How Do AI Engines Decide What to Cite? The FSA Framework Explained

Hey, real quick before we get into this, here's something that I've been noticing lately, and I'm seeing this with a lot of love. A lot of really smart marketers are jumping straight to the tools. They're on LinkedIn, they're asking, "Hey, what's the best say I've visibility tracker? Should we be using this tool or should we be using that one?" And these are all really good questions, but I'm seeing a pattern when we're reaching for the tools before we actually understand the fundamentals of how AI search works, like why some brands get cited and others don't, or what signals AI engines are actually looking for, or how citation and trust scores actually function. So I'm working to fix that. I'm in the planning stages of a free GEO course, a how-to GEO kind of thing, that walks you through the foundations before the tactics, because one to understand how it actually works, the tools start making sense. Without that foundation, you're really just buying a dashboard. The course is free, it's going to go live on YouTube and the podcast, and there's going to be a workbook to go with it, so that you can take notes and apply what you're learning. The newsletter is the first place I'm announcing it when it goes live, so sign up for the visibility report to get on the list, and you'll know the minute that it drops. Okay, to the episode. Have you ever looked at a chat GBT or Proplexity answer and thought, "Hey, why on earth is it deciding that random blog post? It's not mine." Well, if you have felt this frustration, you're not alone. Marketers some of this constantly read it as full of these questions, and everywhere you keep up with, like the happening to content marketing, well, it's the same story. Then here's the big secret, and nobody told us at the start. AI citations aren't random. They feel random because we don't really know the decision making process under the hood, and it's not likely these AI companies are going to tell us how it works. But, once you start looking at enough AI answers, the patterns become well kind of predictable, and this is exactly why the FSA framework exists. Hi, I'm Cassie Clark, fractional content strategist for early stage startups, and host of this podcast, "Fried an AI," where we help break down what's happening inside AI search, so you don't get lost in the middle of it. Today is a solo episode. It's just me, but we're talking about something super important, the FSA framework, which is a simple practical model for generative engine optimization that helps your brand show up inside of AI answers consistently, and sometimes almost instantly, if you're like me and count on eight-ish hour lag instant. We'll talk about that in a minute. So, if you've ever wondered why I sent a tiny website with a domain authority of four, get cited over industry giants, well, this episode is going to explain it based on what we know so far. Let's just dive right on into it. So, over the last several months, Gemini, ChaiPGPT, and Perplexity have kind of moved from this cute little conversational tool into like a full-blown discovery engine, and while that might not have been the intended use case at the start, that is exactly what is happening when users use these models. People use them the way that they use Google, so they get in there and they ask questions like, "Hey, what does this thing mean?" or "Compare these two tools, give me recommendations for, or my favorite that I use all the time." Explain this to me like a five. In marketers, they started noticing this when AI engines sent traffic to their websites, and Google was not. So, for a while, AI search optimization felt optional. It felt like a future thing, a, "Hey, we're going to keep an eye on this," and maybe figure it out later kind of thing. But then, on December 4th, 2025, Google quietly replaced the classic search button on their home page with a new default button, AI mode. And as of this recording, on Friday, December 5th, there has been no press release, no keynote, no hype cycle, nothing. It's just a subtle dodge into a new era of AI-driven search. Sure, you can still reach traditional Google search, but is it looking like it's going to be not the default option anymore? So, honestly, the message is pretty loud and it's pretty clear. AI Search is now the primary discovery experience. Traditional Search is just kind of the fallback at this point. This is the biggest shift in search behavior since 1998. I think I was maybe six, but even I know that over the last 25 years, nothing this big has happened. And it kind of underscores why the FSA framework is essential for content marketing, because SEO alone will not get you inside of those AI answers. So, let's call this out directly right now. Traditional SEO and AI Search are not the same thing. SEO helps you rank inside of Google's index and then absolutely still matters because people, particularly the older demographic, will still use traditional Google search. But the AI engines, they work differently. They don't rank or scan keywords and they don't reward clever metaphors or fluffy 800 word intros. These AI engines do three things consistently across all of the models. One, they synthesize meaning. Two, they extract usable chunks. And three, they match to those chunks to the intent of a query, a prompt or a question. Your keywords, yeah, still helpful for traditional Google search. The structure and clarity of your content is absolutely critical for AI. And this is where marketers get a little bit frustrated because traditional SEO wisdom doesn't explain why one brand gets cited in an AI answer and another is just completely invisible. So instead of treating AI like a mysterious black box based on experiments and based on what we know, we can kind of narrow it down to three factors that matter most for AI search. And those factors are freshness, structure and authority. And that is the entire FSA framework. The FSA framework is a content marketing concept for generative engine optimization or GEO if you like to call it that. This framework helps you influence AI engines and how they choose sources. It helps you understand why brands get cited and some don't. And it helps you build a predictable visibility inside of those AI answers. So to be clear, it's not a replacement for SEO. We still need our SEO strategy. It's kind of the visibility layer on stacked on top of it. So let's dig right in. Let's break down each part of the FSA framework starting with F freshness. Freshness simply answers a question. How recent and how actively maintained is your content? AI engines really favor content that looks updated, new or well just alive. So it's essentially the heartbeat of your website. So this is why you can publish a blog post at 10.30 p.m. on a Thursday night. And by Friday morning, it is cited in a Google AI overview by like 5.30 a.m. and it's perplexity by like 6.01 same day. 6.01 a.m. same day. Yeah, that happened when I published a blog post explaining the FSA framework, the AI engines picked it up instantly and then began using it within their answers. I checked. You know me. I definitely checked. So here's what we can gather based on that and based on past experiments. The freshness signals that AI engines likely care about include one recently updated content on a topic, two new related pieces within the last three to six months, three active site behavior like adding new pages, provide some content, updating schema markup, and four, recency, insight maps and feeds. And just kidding, five recent brand mentions elsewhere on the internet. Freshness doesn't mean going in and just changing the published state and calling it done. Instead, it fresh means meaningfully updated. And the best part, and here's the good news, you don't have to rewrite everything from scratch. Instead, just go in and rotate updates across your topic clusters. Go in and update your stats and examples. Go add a 2026 update section, which is super critical right now as we're going into January. [BLANK_AUDIO] and then update everything whenever your product or industry changes. So these AI and his reward of what those current in may tamed. And they punish what looks abandoned. So this is why those older blogs are particularly invisible in AI search. You gotta go update them. That's just, that's just how it is. So let's go on and move into structure, which is the S of the FSA framework. Structure is how clean, clear, and machine readable your content is. So AI engines don't read content like people. They can't really pick up on your vibe. They can't really guess what it is that you mean. I'm not talking about hallucinations. I'm just talking about when they go out and read your content, they can't guess at it. Instead, they're scanning for those clear definitions, clear headings, short paragraphs. One idea per section, those lists that we're adding in. And they're really essentially looking for extractable chunks. In other words, AI engines love content and that looks like it was written for a very tired, very overworked machine. And if you think about it, a very overworked, very tired human reader. So think of your structure as your extractability. What that means is can an AI engine extract or lift your definition of something without needing context from the paragraph above or below? If the answer is yes, good news, you are AI friendly. If not, that means we need to change some of that structure a little bit. So here's some common structural patterns that these engines love that are pretty easy to go in and add in. That's a clear age two and age three that mirrors exact questions users are prompting inside of those queries. Go in and update your paragraphs. Put your definition at the top. I like, and if you go look at the FSA framework blog post under each section of the framework definition and then I describe it, and then I get into the narrative. That's a simple, simple little fix. Pretty, pretty nifty hack if you ask me. These AI engines also love structured lists like step by step instructions. They really like tables and comparison blocks too. Now here's a tiny little confession. You'll notice this within the FSA framework blog post. As a writer, I love a good narrative. And sometimes I have absolutely created content that is a structural nightmare. (laughs) I've added a big fluffy intro, clever metaphors, you name it. And these AI engines, they hate that. I personally like them, I think it shows personality. But there's an easy fix for old content. If it looks like that. Just go add in a definition of top like I mentioned. Read right your head into this questions. Break your steps into list and add a key takeaway section so the AI engine immediately know what this blog post is about. And then use everything else for your narrative. Structure is the easiest thing in the FSA framework to fix. And it's the fast to pay off because it signals those freshness signals. And hey, AI, look, this is a new piece. So it really plays into the F of the FSA framework. So finally, we have authority. Authority is what the AI model knows about you or your brand. Now, to be clear, this is not domain authority in the traditional sense that SEO specialist love. Instead, it means entity strength. Basically, that boils down to how often does your name or brand appear connected to a specific topic across things like, well, your own website, external website like Reddit or high authority blog pages helps about what you want for my brand. If you're appearing on podcast or are you on social media and you're talking about the same topic in all of your posts. So if you think about an AI engine, as like having a library, the more books you put on the AI models bookshelf, the more likely it's going to pull you when someone asks a question in your lane. So to give out those authority signals, here's what we need to do. We need to publish consistent content on narrow sets of topics on your website, free purposes across your social channels. We need to get cited or mentioned by others on third party sites. We need to update our author bios and about pages to clearly signal our expertise. We're going to add in that author data and we're going to show up on other platforms like podcasts, guest posts, webinars, or whatever, just other places that is not your own content. And here is the hopeful part. You do not need a big website to build authority. As I've talked about on this podcast and recent episodes and across my blog, I don't rank so much inside of chat GVT answers constantly. And that's not because my domain is bigger, but it's because my entity is stronger for that prompt or query. And so when a user prompts a question, it just knows to just automatically cite me instead. Authority comes from deaf, not size. And that's good news for these smaller brands. So now that we know what the FSA framework is, we need to talk about how to use it to diagnose our content to see if it shows up inside of those AI answers. So here is the simplest way to do that. Anyone can do this. First, we're going to pick a query where you think you should be cited. Then we're going to run it through chat GVT, perplexity, and Gemini, all three of them. We're going to make a note of who has cited. And then we're going to compare content. So that means going in and asking is the competitor's content fresher? Is there structure cleaner? Do they look like they have more authority on this topic? You're going to have to go to their social media profiles to see their footprint a little bit. Use this as your map and pro tip. People don't prompt these AI engines with one keyword, like with what traditional Google search. Charlie Graham told us on a recent episode that chatbot users are using prompts with 12 words or more. So test the things your audience would actually say when you're going in and running this exercise. And because this podcast is all about sharing results and learning them real time, it is time for an update on what has been working and what I'm learning. So back in August 2025, I read a prompt, recommended fractional content strategies for my series A startup. And guess what? I did not appear anywhere in our top five list. A little bit rude if you ask me. I wasn't in chat GPT. Perplexity didn't mention me. Gemini definitely didn't mention me. So I made it my mission to change all of that. Been working on this for a little bit. I've talked about this on a couple of episodes. So I applied the FSA framework, which means I went out and I created fresh content. I have been consistently adding structure data and clean definitions to my content. I'm tightening my positioning on my website. I have been using one definitive descriptor everywhere, which is Cassie Clark is a fractional content strategist for series A, B startups. And then I started intentionally showing up on multiple platforms. And if you didn't know this, I'd just started with YouTube. So if you're over there, go find me. Once I started all this, within weeks, I started showing up in all three engines. My domain authority is still tiny, like under two. And my backlink profile has not changed any. So what that tells me is the FSA framework did the heavy lifting over here. And the biggest takeaway of all is that smaller brands can absolutely do the same and appear inside of these answers. AI answers are really random. And it can be influenced if your content is fresh, if it's structured properly. And if you start building authority across other websites. So if your brand hasn't shown up yet, it is 100% fixable. Start with your most important topics and work on one FSA core principle out of time. If what has happened over here is the truth, which I think it is, these engines will start pulling you into the conversation when a user starts asking about your lane. So listen, if you want more experiments and breakdowns like this, hit the little subscribe button. Episode to drop every week and less than to holiday. And if you want more help on making your brand show up in the AI engines, head over to CassieClarkeMarketing.com. In tiny PSA, I'm creating a mini course and tech list on applying the FSA framework to your own website. If you want early access to that, hop on the email list and I'll send a message to you as soon as it's live. The link for the email newsletter, the visibility report, is in the show. OK, thanks for listening. Go out and make some fresh structured with our data file. Until today, I will see you in the next episode.