Search as a Channel
Explores how discoverability powers modern marketing from SEO and paid media to social discovery and product growth.
Search as a Channel
Mine Your Experts
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
If Google or ChatGPT searched your category today for expert advice, would it find actual experts attached to your brand, or just polished copy? That question is reshaping how marketing leaders should think about content. We unpack why AI search is rewarding lived knowledge over coverage, why SME mining is more than a content tactic, and how to build a repeatable system that turns internal expertise into durable search assets. The era of SEO led content factories is fading. The era of expert mined brand intelligence is just beginning.
Imagine um imagine you are running this massive multimillion dollar marketing team. You have the absolute best writers on staff, a highly optimized website, and like these top-ranking keywords that you've been aggressively defending for years. Aaron Powell Right.
SPEAKER_00The dream setup.
SPEAKER_01Exactly. The dream setup. Yeah. And then overnight, literally overnight, a single AI architecture update from Google renders, like 85% of your organic traffic completely invisible.
SPEAKER_00Oh, yeah. And I mean, that isn't some hypothetical thought experiment either.
SPEAKER_01Aaron Powell No, not at all. If you are tracking the digital landscape right now, that is the reality playing out across almost every single industry. We've basically crossed this threshold where the way consumers find answers has fundamentally broken the old internet model.
SPEAKER_00Aaron Powell It absolutely has. The whole era of the self-serve research model. You know, where a search engine just hands a user a list of 10 blue links and basically forces them to click through, read, and manually synthesize the info themselves, that's effectively over.
SPEAKER_01Aaron Powell Yeah, the machines are doing it for us now.
SPEAKER_00Right. Answer engines are now doing all that heavy lifting of synthesis right at the point of the query.
SPEAKER_01Aaron Powell, which is exactly why we're doing this deep dive today. We are so thrilled to have you with us as we tear down the mechanics of what the industry is calling generative SEO. Because I mean, if AI is doing the reading and summarizing for your potential customers, how do you mathematically and structurally ensure that your brand's insights are the ones actually being cited?
SPEAKER_00Yeah, that's the million-dollar question.
SPEAKER_01Aaron Powell To figure this out, we have a really phenomenal stack of research in front of us today. We're looking directly at official documentation from Google's Search Central and their internal keyword blog.
SPEAKER_00Which is uh super revealing, by the way.
SPEAKER_01Oh, completely.
SPEAKER_00Yeah.
SPEAKER_01And we are cross-referencing that engineering data with these amazing strategic deep dives from Search Engine Journal, SEO analyst Harry Clarkson Bennett, Nicole DeLeon, and the latest citation data from Aerops.
SPEAKER_00It's a really critical stack, I think, because it bridges that gap between what the engineers at Google are officially telling us to do and the actual, you know, boots-on-the-ground reality of how large language models are retrieving and citing information in the wild.
SPEAKER_01Okay, let's unpack this. Because the glaring theme we keep seeing across all of these documents is that the classic SEO playbook, like pumping out highly optimized, generic 500-word articles just to blanket a keyword is officially a liability. The commodity content factory is closed.
SPEAKER_00Aaron Powell It had to close. And if we connect this to the bigger picture, the reason really comes down to the economic collapse of average writing. Large language models have basically driven the cost of producing grammatically perfect, completely average information to absolute zero.
SPEAKER_01Aaron Powell Because anyone can just spin it up in seconds.
SPEAKER_00Exactly. Anyone can now generate a perfectly synthesized summary of existing knowledge in like 20 seconds. So originality isn't just a no syshetic bonus for your brand anymore. Originality is the entire moat. It is literally the only structural defense you have left.
SPEAKER_01Aaron Powell I really want to look under the hood of that defense. Because to understand how to win in this new era, we can't just talk about content, right? We have to talk about the actual mechanisms of AI answer engines.
SPEAKER_00Right. The architecture.
SPEAKER_01Yeah. And Harry Clarkson Bennett is incredibly blunt in his assessment here. He argues that commodity content is doomed for two very specific architectural reasons. The first is that LLMs are basically trained on the entire historical internet.
SPEAKER_00So they've already seen it all.
SPEAKER_01Exactly. If you publish an article that just rephrases what five other sites have already said, the AI already knows it. It just collapses your generic article into its existing weightings. And then the second reason is the zero-click environment.
SPEAKER_00Right. Because if the answer engine synthesizes your generic advice directly onto the user's screen, the user has zero incentive to click through to your domain.
SPEAKER_01Aaron Powell Why would they? They already have the answer.
SPEAKER_00Trevor Burrus Exactly. You've provided no unique hook. You are essentially building a generic widget in a marketplace that now possesses infinite free generic widgets. It's just a guaranteed way to bring your marketing budget.
SPEAKER_01Aaron Powell But um how is the AI actually deciding what to synthesize? Because I think a lot of people assume the AI is just sort of thinking of an answer based on its training data from a year ago.
SPEAKER_00Aaron Powell Yeah, that's a huge misconception. The Google Search Central documentation breaks down the actual mechanism, which is retrieval augmented generation or rag.
SPEAKER_01Aaron Powell Raggy, right?
SPEAKER_00Yes. And understanding RAG is non-negotiable if you want to survive generative SEO. Without RAG, asking an AI a question relies purely on its parametric memory, like what it learned during its initial training. But because Google needs to provide real-time, accurate answers, they use ARG as a grounding mechanism.
SPEAKER_01Aaron Powell I like to think of RAG as forcing the AI to take an open book test.
SPEAKER_00Oh, that's a good way to put it.
SPEAKER_01Yeah, because without it, the AI is taking a closed book history test based on whatever it crammed into its head months ago.
SPEAKER_00It might panic and just hallucinate a completely fake historical date.
SPEAKER_01Aaron Powell Exactly. But RAG changes the rules. It forces the AI to actively open the textbook, which in this case is Google's live search index, find the specific up-to-date paragraph, and read directly from it to construct the answer.
SPEAKER_00Aaron Powell That is a perfect analogy. The model retrieves these highly relevant live web pages, brings them into its temporary context window, and then generates a response that is bounded by those specific facts. But uh here is where the architecture gets even more aggressive. Google doesn't just run the user's single query, they use a mechanism called query fan out.
SPEAKER_01Oh, this blew my mind when I read it in the documentation.
SPEAKER_00It's a massive shift in how search operates. When a user asks a complex question, the AI model generates a series of concurrent related queries in the background to fetch a multidimensional set of results.
SPEAKER_01Like all at the exact same time. Yes.
SPEAKER_00So if a user types how to fix a lawn that's full of weeds, the AI isn't just looking for pages with that exact phrasing. In a fraction of a second, the AI secretly fans out and searches for chemical breakdown of commercial herbicides and organic clover removal techniques and soil pH balance for grass health.
SPEAKER_01So it's casting this huge intelligent net.
SPEAKER_00Right. It pulls all of those retrieved documents back into the context window to build a really robust, multifaceted answer.
SPEAKER_01And we are actually seeing Google change the user interface to match this behavior. If you look at the search keyword blog, they aren't just hiding these citations anymore. They are actively embedding the links directly next to the synthesized text.
SPEAKER_00Yeah, the hover previews.
SPEAKER_01Yes. They're rolling out hover previews on desktop so a user can see your website's title and metadata before they even click, which acts as this huge trust building mechanism. And they're pulling in perspectives from first-hand sources too.
SPEAKER_00Which is so important for building that authority.
SPEAKER_01But wait, I have to challenge this premise for a second. If you're listening to this right now, you might be looking at your company's massive 2024 SEO keyword calendar and thinking, do I need to scrap all of this? If query fan out means the AI is doing the multidimensional searching, our generic content is dead. Wait, so are search volumes completely useless now?
SPEAKER_00It is a totally valid concern, but no, you shouldn't throw out your keyword data. You just have to change how you interpret it. Search volume is no longer a proxy for guaranteed clicks.
SPEAKER_01Okay, so what is it?
SPEAKER_00Search volume is now strictly a proxy for hean demand. Clarkson Bennett uses a really great example with the search term family holidays. If you look at the raw data, searches for family holidays reliably spike every single January.
SPEAKER_01Because everyone is stressed and wants a vacation.
SPEAKER_00Exactly. The volume tells you exactly when the human psychological need arises.
SPEAKER_01Ah, I see. So the demand is a real measurable force. But the goal isn't to just build a generic net to catch a cheap click.
SPEAKER_00Exactly. The goal is to ensure that when the AI detects that January demand and begins its complex query fan out, your brand provides the highest quality, most original raw material for the AI to ingest. You aren't writing clickbait anymore. You are feeding an intelligence engine.
SPEAKER_01Okay, so let's say the generic widget factory is closed. We can't just look up a high-volume keyword and pay a freelancer to summarize what already exists on page one of Google. Where do we actually get this new high-quality raw material?
SPEAKER_00This brings us to Nicole Dillion's analysis, and she introduces a framework that I think should be like mandatory reading for every organization right now. She calls it SME mining or subject matter expert mining.
SPEAKER_01SME mining. Okay.
SPEAKER_00Her core argument is that the most valuable AI-proof content your organization possesses isn't sitting in some marketing brainstorming document. It already exists. It is happening naturally every single day inside your company's internal communications.
SPEAKER_01So what does this all mean? It means we have to start treating our own company's Slack channels, support tickets, and water cooler debates like an untapped gold mine. It's essentially brand intelligence extraction.
SPEAKER_00Precisely. The raw material is in the sales call recap, where an engineer explains exactly why a six-figure deal fell through because of a highly specific software integration issue. It's in the customer support escalation log. It's the impassioned debate your product team had last Tuesday about a new feature rollout.
SPEAKER_01The stuff that never makes it to the blog.
SPEAKER_00Right. DeLion argues that you have to turn this lived experience into structured assets.
SPEAKER_01But let me guess, the worst way to do this is to walk up to your brilliant lead engineer and ask them to write a 1,000-word blog post about it.
SPEAKER_00Oh, it's a guaranteed failure. They are busy being experts. They don't want to write marketing copy. If you give an SME a blank page and a deadline, you will wait six months and get absolutely nothing.
SPEAKER_01So true.
SPEAKER_00Instead, DeLion outlines what she calls a lightweight interview engine.
SPEAKER_01Walk us through the mechanics of that engine. How does a strategist actually extract this?
SPEAKER_00Well, you don't ask for a draft. You set up a highly focused, 20-minute recorded conversation, you send them sharp, provocative prompts in advance, and then they just talk. Your job as the strategist isn't to write, it's to act like an investigative journalist.
SPEAKER_01Oh, I love that.
SPEAKER_00You record the call, transcribe it, and then you aggressively mine that transcript for tension. You look for the trade-off, you look for the unexpected, counterintuitive customer stories.
SPEAKER_01So, okay, if you sell cybersecurity software, the boring generic content approach is writing an article that says our software is very secure and uses encryption.
SPEAKER_00Yeah.
SPEAKER_01The AI completely ignores that.
SPEAKER_00Right. It's white noise.
SPEAKER_01But the SME mining approach is finding out from the sales rep that your software is the only one that patches a highly specific firewall leak that happens during automated server reboots at 2.0 AM. That is the tension.
SPEAKER_00Yes. And from that one 20-minute interview, you can spin out an entire ecosystem of assets. That single insight becomes a high-density blog post, a contrarian LinkedIn thread, an addition to your technical documentation, and a targeted sales enablement PDF.
SPEAKER_01But let me play devil's advocate here for a second. Even if you extract that great 2.0 AM firewall story, if an AI model can technically generate a passable, grammatically perfect article about cybersecurity in 20 seconds, why does the underlying architecture of a search engine actually care about your interview? Like why would a buyer or an AI treat your version as special?
SPEAKER_00Because of a deeply embedded architectural concept called information gain. And this is vital. Google actually holds patents on estimating and scoring information gain. It isn't a fluffy marketing term, it is a literal mathematical calculation.
SPEAKER_01Aaron Powell Let's break that math down conceptually, because I think a lot of people miss this.
SPEAKER_00Sure. So if the AI looks at 10 existing articles about cybersecurity and they all say the exact same basic facts, the 11th article that repeats those facts has an information gain score of absolute zero.
SPEAKER_01Aaron Powell Because it doesn't teach the model anything new.
SPEAKER_00Aaron Ross Powell Right. It adds no new vectors of understanding to the model. But if your interview-based article introduces the 2.0 AM server reboot anomaly, a fact that does not exist anywhere else in the model's current weights or the retrieved index, the math immediately flags that as a positive delta. The algorithm actively calculates your uniqueness and pushes you up the citation hierarchy because you are expanding the knowledge graph, not just echoing it.
SPEAKER_01Aaron Powell That is fascinating. The AI literally calculates your uniqueness. It rewards effort and insight that is difficult to replicate.
SPEAKER_00Exactly. The true value that agencies and in-house teams provide now is not content output volume. Their value is capturing institutional expertise before it evaporates into the ether.
SPEAKER_01Okay, so let's say you've done the work. You've mined the gold, you have the high information gain, you have the tension. But if you just dump a raw, unstructured transcript onto your website, the AI still isn't going to parse it effectively, and a human certainly isn't going to read it. How do you refine this raw material?
SPEAKER_00This is where we bridge the gap between human psychology and machine parsing, looking at some great insights from Search Engine Journal. Before you worry about the algorithm, your content has to trigger the human limbic system first. Emotion has to grant permission before logic can even engage. It perfectly illustrates the shift. Think about generic informational copy. Our coffee shop is open 24 hours and uses high-quality beans sourced globally.
SPEAKER_01Yeah, that's just boring. Factually accurate, but totally forgettable.
SPEAKER_00Right. And LLM can generate that in its sleep. Now, compare it to visceral storytelling based on actual customer observation. For the late night grinders and the early risers, fuel that traveled 4,000 miles to keep you going were awake when you are.
SPEAKER_01Oh wow. It completely changes the atmospheric pressure of the text. You aren't just listing a business hours schema, you are identifying the specific psychological state of your consumer.
SPEAKER_00But, and this is crucial, emotion alone will fail if the structure contradicts how information is actually consumed online. Clarkson Bennett breaks down the reality of audience behavior into four types. You have answer seekers who just want a binary fact and will bounce immediately. You have visual or audio consumers, you have deep readers who will read an entire 3,000-word essay top to bottom, but they make up a tiny, tiny fraction of a percent of your audience. And finally, you have the scanners.
SPEAKER_01And scanners make up over 90% of internet users. I mean, people do not read web pages like they read novels. They treat web pages like topographical maps. They are scrolling rapidly, looking for structural landmarks to tell them if the page is even worth their time.
SPEAKER_00And here is where human psychology mirrors machine architecture perfectly. Large language models suffer from an architectural limitation known as the lost in the middle effect. Because of the way an LLM's attention mechanisms work across its context window, it heavily weighs information at the very beginning of a prompt and at the very end. But if you bury your most critical high information gain argument right in the middle of a massive 800-word block of unstructured text, the attention weights literally drop off.
SPEAKER_01That is wild.
SPEAKER_00The AI model loses the ability to accurately retrieve and cite that specific fact, exactly like a human scanner whose eyes just glaze over a wall of text.
SPEAKER_01So the solution for both the human scanner and the AI model is identical. High scannability. You must use bullet points. You must use semantic HTML, meeting proper nested use of H1, H2, and H3 headings, these aren't just design choices anymore. They are structural signposts.
SPEAKER_00Exactly. If you don't front-load your most important insights and break the text up visually and semantically, the AI mathematically struggles to cite it and the human just refuses to read it.
SPEAKER_01But let me push back on this because it feels like a bit of a paradox. If we are forced to format absolutely everything with bullet points and bolded keywords and rigid heading schemas just to appease an LLM's attention mechanism, aren't we just stripping the soul out of the writing? Aren't we going to end up sounding like robots ourselves?
SPEAKER_00What's fascinating here is that the exact opposite is true. The strict formatting actually elevates the need for human emotion. Search Engine Journal refers to this as the human in the loop requirement.
SPEAKER_01Human in the loop.
SPEAKER_00Right. The machine provides the scale and the structural formatting, but the human must provide the meaning. They highlighted a really wild case study to prove this, a short film called Lily that won a million-dollar prize at an AI film summit.
SPEAKER_01Oh, this story completely reframed how I view AI generation. Yeah.
SPEAKER_00The entire film was generated using Google's AI video tools. The narrative follows a lonely archivist who discovers a doll at the scene of a hit and run, and the doll slowly becomes the silent psychological witness to the archivist's guilty conscience.
SPEAKER_01Reapy but cool.
SPEAKER_00Very. And the AI generated all the gloomy visual aesthetics and the character movements flawlessly, but the AI did not invent the concept of guilt.
SPEAKER_01Right. The AI is essentially just a highly advanced paintbrush. The human is the artist deciding what to paint. Because, I mean, an LLM can render the image of a sad doll, but a machine has no foundational understanding of what sadness actually is.
SPEAKER_00It doesn't feel it.
SPEAKER_01Exactly. It doesn't understand why a physical object at a crime scene becomes unbearable for a human to look at. The human creator brought the elemental, lived emotion of isolation and confession. The AI simply executed the structural vision.
SPEAKER_00Precisely. Humans provide the meaning, machines provide the scale. And speaking of machine execution, we really need to highlight some crucial myth busting directly from the Google search central documentation.
SPEAKER_01Oh, about formatting.
SPEAKER_00Yeah, there is a lot of panic right now among technical SEOs about how to code for AI. But Google explicitly states you do not need to create special machine readable text files, like the heavily debated elms.txt concept. You don't need to artificially chunk your content into tiny disjointed fragments.
SPEAKER_01That is such a relief. You don't have to write in binary just to get noticed.
SPEAKER_00No, not at all. Because modern AI natively understands synonyms, semantic relationships, and natural language flows. If you write viscerally for humans, but structure it with clear logical landmarks, the AI's natural language processing will figure out the rest.
SPEAKER_01Aaron Powell Okay, but structured, highly emotional content with incredible information gain is fantastic. But what if it was published two years ago? Or what if it exists in a vacuum on a brand new website with no audience? Here's where it gets really interesting. Because the foundational rules of authority have completely shifted. We have to talk about velocity, off-site signals, and the literal race to be first.
SPEAKER_00Let's look at the hard data from AirOps on this. They analyzed thousands of live AI queries to see what actually gets cited in the wild, and the metrics are just staggering. 70% of the pages actively cited by AI models were updated within the past 12 months.
SPEAKER_0170%. That's huge.
SPEAKER_00And it gets more aggressive. Content that was updated within the last three months is three times more likely to be pulled into a generative answer. Yeah. You have to understand why the architecture demands this. LLMs are notoriously prone to hallucination. So to combat this, the grounding mechanisms are heavily weighted toward freshness. If your definitive industry guide hasn't been updated in two years, the model's confidence threshold just drops.
SPEAKER_01It just assumes things have changed.
SPEAKER_00Exactly. It assumes the data is stale and skips right over you to find a more recent, albeit potentially inferior, piece of content. Speed and freshness are the new technical authority signals.
SPEAKER_01Aaron Powell Which perfectly explains the incredible API design success story from the SEJ playbook. You have this small 10-person startup trying to break into the API software space. And the space is completely dominated by massive legacy tech giants with, you know, bottomless marketing budgets.
SPEAKER_00Right. Impossible to compete on traditional terms.
SPEAKER_01Aaron Powell Totally impossible. The startup knew they couldn't outride the giants on established high-volume topics, but they surveyed their niche audience and noticed a new phrase bubbling up in the community: API design.
SPEAKER_00And at the time, if you put API design into a keyword research tool, it showed zero search volume. The legacy competitors were ignoring it completely because the historical data didn't justify the investment.
SPEAKER_01Exactly. So the startup didn't spend eight months committee reviewing a comprehensive award-winning masterpiece. They just interviewed their internal expert, threw together a highly structured 1500-word page, and they published it first.
SPEAKER_00They just got it out there.
SPEAKER_01They got out of the blocks before the Giants even realized a race had started. And a year later, when the actual human demand exploded and the cost per click for that keyword skyrocketed to like $200, the startup owned the entire space. Perfection is the absolute enemy of good when it comes to capturing a new concept in an LLM's index.
SPEAKER_00It really proves that information game isn't just about having the most profound thought. Mathematically, it is often just about having the first thought on the public record. But even then, we really have to address the most uncomfortable truth in the AirOps data.
SPEAKER_01Over third-party stuff.
SPEAKER_00Yes. Publishing brilliant, fast content on your own domain is no longer sufficient. 85% of brand mentions inside AI-generated answers actually stem from third-party sources, not the brand's primary website.
SPEAKER_01I want everyone listening to stop and process that. 85%. That is a massive systemic blind spot for most marketing teams. We spend almost our entire budget obsessing over our own landing pages, tweaking button colors, and agonizing over our own corporate blog copy. Meanwhile, the AI architecture is out there scraping third-party listicles, independent review sites, and Reddit threads to determine what the broader ecosystem actually thinks of us.
SPEAKER_00It is the ultimate validation check. If you are selling a B2B software tool and your brand is not mentioned in the top three results of an independent third party best software tools of 2026 listicle, you are essentially invisible when an AI runs a comparison query. The AI cross Cross-references your claims against the consensus of the Internet. Trevor Burrus, Jr.
SPEAKER_01Which makes sense from a trust perspective.
SPEAKER_00Trevor Burrus, Jr.: Exactly. Furthermore, the data shows that community platforms like Reddit and YouTube now make up nearly 48% of all AI citations.
SPEAKER_01Almost half of all citations come from community platforms. But why does the AI crave Reddit so much? I mean, it's just a bunch of random people talking. Trevor Burrus, Jr.
SPEAKER_00Because LLMs are, at their core, probabilistic prediction engines trained on massive data sets of human conversational text. When they look for authority, they are not looking for sterilized, risk-averse corporate mission statements. They are looking for messy, high-density sentiment.
SPEAKER_01Aaron Powell Uh, the lived experience again.
SPEAKER_00Right. They recognize the structural patterns of a passionate Reddit debate, the upvotes, the specific technical arguments, the authentic lived experience as a much closer proxy for the truth than a biased corporate sales page.
SPEAKER_01Aaron Powell, which means the way executives measure marketing success has to fundamentally change. You cannot walk into a board meeting and report on generic page views or bounce rates anymore. Those are relics of the 10 blue link era.
SPEAKER_00Absolutely. Those are vanity metrics in a generative world. You have to track behavioral and syndication signals now. You track watch time on your video assets to see if the limbic system is actually engaged. You track scroll depth. And most importantly, you track answer referral traffic.
SPEAKER_01Answer referral traffic.
SPEAKER_00Yeah. When a user asks an AI a complex question, reads the synthesized paragraph, and then explicitly clicks the embedded citation link to visit your site. That is the new organic click. That is the highest intent traffic you can possibly capture because the AI has already pre-qualified your authority.
SPEAKER_01Okay, we have covered an immense amount of architectural and strategic ground today. Let's synthesize this journey for everyone. The era of the SEO content factory, blindly churning out average articles to capture historical search volume, is dead. The winning playbook is generative SEO.
SPEAKER_00R.I.P. to the factory.
SPEAKER_01Seriously.
SPEAKER_00Yeah.
SPEAKER_01It requires you to operate like an investigative journalist inside your own company, extracting the raw, lived expertise of your team before it vanishes. You have to format that insight with ruthless scannability so both human eyes and machine attention mechanisms can actually process it. You have to push those insights out of your own walled garden and into community platforms, and you have to update your ecosystem relentlessly to prove to the modders that you are the freshest, most authoritative node on the graph.
SPEAKER_00Beautifully summarized. And this raises an important question, which brings us to a final critical thought I want to leave everyone with. It is based on a brief, highly technical mention tucked away in the Google Search Central documentation regarding agentic experiences and the universal commerce protocol.
SPEAKER_01Okay. Yeah.
SPEAKER_00Throughout this entire deep dive, we have been talking about AI summarizing answers for human beings to read. But the immediate next phase of this architecture is autonomous AI agents.
SPEAKER_01This is where the science fiction becomes corporate reality. Walk us through it.
SPEAKER_00Imagine an AI agent that is given a localized budget by a human user. A human doesn't ask for a summary of software options. The human instructs the agent to autonomously research, compare, negotiate, and actually purchase a software platform or book a B2B service.
SPEAKER_01Wait, autonomous? So the human isn't reading the output at all?
SPEAKER_00Exactly. The machines are communicating via the Universal Commerce Protocol, negotiating with other machines based purely on the structured data they retrieve. If your brand's digital presence is not built on clear, unique, high information game expertise that is explicitly machine readable, the AI won't just fail to cite you in a paragraph.
SPEAKER_01It won't even know you exist.
SPEAKER_00It will literally fail to buy from you. Your revenue will just quietly vanish.
SPEAKER_01That is a staggering look at where the underlying architecture is heading. You simply cannot afford to be invisible to these systems. Well, thank you so much for joining us on this deep dive today. As you head into your office or you know, do jump onto your next internal slack huddle tomorrow morning. I want you to look at your team and ask yourself who is the expert we aren't listening to? Because if you want to survive the shift to generative SEO, you have to feed the intelligence engine a story it simply cannot ignore. We'll see you next time.