselves, we get different results sometimes. So that’s really something that you don’t want to risk. So sometimes we have shared factors in the optimization strategy, but they’re weighed differently, if that makes sense as well.

Lauren Welles Medley (00:13:44):

And I would say even some of the new AEO tactics also work in reverse and are beneficial for SEO, still as well. We talk about how content should be more conversational because you are having these interpersonal conversations with these AI engines and therefore you should be structuring your content in the same way. But in the traditional SEO space, we’re seeing those conversational-type queries translate over to traditional search because of this behavior change as well. So I think it is very much, even though we have designated quite a bit of directly lands in that overlap gap, it’s even some of the items that fall on either side that do actually kind of still support one another.

Lily Ray (00:14:37):

And one last thing I want to add here, too, is that these large language models, in many cases, are pulling from search results. So being visible in search results is essential for visibility in the main large language models, right? Perplexity looking at Google and ChatGPT is pulling from Bing. I’ve seen some links show up in ChatGPT that are obviously from Google. You have to be in the search engines invisible in the search engines to appear when chatbots are basically accessing search results.

Lauren Welles Medley (00:15:07):

Yeah, I forget the percentage off the top of my head, but isn’t it like if you rank on page one, the likelihood that you appear in the AI Overview is 25% or something like that, right?

Lily Ray (00:15:20):

If you rank position one on Google, it’s 25%.

Why your brand’s reputation matters in AI Search

Lauren Welles Medley (00:15:21):

Position one, 25%. Yep. Alright. So Romain, I want to pass this one to you, which is AI engines as similar as they are in how we approach optimizing between AEO and SEO, AI engines are fundamentally different than search engines and how they gather information. Can you talk a little bit about that?

Romain Damery (00:15:51):

Yeah, sure. So just like said, there’s also some overlap there always, but one of the important differences between SEO and AEO is really understanding how the chat bots fundamentally work and source their information. So that’s giving us additional insights as to how we might be optimizing for them.

So while the search engines like Google and Bing generally they use their quote live index to share results. The chatbots are either looking within their own pre-trained and fine-tune datasets or real-time external data sources, or usually a combination of both through a process that’s known as retrieval, augmented generation or RAG. So on the one hand, like I said, there’s the LLM’s own core and pre-trained data here you can see on the graph for an example, for GPT3, which is the acquired the static knowledge without connecting to the internet, where most of the content discoverability for LLMS actually come from. So they come from massive data sets.

So a huge one is the data from Common Crawl, which is an organization that crawl the web and free provides its archives and data sets to the public. And that’s estimated to be as large as 62% here in the case of GBT three, they’re also sourcing from scrape resource heavy sites, which can be structured or not like Wikipedia or large books, data sets like bookkeepers or lip gen for scientific papers, but also forums like Reddit’s or Stock Exchange that end up in different data sets themselves called Web Text.

So that’s their base, that’s their core index if you will. And then on the other hand, we have external live sources that are leveraged through RAG, which is a technique that enables the LLMS to retrieve and incorporate new information. So with it, the LLMs, they don’t respond to user queries until they refer to a specific set of documents if they feel like the query or the prompt deserves some freshness, for instance, or if the LLM doesn’t know, or if there’s ambiguity.

So it’s a more dynamic process, and that includes accessing, like Lily said, like search results. So Google, in the case of Gemini, or Bing, in the case of GPT, or specific APIs, or other external databases that they may or may not have license rights to as well. And so accessing those as necessary to retrieve more relevant and up-to-date information, but also ground, as we said, the AI-generated answer to reduce the hallucinations. And that’s especially when specific content strategies and technical SEO also come in. So that’s what we’re going to be getting into next.

How to build an AI-friendly content strategy

Lauren Welles Medley (00:19:13):

Yeah, so how does this change how we approach content, Lily?

Lily Ray (00:19:19):

Yeah, definitely. And to Romain’s point, one of the things that makes Google’s, especially AI Mode, its new products, really interesting, is they can tap into the different kind of data sets and information that Google has to answer your question based on what they think you’re really getting at when you ask that question in AI Mode.

So what Google’s doing is, according to them, they’re using what they call a query fan-out technique, which basically breaks down your question into multiple subtopics and issues, a multitude of queries simultaneously on your behalf. And that can come from any data source that Google has at its disposal. So that’s why here, you might see them tap into the knowledge graph, or into real-world locations, or into weather. And what they’re doing after that is they’re actually conducting numerous searches simultaneously based on what they think you’re kind of implying in that search result or what they know about you and your search history.

So they’re actually going to be able to personalize these results. They say that that’s going to be rolled out this summer, so they can say, based on this user’s Gmail history or their search history or their personalization, what we know about them, we can also anticipate the best possible answer for them. So traditional search kind of looked at one keyword and kind of map that to the most relevant results. Now with Query Fan out, it’s really anticipating what are the dozens of different queries that this user might have baked into that search and simultaneously answering all of them in one answer

Romain Damery (00:20:40):

Based on that intent and the follow-up queries that will come after.

Lauren Welles Medley (00:20:47):

I think what’s really interesting, and one of the things that I’m really excited about with this move towards AEO, is that it’s no longer just about your website. You also have to be considering how your brand appears across the entire web because oftentimes it’s not your website that’s going to be cited, it’s pulling from a million different sources to come up with the best possible answer. So I think we have some data here that Lily, I’d love for you to take us through around some of the top-sided domains, across these different platforms.

Lily Ray (00:21:29):

Yeah, absolutely. And we mentioned before that we’re partnering with Profound, Profound is doing millions or hundreds of thousands of searches every day across different large language models and collecting the data to understand who’s being cited the most and the answers who’s being linked to the most. So this is a visualization of what they’re seeing with all of their tracking to understand who are the top-sided domains across the different large language models.

So with ChatGPT, we see that 47% of the time it’s pulling from Wikipedia, followed by Reddit and Forbes, and G2 with Google’s AI Overviews. They’re predominantly actually pulling from Reddit, Forbes, and YouTube and Quora.

So a lot of user-generated content there. And Perplexity is also pulling from Reddit. So, a couple of these companies are actually partnering directly with Reddit, which is why we’re starting to see a lot more Reddit answers and Reddit conversations appearing in AI responses.

But it also speaks to what you mentioned earlier, Lauren, it’s about what a lot of people are saying about a lot of things across the internet. And I think that these tools are very focused on that user generated content because at the end of the day, what they’re trying to do is make these chatbots conversational, make them feel like human and make them kind of answer questions the way that a human might answer, and where do they get that content?

They’re getting it from Reddit, they’re getting it from Quora and YouTube and LinkedIn. So it’s really interesting to see how these chatbots are really skewing towards pulling information from user-generated content.

Lauren Welles Medley (00:22:49):

So we have to think about what our strategy is across all of these different platforms and channels, which again is what I’m so excited about this. We are collaborating here at Amsive with our social team on strategic playbooks for managing Reddit in a way that accomplishes both brand awareness and brand building, but also how do we rank in those forums and discussions results where Reddit is everywhere in organic search. So it sort of serves two purposes, which is pretty cool.

But brand building is obviously huge, and that really goes, this is where AEO goes beyond SEO, and we partner with all of our other channels here at Amsive to accomplish these things. So, like, upper-funnel, paid media efforts to drive up brand demand, that community management that I just mentioned, digital PR is going to be huge. It has never gone away. For SEO, obviously, developing that authority and that strong backlink profile is really critical.

But again, these LLMs are looking for your brand to be the source. So the more places you are cited across the web, the better chances you’re going to have to appear in them. Now we’re going to jump pretty quickly into some really specific and tactical recommendations. I’m going to toss it over to you, Romain, if you can walk us through some of the technical foundations that remain really critical or some new things that are really important for LLM visibility.

Technical strategies for AI Visibility

Romain Damery (00:24:33):

Yeah, so here again, a lot of shared best practices with SEO, but even more important for AEO, the goal is really to make sure that your content is very clear and is easy to be understood and read by bots. So the content has to be found really ideally in the raw, the source HTML, so you don’t have to rely on JavaScript to retrieve that content. So think about optimization strategies like server-side rendering or using static site generators, but also the way that you build your pages beyond that.

So think about some of the typical issues, technical issues that you have with SEO, that’s even having to click on a read more button to see the rest of the content, or collapsible elements that may be requiring JavaScript through Ajax calls to actually retrieve the content itself. That’s really not going to work well with AI chatbots.

You also need to make sure that your content is marked up with schema, but also that the information that you are marking up, just like for SEO, is found on the page, the rendering of the page itself, right? That’s still pretty cool. JSN LD is the easiest one to implement and the official preferred format for schema, and we don’t use just general labels like thing or webpage. The more precise you are, the easier it is for such engines and AI tools to understand what your page is actually about. So make sure that as you are using product schema, including images, prices, reviews, and ratings, that’s still very important.

FAQ page is another big one, especially for AEO, how-to schema as well. But also events or software applications, local businesses, all of those stay critical. There’s also no harm in using speakable markup to identify sections within an article or webpage that’s best suited for audio playback if you want to be picked up by AI assistance or things like that. And also, if you’re in eCommerce, particularly, don’t forget your Google Merchant Center product feeds as well, especially for Gemini.

Lauren Welles Medley (00:27:04):

You have to sign up.

Romain Damery (00:27:06):

For Open AI, yeah, and Perplexity. You have forms where you can opt in and they’ll let us know when it opens up, but, so, we have yet to hear confirmation from them, but there are opt-in mechanisms already. For product feeds, strengthen your personal organization entity, so that’s using the same as properties pointing to your LinkedIn, your Wikipedia profiles, things like that. It’s very, very important using semantic HTML5 elements for the template structure itself as well.

So that’s like header nav, main section aside, footer, the structure of the template of the page. That’s really even more important for the chatbots. And then the regular ones that we have for SEO for the structure of the content, like header tags with a logical hierarchy. If you have data comparison, make sure that they’re in proper HTML tables that you’re using proper tags for lists and other content attributes like date published or date modified LLMs. And chatbots also really want to sell fresh content, so make sure that you are not forgetting those.

And then a good internal linking structure and clustering your topics properly. That’s even more important for LLMS and chatbots because they don’t crawl the way that Google Bot or chatbots do. And that’s why on the first diagram, you could see XML site maps. That’s more of an SEO thing. The LLMS themselves, they don’t crawl sites like that or discover content in this way. So the internal linking structure and clustering your topics properly is going to be key as well.

Page group times, not just for Google, especially if you’re requiring JavaScript, the chatbots, they don’t have as many resources as Google to wait around for your content to load. And also, it’s even more important to speed is even more important than ever because for conversion rate optimization, really, you don’t want to waste the reduced amount of traffic that you’re going to be getting.

So that’s still key. And then obviously don’t block the known crawlers and the chatbot user agents, but also include things like the CC bot or things like that. That’s what the Common Crawl is using. So there’s also these ones, don’t forget about those. And obviously robust TXT file checking, your CDN, your web application firewalls for bot protection rules, and so on, so.

Lauren Welles Medley (00:29:59):

We saw this a lot early on, I think because people were maybe scared, didn’t want AI crawlers to be scraping their content.

Romain Damery (00:30:11):

With the robots specifically. But also, people may not know that their CDN provider may be blocking them as well. So it’s good to at least know what’s going on and monitor that so you make sure that your content is accessible to at least the big, legit ones. And then bonus points, if your HTML can pass the bonus three C standards, because I know it could be a lot of work, but it can save you a lot of headaches down the road. So just wanted to mention that as well.

Content structure and optimization tactics for answer engine visibility

Lauren Welles Medley (00:30:47):

So Lily, Romain covered some HTML structure that is good, required, but also nice to have to make these things, make your content more digestible. But are there other content structures that brands should be considering as they’re looking to update their content?

Lily Ray (00:31:08):

Yeah, definitely. So this is something that we’re all kind of digging into as we really start to understand how these tools are pulling in relevant snippets. And I think that word snippets is kind of one of the most important takeaways, is that these tools are basically pulling in short passages or they’re analyzing short passages from the page in order to look for a response to the question.

And you can even see with Google’s AI Overviews, in many cases, they’ll append the kind of block of text that they’re pulling from directly into the URL. And when you click on that URL or click on Google, you’ll be taken specifically to that block of text and be highlighted where they’re actually pulling it from. So it’s really essential to pay attention to what those snippets are. There’s also some kind of new research in this space.

There’s a man named Dan Petrovic who does a lot of testing, and he has pulled in some data specifically from Gemini that shows that Google’s AI Overviews and AI Mode are really looking at an 160-character block of text to kind of look for the answer to that question.

So, it’s really important if you’re trying to target an AI answer specifically to look at what are those snippets that they’re pulling in from your competitors, what are the facts and entities and everything that are mentioned in that answer, and can you kind of structure your own content to match that or even be better than that?

So, optimizing the key message within 160 characters, front-load, your strongest value proposition, names and numbers, and everything early in the headings. You want to have complete, context-rich sentences that communicate the benefits of your brand. So even if it’s truncated, the snippet still remains meaningful. Romain spoke a lot about semantic and structural HTML, but we’re seeing that plays a big role as well. Even sometimes bolding the key entities or the key answer to the question.

And then using keyword-rich, customer-focused phrasing. So high-density, relevant keywords as close to the beginning of the snippet as possible, thinking about writing in the second term, in a second person, when it makes sense to do so, your team, or these are the benefits that you’ll get. And then we’ve talked about this a lot in the SEO space over the years, but kind of thinking about natural language processing and using what we call semantic triples. So basically, subject-predicate-object, very clear sentences, so the man went to the store, like reducing ambiguity that really seems to be rewarded in AI answers.

Lauren Welles Medley (00:33:25):

What about content strategies overall? This is something that you talk about all the time, and I think that it’s a philosophy that carries over from really good SEO that’s rooted in E-E-A-T. So I think that this is a really important piece.

Lily Ray (00:33:43):

Yeah, definitely. This is actually a visual pulled from a number of different conference presentations that I’ve done this year because this has really been, I would say, one of the biggest takeaways in the SEO space this year. Not just for AI search, but for SEO in general.

Taking original content that you’ve produced, let’s say survey data or statistics or something that you have that’s kind of first party to your company, and then repurposing that into many different formats is a really great way to be visible both in search engines and AI search. Again, a lot of the time, they’re pulling from these more conversational and user-generated content-type sites. So the more that you can take that content, repurpose it into many different formats. We also know that Google’s AI Mode is actually able to digest information from different, what we call multimodal formats.

So video, audio, podcasts, images, everything like that. So, taking that original content and going out there and doing podcasts about it, they can use YouTube videos and transcripts and podcasts to kind of factor into the results, making sure that it’s talked about on Reddit, having supporting visual references that can help to show up in Google Lens, for example.

Perhaps you can do presentations about it and upload your presentation slides to something like SlideShare. You can do a long video that’s published to something like YouTube and then splice it up and publish it on TikTok and YouTube shorts, and Instagram reels. And then, of course, just kind of echoing that throughout social media. So, large language models are pulling a lot from LinkedIn. They’re pulling a lot from Medium and Hub pages and X and BlueSky. So, the more that you can kind of put that information out there and have people talking about it, the more chances you have to appear in AI responses.

Lauren Welles Medley (00:35:22):

This touches on all the things that we love. It is, I think, about citations and the likelihood to be cited in the same way that we have thought about SEO content marketing for a long time. That’s rooted in developing content that’s shareable, that’s going to be linked to as a source for different data. So that original research is really important, but then it also touches on what we were talking about before with the need for this integrated multichannel approach and really getting the most bang for your buck here. Let’s do more with less. Let’s do the research once and repurpose it across the web in all different formats.

Lily Ray (00:36:09):

Absolutely.

How AI search is changing local SEO and Google Business Profiles

Lauren Welles Medley (00:36:12):

Something that I thought was interesting, as we’re digging into AI Mode, goes back to Google wants to keep you on Google, but sort of emphasizes some elements of SEO that you wouldn’t necessarily think are as important as they actually are.

Lily Ray (00:36:31):

Yeah, we were a bit surprised, I would say, to see this, the extent to which Google is really pulling from Google Business Profile and Google Maps specifically in its AI Mode, results, and responses. So what that means is that, for example, here I typed best companies for life insurance, and Google’s AI Mode is recommending a handful of different companies, but that dotted underline means that they’re not linking to an external website. It’s not linking directly to those companies’ websites. It’s instead embedding a Google Maps result for that business and also geolocating it to where I’m searching from.

So, for example, MassMutual Financial Group was recommended as a great insurance company, but in many cases, it’s going to pull in the closest Google Business Profile result to the searcher. So I’m actually seeing Google get this wrong. AI Mode is still, it’s still improving. So sometimes they might get it wrong, they might actually pull some business with a similar name that’s kind of geolocated to where the searcher is searching from.

But I would say the main takeaway here, and something that we’re excited about at Amsive because we have focused so heavily on local SEO for many years now, is making sure you’re putting your best foot forward with Google Business Profile optimizations, making sure that the pictures look good, making sure that you’re building out your reviews for different locations. Because even if you think AI Mode is going to show your corporate headquarters, it actually might link to some practitioner’s office that’s closer to where the search is searching from. So you need to make sure you’re buttoned up across all your Google Business Profile results.

How to measure AEO success and track AI visibility

Lauren Welles Medley (00:38:08):

So I want to shift gears a little bit and get into tracking and measurement. Now, obviously, we’re seeing huge shifts in where people are searching, how they’re searching, and how they’re engaging with different platforms, and I think that it’s really important for us to talk about how we need to evolve from a tracking and measurement perspective. And there’s a lot of new technology coming out, right? Lily, you’ve kind of dabbled in all of these. I know we’ve landed on Profound as our partner, which we’re really excited about, but there’s a lot happening here.

Lily Ray (00:38:42):

Yeah, we’re using numerous tools. We’re really excited about Profound, but we are using a few of the other tools shown here, and there are just so many new tools in this space as more and more people try to basically emulate what a real person might see in their experience using a large language model, which is at the end of the day, impossible to do perfectly because by definition, large language models provide slightly different answers to questions depending on context and history and everything like that.

But what these tools are trying to do is really emulate the general kind of consensus-driven answers that you might get to questions about your brands, your products, or services. So they’re conducting many, many different prompts and collecting the answers throughout the day, and kind of counting and aggregating how frequently brands are mentioned there, what is the context in which their products and services are being mentioned, what is the sentiment around those products and services?

So I would say one of the main takeaways we want to get across today is that this is really what you need to be paying attention to as far as AEO goes, because if you’re just doing traditional rank tracking and just looking at Google results and Bing results the way that we have in the SEO space for 25 plus years, you’re not going to capture how your brands are appearing in large language model responses. So we’re really excited to be using some of these tools to get that data out there and update our reporting for our clients.

Lauren Welles Medley (00:40:01):

And the traffic isn’t necessarily going to be there anymore.

Lily Ray (00:40:05):

Yeah.

Lauren Welles Medley (00:40:06):

It’s being mentioned, right?

Romain Damery (00:40:09):

Yeah, there can be some surprises as well. You think you’re doing great in SEO in search and organic search, and then you find out the hard way that you don’t have that much visibility in LLMs. Sometimes it can be very different based on the LLMs; you don’t have the same visibility across all of them. So yeah, definitely make sure you’re tracking that across the,

Lauren Welles Medley (00:40:32):

We can get some data now out of legacy tools. So, traffic and conversions.

Romain Damery (00:40:42):

Besides the AI visibility tracking platforms right now, at the very least, it’s critical that marketing teams set up custom channel reporting in GA4 or Adobe or whatever analytics platform they’re using to track the traffic that’s coming from those sources and understand how users are already finding them and also how they’re not finding them. And that can really give you an idea of what LLMs you already have some visibility for and which ones you should spend time on. Yeah, so most importantly, tracking conversions and revenue from referral sources.

That’s just a few stats from some of our clients in those two verticals, where we found that conversion rates can be much higher, actually, than for organic search. If you think about it, they’re doing so much like top-of-the-funnel research ahead of time, ahead of clicking and arriving on your website. So it’s not uncommon to see higher conversion rates, but also that means you need to make sure that your site is fast, that you also always have some ongoing CRO IB testing going on. You really need to make the most out of the traffic that you’re going to be getting now.

Lauren Welles Medley (00:42:20):

Yeah, I think this is also a point of why to pay attention to this now, because it’s clear that even if the traffic volume is slow, it potentially could be that you have to look at your own data, but it could be converting at an even higher rate than organic search. So there could be untapped potential if you sort of shift and pay attention to what’s happening there and how you can get more out of that. But we do still have some sort of gray areas in terms of tracking, as Google looks to do for us. You want to talk about that?

Lily Ray (00:42:52):

Yeah, definitely. So this has been kind of an ever-evolving conversation in the SEO space, but where we’ve landed as far as reporting directly from Google as it relates to AI Overviews and AI Mode is that we do get the impression data from these products in Google Search Console. Unfortunately, we don’t get any type of breakout that indicates that something comes from AI Mode specifically, or something comes from AI Overview specifically. So it’s grouped into our larger search report in Google Search Console.

A lot of people have been asking Google many times if we can have more granular data, but it doesn’t seem like we’re going to be getting that. So Google does have some documentation if you want to understand how the positions are tracked within AIO and AI Mode, because there are some nuances as far as in AI Overviews, the upper-right list of links are all tracked as position one in Search Console. So, if you really want to understand that, I would recommend reading Google’s guidance. But the main takeaway here is that we are seeing that traffic appear, we just don’t necessarily know which search product it’s coming from.

Lauren Welles Medley (00:43:56):

Yeah. Alright, so I want to get us into some of the specifics. We got some really great data from our partner Profound related to specific industries. So I want to get through that. But Romain, real quick, you want to give everybody an overview of Profound in our partnership with them and how we’re going to be using their platform for our clients?

Romain Damery (00:44:22):

Yeah, real quick, that’s definitely our AI visibility platform of choice where we track brand products, services, everything across four or five different large language models simultaneously, even internationally. If you’re doing business outside of the us, that’s not a problem at all. And yeah, it gives us a lot of insights into the platform-specific performance that are really crucial for an AEO strategy development, especially considering that, as you’ve seen they may, each LLM platform kind of pulls from slightly different sources. So you need to be tracking all at once if you want to be successful.

AI visibility trends by industry

Lauren Welles Medley (00:45:15):

Cool. So let’s get into some of the data that they were able to graciously provide us with. First up, we have a look at banks and credit unions is the vertical.

Lily Ray (00:45:28):

So what I’m going to be doing here is just showing some of the types of insights that we are working with as it relates to Pofound data. So this is kind of high-level information about share of voice and visibility within the category of what we’re looking at, is combining visibility across four different large language models. I believe it’s AI Overviews, AI, five large language models, AI Overviews, AI Mode, Perplexity, ChatGPT, and Copilot. Those are the five that we’re looking at. So it’s kind of aggregating how visible different how many times these different brands are mentioned across those large language models for thousands of queries and prompts that Profound is looking at every day. So here, this is one visualization that we have that looks at the trends over time as far as which banks and credit unions are the most visible in this space.

Finance

And you can see that Bank of America is the market leader, followed by SoFi and then LightStream, and Capital One. But this is just one of many of the reports that we’re kind of paying attention to as far as who’s being cited the most within each category. Another view that we look at within Profound is sentiment. So what it can tell us is not only who’s the most visible, but as it relates to that company, what are the different attributes that people might be talking about as it relates to their products and services and their offerings, and what’s the sentiment around those products and services?

So we can kind of work with our clients to say, Hey, people really love that you have no fees. They love that you have a lower fixed APR, but a lot of people are struggling with the lack of personal loans or no unsecured personal loans. And that’s where we can kind of do some kind of proactive messaging and marketing to make sure that we’re really hitting on those pain points that we’re seeing come up in AI responses related to our brand.

Auto insurance

So, for auto insurance, another thing that we’re looking at in Profound is who is being cited the most in the answers? Because this is actually kind of the source of truth for how chatbots are getting the information that’s eventually funneling into their answers. So if we know that US News and Forbes and NerdWallet, and insurance.com are the most heavily cited pages within the auto insurance space, that’s kind of a digital PR play we can think about, how can we make sure that we’re mentioned on those heavily cited sites?

What does that mean from a PR perspective? If it’s eCommerce, what do we need to do to get our products in front of those publications? So this is kind of where the digital PR angle comes in to make sure that number one we’re being mentioned, and number two, our brands are being mentioned in a positive light on those sites.

This is something we thought was kind of interesting. So this is one automotive company, I won’t name names, but if you want to look at sentiment around some of the concerns that your audiences might be having with, let’s say, your automobiles. So maybe there’s production challenges, maybe there’s safety concerns, regulatory scrutiny.

This is what we want to talk to our clients about when we say, Hey, we see that you’re appearing in large language models, but people are having these types of questions and concerns, and there might be more work that you can do both on your website as far as along with your social media accounts and the ways that you communicate with your customers to reconcile and get ahead of some of those difficult conversations.

Health systems & hospitals

And with health systems and hospitals, we thought it was interesting here that the sites that are most heavily mentioned and cited in this category really echo and kind of emulate what we’re seeing in Google’s results over the years as well. So the more authoritative health institutions like the Mayo Clinic, and Cleveland Clinic, and Johns Hopkins Hospital, the content from those websites, chatbots are really leaning into that content.

So this is where we can start to see that they’re kind of using their own version of what we call E-E-A-T experience, expertise, authority, and trust to make sure that, as it relates to health content, they’re pulling from these trusted health and medical institutions, software, and SaaS. This is another visualization that we have in Profound, so we can look specifically at the different prompts. For example, in software and SaaS, it would be what accounting software integrates best with other business tools? And then you can basically see who’s being mentioned the most.

So here we have the black logos, QuickBooks, the blue one, Zero, the colorful one, Zoho, F is FreshBooks, and then the penguin is Wave. And you can see how they’re ranking across these different prompts that we’re monitoring over time, and that can give you a good clue about where do we need to market ourselves a little bit differently or get ahead of the conversation to make sure that we are as visible as our competitors across these different questions that people might be asking in large language models for digital publishing, this is another example where you see almost more mid-sized publishers that are really being cited heavily in the responses. So this is also a digital VR play where we can say, Hey, we’re a digital publishing company.

We want to be mentioned on the best online platforms when people search for that in chatbots, but it looks like ChatGPT and the other large language models are really leaning into this Indie Media Club page. So what can we do to make sure we’re present there and present across the other main cited most heavily cited citations within that space?

Lauren Welles Medley (00:50:21):

Barnacle SEO.

Retail & consumer goods

Lily Ray (00:50:23):

Same concept. Exactly. And then our last couple of examples. So retail and consumer goods, no big surprise, but Amazon is cited very heavily along with Walmart, Target, Costco, and Best Buy. And then you can also see the breakdown of the different topics people might be searching for in eCommerce, who are the most affordable options, like Target’s number one for that, best online shops for consumer goods delivery, Amazon’s best for that. So you kind of get this consensus around the different questions and different attributes that people might be asking about your products and services.

Higher education & EdTech

And then lastly, just another view of in the higher education and EdTech space, you can see who’s trending over time. And it’s interesting to see here. Harvard really started to take off in the last few days as the most visible site in this space. But again, these are very authoritative institutions.

Obviously, we’re looking at content from a lot of ID leads and major universities here. So it looks like large language models are also really tapping in to as it relates to higher ed and ed tech, more authoritative content from these trusted websites.

Final AEO takeaways: what you can do today, and what to watch for

Lauren Welles Medley (00:51:31):

Thanks, Lily. So I want to wrap us up real quick, and then we’ll open it up for questions. This is sort of a summary of takeaways of what we’ve talked about, but I think that really in order to develop a meaningful strategy around AEO and to evolve your current SEO strategies to align with these new platforms, you have to first be investing in new technology.

The KPIs around AEO are different, the measurement is different. So we need to first be monitoring website visibility and competition within those LLMs directly because your traffic data and your conversion data is only going to get you so far because oftentimes people are not going to click.

It’s just about your brand being visible. Don’t abandon your traditional SEO strategies just yet. The foundation is very much still rooted in those SEO best practices as we saw earlier with our Venn diagram.

There’s so much overlap that very much that foundational SEO still matters, adjust your content strategy, but again, don’t change it entirely. A lot of the new content structures and formats, and strategies around developing new content are also going to continue to help you with SEO. So, focus on those things that continue to matter while tweaking content here and there to be more aligned with what AI engines and AI bots are looking for. It’s really important to think beyond your website.

We talked a lot about integrated approaches and cross-channel strategies to build your brand and appear everywhere across the web. That’s a big difference, is that really AEO is going to be pulling that information from all different types of sources. So your brand has to be everywhere.

Track your traffic. You can do that. That’s something you can do literally today. We’ll have a follow-up resource for all of you after this and test and learn, because it’s all very new still, and it’s going to continue to change, and it’s going to continue to evolve. New tips and tricks are going to come out every day, and those who are willing to test and learn and evolve as this all changes are going to be the ones that continue to win. So I think we’ll open it up for questions.

Q & A

Leih Boyden (00:54:07):

The first question is, how can brands with smaller budgets tackle this new requirement to be omnipresent? How do they prioritize where to show up if they can’t invest in Reddit, YouTube, digital, pr, et cetera, all at once?

Lauren Welles Medley (00:54:25):

That’s a really good one. I have some initial thoughts. I see your wheel spinning too, Lily, but I think that this applies very directly to what we were talking about, focus on strategies where you can get the most bang for your buck. So if you’re developing content, develop content that can be repurposed in multiple different ways and across different platforms. I don’t know, Lily, if you had another thought.

Lily Ray (00:54:50):

Yeah, no, that was exactly what I was going to say. It really feels like, at least for now, large language models are leaning really heavily into a small handful of user-generated content sites. So I would say Reddit’s, generally speaking, a little bit harder to influence than the other ones because they have really strict moderation. So you do want to be careful about how you approach Reddit in particular.

Not that it’s impossible, there’s a lot of work you can do there, but as it relates to LinkedIn, Medium Hub pages, Instagram, you can take a piece of content and kind of repurpose it for those different platforms. You can even use AI to help you take a piece of content that you have and tell it to help me create different messaging for these different platforms, and it’ll help you kind of splice up what you already have, just kind of leaning on languages that are pulling heavily from is a good place to start.

Lauren Welles Medley (00:55:43):

Is there something from the tech side, Romain, that if you had to pick one thing? I know that’s such a hard question because it’s different for every site, but if you had to focus on one area

Romain Damery (00:55:58):

With limited budget?

Lauren Welles Medley (00:55:59):

With limited budget.

Romain Damery (00:56:05):

I mean, I think there’s so much you can do just by optimizing your site architecture, your internal linking your template structures, that is not too much of a lift. So I would definitely start there.

Leih Boyden (00:56:32):

So the next question is, how does a large strategy from a bigger brand work when they’re also trying to focus on small local markets where their businesses may be located? So it’s that balance of an overarching strategy that has a local presence.

Lily Ray (00:56:54):

Yeah, I mean, the way that, especially Google’s AI Mode is working is that it’s answers to questions, even when the user doesn’t necessarily say that they’re looking for a local result. So I’m in Brooklyn, New York. If I type best renter’s insurance company, it’s going to give me answers that are based in Brooklyn, New York. So even if you are a national company or multinational company, if you have those kind of localized business listings or that content that speaks to local audiences, a lot of the times the AI tools are going to pull that up for the user depending on where they’re searching from.

Lauren Welles Medley (00:57:31):

And if it’s a hugely multi-location business, the citation management tools and platforms, technology is critical to do all of that at scale.

Leih Boyden (00:57:49):

One final fast question is, as Reddit rises in importance, and we want the content to sound more conversational, how do you balance high E-E-A-T content that is high topical authority and you’re writing authoritatively while still providing content in a really human way, sometimes it can feel at odds, and how do you balance those two things? For our final question,

Lauren Welles Medley (00:58:15):

That’s probably a good one for you, Lily.

Lily Ray (00:58:17):

Yeah, I think so. Especially what Google’s doing, Nerdo idoing is supplementing that with Reddit because a lot of times people who let’s say have a medical condition might be going on Reddit and talking to others in their community that suffer from the same medical condition. So they’re trying to also showcase that information. So I would think of it more as supplemental to the work that you’re already doing to get high-quality, trustworthy information out there. But think about what are the conversations that real people are having, because there’s always going to be that mix of authoritative content and user-generated content and the answers.

Lauren Welles Medley (00:58:57):

And I think that there’s a place also for firsthand experience articles in those types of situations. So having that really highly authoritative, more like thought leadership type content, balanced with maybe individual customer stories, those sorts of things, alongside also the UGC that’s happening offsite as well. I think that was our last question.

So we’ll go ahead and wrap it up in our last minute. We just want to say thank you all for joining us. This is what we’re talking about every single day, so we’re excited to share our insights and approach with you guys. We’ll be sharing this webinar on demand and the materials via email. I sort of tease that there’ll be some other resources that we share out afterwards, but if you want to learn more or get in touch or talk about your search strategy, come talk to us. It’s what we do every day and we’re all very passionate about it. So thank you all for joining us and hope everyone has a good day.

Lily Ray (01:00:17):

Thanks everyone. Thanks.

The importance of paid media is ramping up in the digital landscape. Discover actionable insights for AI-powered paid media in our upcoming webinar with Google, or let’s talk about how Amsive can help you future-proof your marketing strategy.