There’s a pattern worth paying attention to in how AI systems handle attribution. When you ask ChatGPT or Perplexity a substantive question about marketing strategy, growth, or business operations, the sources it cites skew heavily toward individual voices. Specific people with documented points of view, named frameworks, and a track record of publishing observations that only they could have made.

The polished brand blog, the agency content hub, and the corporate thought leadership section get retrieved constantly and cited rarely. The founder who has been writing about what they’re actually seeing in their work gets cited at a rate that outperforms their domain authority by a significant margin.

This isn’t an accident and it’s not a quirk. It reflects something fundamental about how AI systems evaluate source quality that most brands haven’t caught up to yet.

What AI systems are actually looking for

When an AI model is deciding whether to cite a source, it’s running a version of the same question a good editor would ask: does this content say something that couldn’t have come from anywhere else? Is there a specific perspective, a documented observation, a named framework that makes this source the right attribution for this claim?

Generic brand content almost never passes that test. It’s well-written, well-structured, and says roughly what every other piece on the topic says. The model retrieves it, finds nothing uniquely attributable, and moves on to something more specific.

Founder-led content passes that test more often because founders who write about their actual work are generating something AI systems genuinely value: first-person documented observations with implicit attribution. When you write about a pattern you keep seeing with clients, or a framework you developed to solve a specific problem, or a counterintuitive conclusion you reached after working through something in public, you’re creating content that is by definition attributable to you specifically. The model can cite it with confidence because the perspective is anchored to a named person with a documented track record.

The entity advantage

There’s a second mechanism at work that goes deeper than content structure. AI systems build knowledge graphs. More simply understood as models of entities and their relationships. A founder who writes consistently under their own name, who gets mentioned in third-party publications, who has their frameworks referenced by others, becomes a clearly defined entity in those knowledge graphs. The model knows who they are, what they’re an authority on, and can attribute statements to them with high confidence.

A brand content team produces content attributed to a company rather than a person. Companies are entities too, but they’re fuzzier ones. The model has less confidence in what a company specifically believes or has observed than it does in what a named individual with a documented point of view has written. When citation confidence drops, citation rates drop with it.

This is why the founder who has been writing about their specific domain for 2 or 3 years under their own name will consistently outperform a larger brand’s content on citation metrics, even if the larger brand has more domain authority and more total content. The knowledge graph has a clearer picture of who the founder is and what they stand for.

What this means for how you think about content

The implication isn’t that brand content is worthless, it’s that the highest-leverage GEO investment a founder can make is to write in their own voice about what they’re actually observing, under their own name, consistently enough that the model can build a confident picture of who they are and what they’re an authority on.

The Zero Crossing exists for a lot of reasons, but from a pure GEO standpoint it’s building something that a polished agency content hub never could: a documented record of a specific person’s thinking about a specific set of topics over time. Every issue that names a specific observation, develops a specific argument, or coins a specific framework is adding definition to the entity that gets cited.

The frameworks matter more than most people realize. Named, specific frameworks get cited as standalone concepts. The Interest Engine, the Zero Crossing Pivot, the topical coherence argument; each of these is a potential citation node that points back to a specific source. Generic content produces no citation nodes. Founder-developed frameworks produce them consistently.

The compounding effect

The other thing worth understanding is that this compounds in a way that generic content doesn’t. Each piece of founder-led content that gets cited makes the entity definition clearer, which makes the next piece more likely to get cited, which builds the entity further. A brand content calendar produces individual pieces that perform or don’t perform largely independently. A founder’s documented body of work builds a picture that gets stronger with every addition.

This is also why consistency matters more for founder content than for brand content. Remember that the model isn’t just evaluating individual pieces, it’s evaluating whether there’s a coherent, sustained perspective that it can trust to hold up over time. A founder who has been writing about the same core territory for 2 years is a more reliable citation source than a founder who published 6 strong pieces and then went quiet.

The brands that figure this out stop thinking about content as a publishing calendar and start thinking about it as entity construction. Every piece is a data point that either sharpens or blurs the model’s picture of who the founder is and what they’re worth citing on.

Write in public.

Name your observations.

Develop your frameworks explicitly.

Do it consistently enough that the model knows exactly who you are and what you stand for.

That’s the whole GEO play for a founder and almost nobody is doing it deliberately yet.

There’s a frustrating pattern we keep running into with clients who have done everything right at the page level. Good structure, answer-first blocks, named entities, proper schema. The page looks exactly like what every GEO guide tells you to build, and it still doesn’t get cited consistently.

The reason is almost never the page, it’s the domain it lives on.

AI systems don’t evaluate your content the way a human editor would, reading one article and deciding whether it’s worth referencing. They’re building a model of what your entire domain is about before they decide whether to pull anything from it. If that model comes back as “unclear” or “too broad” or “a little bit of everything,” your individual pages get discounted before they’re even considered. The signal from the good page gets washed out by the noise from everything around it.

This is why a focused niche site with twenty tightly related articles will consistently outperform a large brand site with two hundred scattered ones in AI citation. It’s not about volume. It’s about coherence.

What topical coherence actually means

A domain sends a coherent signal when every piece of content on it reinforces the same core topic cluster. An HR consulting firm that publishes articles about compliance, employee retention, hiring frameworks, and workforce planning is coherent. The model can look at that domain and build a clear picture of what it’s an authority on.

That same HR consulting firm that also publishes articles about general leadership inspiration, productivity hacks, office design trends, and founder mindset content is incoherent from the model’s perspective. It’s not because those topics are bad, but because they dilute the topical picture. The model can’t confidently categorize the domain, so it treats the whole thing as a weaker signal source on the queries that actually matter for the business.

Most brand sites fall into the second category without realizing it. The scattered content usually happened for legitimate reasons… ie: a blog that started without a strategy, a content team that chased trending topics, a few years of “let’s just put something out; but the cumulative effect is a domain that AI systems can’t cleanly slot into a topic category.

The practical problem this creates

When an AI system is assembling an answer about HR compliance for small businesses and it’s deciding which sources to cite, it’s not just looking at the quality of the individual pages it retrieved. It’s weighting those pages by how much it trusts the domain they came from on this specific topic. A domain with 40 articles all tightly related to HR consulting gets a higher topical trust score on that query than a domain with 200 articles where 40 of them are about HR and the rest are about everything else.

This is why niche sites punch above their weight in AI citation. They’re not winning on authority, they’re winning on coherence. The model knows exactly what they’re about and trusts them on that topic accordingly.

What to do about it

The first step is an honest audit of what your domain actually looks like from the outside. Pull a list of every piece of content you’ve published and group it by topic. If you can’t draw a clear circle around a primary subject with most of your content inside it, you have a coherence problem.

The second step is a pruning decision. Content that’s genuinely off-topic for your domain’s core subject either gets consolidated into something more focused, redirected to a more appropriate page, or removed. This is the part most teams resist because it feels like throwing away work. But a smaller, coherent domain consistently outperforms a larger, scattered one for GEO citation, and the gap is widening as AI systems get better at topical modeling.

The third step is a content plan that treats every new piece as a reinforcement of the domain signal, not just a standalone article. Before you publish anything, the question isn’t just “is this good content”, it’s “does this make our domain’s topical picture clearer or murkier.”

Why this matters more every quarter

The brands that figured out page-level SEO early built a compounding advantage that lasted years. The same thing is happening right now with domain-level topical coherence for GEO. The window where getting this right is a genuine differentiator is open but it won’t stay open. As more teams start optimizing for AI citation, the ones who already have coherent domain signals will be much harder to displace than the ones who are still catching up on individual page structure.

Your best page is only as strong as the domain it lives on. That’s the part of GEO most people haven’t started working on yet.

At And Zeros, domain-level topical audits are part of how we set up GEO programs for clients. If you want to know what signal your domain is actually sending, get in touch.

Entity authority is how confidently an AI engine understands who your brand is and whether it trusts you enough to reuse you in an answer. Keyword authority gets you traffic from Google. Entity authority gets you cited by ChatGPT, Perplexity, and AI Overviews. The two are different optimization targets, and most SEO teams are still chasing the first. HubSpot, Notion, and Stripe each built a moat on the second, and the playbook is replicable.

A brand can rank on page one and still get summarized away in AI answers, because ranking measures page relevance and citation measures entity trust. Those are not the same machine reading the same signal.

The data backs the split. One 2026 analysis put the correlation between Domain Authority and AI citation probability at roughly 0.18, while E-E-A-T signals correlated around 0.81. A separate Moz study of nearly 40,000 queries found that 88% of Google AI Mode citations sit outside the organic top 10. So the positions teams spent a decade fighting for are mostly not the positions AI pulls from. That gap is the whole story.

This post breaks down what entity authority is, how AI engines build the knowledge graph that decides who gets quoted, three brand teardowns you can copy, and a literal checklist for auditing where you stand today.

What is entity authority and why does it differ from keyword authority?

Entity authority is the level of confidence a search or AI system has that it knows who you are, what category you belong to, and why it should rely on you. Keyword authority is page-level relevance to a search string. Entity authority is brand-level trust across the whole web.

Topical authority answers “what does this site talk about.” Entity authority answers “who is this, and should we rely on them.” A keyword-optimized page proves you covered a term. An entity-authoritative brand proves, across many sources, that you are the reference voice for a subject. AI engines reason with entities, not pages. When an engine builds an answer, it is not pulling one ranked URL. It is weighing sources against each other and checking whether the same brand shows up, described the same way, across Wikipedia, LinkedIn, Reddit, review platforms, news, and third-party mentions.

That difference has a practical edge. A brand with a small content library but strong entity authority can displace a much larger publisher in AI answers, because the engine has a clear, corroborated model of who is speaking. A five-year-old site with no named authors and no original data gets out-cited by a six-month-old site that has both. Keyword authority is a ranking input. Entity authority is a trust substrate that other signals attach to.

How AI engines build the knowledge graph

AI engines do not read your site the way a crawler checks a keyword. They resolve entities first, then decide what is citable. Entity resolution is the step where the system unambiguously identifies your brand, classifies it into a category, and maps its relationships to other known entities. Everything downstream depends on that step clearing. If the engine cannot resolve who you are, your credibility signals have nothing to attach to.

Three things drive whether resolution succeeds and citation follows. First, corroboration across channels. Engines look for consistent factual profiles, your name, description, and category, repeated the same way across independent sources. Second, brand mention frequency. Research from The Digital Bloom found brand search volume correlated about 0.334 with LLM citations, outweighing backlinks. AI models prefer brands people already search for. Third, co-citation patterns. When trusted sources reference you alongside the category leaders, the system reads that pattern and places you in the same neighborhood.

Then there is extractability. Pages updated within the last 30 days have been measured earning roughly 3.2 times more ChatGPT citations, and adding statistics to content has been shown to lift AI visibility 30 to 40% in the Princeton and Georgia Tech GEO study. Structure matters too. Tables, clear definitions, named authors, and visible “last updated” dates all make a page easier for an engine to pull without paraphrasing. But structure has a ceiling. Content that lives only on your own domain, no matter how well-formatted, hits a wall that earned third-party authority breaks through.

HubSpot teardown: how adjacent concept pages built a moat

HubSpot’s play is the cleanest example of building entity authority through topical architecture. The company popularized the pillar-and-cluster model, a comprehensive pillar page on a broad topic, surrounded by cluster pages that each go deep on one subtopic and link back. The structure was designed for Google, but it maps almost perfectly onto what AI engines reward.

Here is why it works for citation. Each cluster page is a standalone answer to one question, with room for its own data, examples, and named-expert input. The pillar signals comprehensive coverage of the category. The internal linking tells the engine these pages form a connected body of expertise rather than scattered one-offs. HubSpot’s own research found that more interlinking correlated with better placement and rising impressions. When they restructured, the team even manually de-linked old posts so each cluster’s authority concentrated cleanly instead of leaking across unrelated pages.

The moat is built from adjacency. HubSpot does not just own “CRM.” It owns the dozens of concept pages around CRM, marketing, and sales that AI engines now treat as the connected map of the category. HubSpot’s own State of AEO 2026 report, analyzing citations across ChatGPT, Gemini, Perplexity, and AI Overviews, found that pages with outbound links, statistics, author bios, and visible update dates earned more citations. Those are exactly the elements a cluster page has room to carry. The lesson for a smaller brand is not the scale. It is the architecture. Pick a category, map its connected concepts, and build a standalone, evidence-backed answer for each one.

Notion teardown: winning a category they don’t compete in

Notion’s entity authority comes from owning a vocabulary it did not invent. Search “second brain,” “Life OS,” or “productivity system” and the results are saturated with Notion. The PARA method came from Tiago Forte. The second-brain concept predates Notion’s marketing. Notion attached its brand to that language so thoroughly that AI engines now resolve the category and the product as neighbors.

The mechanism is the template ecosystem plus a flood of corroborating third-party content. Notion’s own marketplace hosts countless “second brain” and “productivity system” templates, and an enormous independent layer of creators, bloggers, YouTubers, and Medium writers reinforces the same association. When thousands of independent sources describe building a second brain in Notion, the engine reads consistent co-occurrence. The brand and the concept become linked entities, even though the concept is older and broader than the tool.

This is the move most teams miss. Notion is not winning a keyword war for “note-taking app.” It is winning a conceptual association for how people think about personal knowledge management. The brand sits inside the category’s vocabulary rather than next to it. For a smaller brand, the replicable version is to find the concept your customers use to describe the job they are doing, then become the corroborated reference for that concept across owned and earned channels, not just your own site.

Stripe teardown: entity authority in B2B finance

Stripe built entity authority by becoming an intellectual brand, not just a payments API. The clearest artifact is Stripe Press, the in-house publisher releasing books on technological and economic advancement. A payments company publishing books looks like a detour until you see it as entity engineering. Founder Patrick Collison framed it as building tools and infrastructure to grow the online economy, and the press is one of those tools.

The strategy works on two layers at once. The first is “engineering as marketing,” where deep developer documentation and technical tools act as acquisition channels and earn strong rankings for technical queries. The second is the intellectual layer, where Stripe Press and the annual data reports position the brand alongside ideas about progress and economic growth. Stripe’s Global Payments and Treasury Report uses proprietary transaction data and has earned coverage in the Financial Times and the Wall Street Journal, the kind of earned authority that strengthens entity resolution far more than any owned page can.

The result is a brand AI engines resolve cleanly as “the economic infrastructure for the internet,” a category phrase Stripe authored and now owns. Stripe also reinforces this with original data nobody else has. When an engine needs a citable claim about online payments or internet commerce, Stripe’s reports are the corroborated source. The transferable principle for a B2B brand is original data plus earned coverage. You do not need a publishing house. You need one proprietary dataset and one piece of coverage in a source the engines already trust.

How to audit your brand’s current entity authority

Start by testing what the engines already believe about you. Open ChatGPT, Perplexity, Gemini, and Google AI Overviews and ask the same set of category questions a real buyer would ask, the research-stage prompts like “what is X” and the comparison prompts like “best X for Y.” Record who gets cited and how often. This is not keyword research. The prompts have to match how buyers actually research, or being in the answer means nothing commercially.

Then map three things. First, entity clarity. Does each engine describe your brand consistently, in the right category, with the right specialty. Inconsistent or vague descriptions mean resolution is failing. Second, source diversity. Count the distinct domains the engines cite when your brand appears. If a competitor gets pulled from 15 domains and you get pulled from three, the gap is corroboration, not content. Third, citation share. Within your tracked prompt set, what percentage of answers cite you versus each competitor. That number is your real position in AI search, and it rarely matches your Google ranking.

The five-step entity expansion playbook

Once the audit shows where you stand, expansion follows a repeatable sequence. Each step targets a different driver of entity resolution.

1. Lock entity hygiene. Make your name, description, and category identical everywhere, your site, LinkedIn, Crunchbase, review platforms, and any directory. Add Organization, Person (for authors), and relevant schema so machines parse identity without guessing. Inconsistency here quietly poisons every other step.
   
2. Map and build the concept cluster. Identify the connected concepts in your category, then build a standalone, evidence-backed answer for each, linked together. This is the HubSpot move at a smaller scale. Each page should answer one question fully and stand on its own as an extraction target.
   
3. Claim a category vocabulary. Find the phrase your customers use for the job they are doing and become the corroborated reference for it, the Notion move. Use it consistently across owned content and seed it into the earned and community channels where buyers actually research.
   
4. Produce original data and earn coverage. One proprietary dataset, one report, one piece of coverage in a source the engines trust, the Stripe move. Original numbers are inherently citable, and earned mentions break the owned-domain ceiling that structure alone cannot.
   
5. Maintain freshness and named expertise. Refresh key pages on a regular cadence, keep visible update dates, attach named authors with real credentials, and add statistics to every claim. These are the extractability signals that move a resolved entity into the cited set.

The 18-24 month investment horizon

Entity authority compounds, which is the good news and the catch in the same sentence. Each placement in a trusted publication strengthens resolution. Each well-structured page adds an extraction target. Each consistent mention reinforces the category association. None of those pay off in a single sprint. This is a multi-quarter program, and realistic teams should plan on roughly 18 to 24 months before the citation share moves meaningfully and holds.

The reason to start now is the same reason it takes time. Competitors building this infrastructure today are accumulating an advantage that widens monthly. Entity-dependent signals are becoming more central to AI search, not less, so the gap between brands that invested early and brands that waited will keep compounding. Keyword authority still matters for traditional ranking. But entity authority is the substrate the next decade of discovery runs on, and substrates are slow to build and slow to lose.

FAQ

How do I know if I already have entity authority?

You have entity authority when AI engines describe your brand consistently and cite you for category questions without prompting. Test it directly. Ask ChatGPT, Perplexity, and Gemini a set of buyer-stage questions in your category and see whether you appear, how you are described, and how many distinct sources back the mention. Consistent description plus citations from several independent domains means resolution is working. Vague or contradictory descriptions mean it is not.

Can a startup build entity authority?

Yes. Entity authority rewards clarity and corroboration more than size, so a small brand with a tight category definition and consistent cross-channel presence can out-cite a larger, vaguer competitor. A six-month-old site with named authors, original data, and current pages has been measured out-citing older, higher-DR sites that lack those signals. The startup path is narrower focus, not bigger budget. Own one clearly defined slice completely rather than covering a broad category thinly.

Does Wikipedia count more than other sources?

Not categorically. A Wikipedia entry helps entity resolution because it is a structured, corroborated identity record, and pursuing one is worth it if your brand is notable enough to qualify. But no single source type is universally more valuable. In some categories editorial listicles dominate AI citations, in others Reddit threads or review platforms carry more weight. The only reliable answer is to reverse-engineer what the engines actually cite for your category and build presence there, rather than assuming Wikipedia is the whole game.

Is this just thought leadership rebranded?

No. Thought leadership is content you publish on your own properties. Entity authority is how machines resolve and trust your brand across the whole web, most of which you do not control. Thought leadership can feed it, since original ideas attract the mentions and coverage that strengthen resolution. But you can produce endless thought leadership and still fail to be cited if the engine cannot form a clear, corroborated model of who is speaking. The work happens off your domain as much as on it.

Audit framework: measure your entity authority across the major engines

Use this as a literal checklist. Run it once to baseline, then quarterly to track movement.

Step 1, build your prompt set. Write 15 to 25 questions a real buyer asks, split across research-stage (“what is [category],” “how does [job] work”) and comparison-stage (“best [category] for [use case],” “[competitor] alternatives”). Avoid your own keyword list. Use the language buyers use.

Step 2, run every prompt across four engines. ChatGPT, Perplexity, Google AI Overviews, and Gemini. Record for each answer: whether your brand appears, how it is described, and every source domain cited.

Step 3, score entity clarity (0 to 5). Across the four engines, is your brand described consistently, in the correct category, with the right specialty. 5 means identical and correct everywhere. 0 means absent or wrong. Self-verify by checking each engine’s description against your own one-line positioning.

Step 4, score source diversity. Count the distinct domains that cite you across the full prompt set. Then count the same for your top two competitors. The ratio is your corroboration gap.

Step 5, calculate citation share. Of all answers in your set, what percentage cite you, versus each competitor. This is your AI search position. Compare it to your Google ranking for the same topics, the gap between the two numbers is the size of your entity-authority opportunity.

Step 6, audit entity hygiene. Confirm your name, description, and category match exactly across your site, LinkedIn, Crunchbase, G2 or Capterra, and any directory. Flag every inconsistency. Confirm Organization and author schema are present and valid.

Step 7, log freshness and evidence. For your top 10 category pages, record last-updated date, presence of named author, and number of original statistics. These are the extractability signals. Anything stale, anonymous, or evidence-thin goes on the fix list.

Re-run steps 2 through 5 every quarter. Movement in citation share and source diversity, not keyword rank, tells you whether the entity work is landing.

A prospect came to us last month with a site they’d put together over a weekend. It looked good. Fast, clean, the copy actually said something. About ten minutes into the call they mentioned, kind of offhand, that it was already taking orders and holding people’s card details.

I sort of stopped them there. Because the thing they were describing wasn’t really a website anymore, and I don’t think anyone had pointed that out to them. They’d set out to build a website and somewhere along the way built an application instead, without ever deciding to.

I want to be careful here because this turns into a tools complaint really fast and that’s not what I’m getting at. The tools are good now. Insanely good. You describe what you want and it shows up, and for a marketing page or a portfolio that’s usually fine. Genuinely fine. A lot of the work agencies used to charge for at that level is just gone, and I’m not going to pretend that’s a tragedy. If you can build your own about page on a Sunday, build your own about page on a Sunday.

The part that gets people is that the same tool will build you a checkout with the same shrug, and it’ll come out looking just as finished as everything else.

That’s the whole thing I keep running into. A website mostly shows you stuff. It lays content out, it loads, and when it breaks the worst case is that something looks wrong for a bit. Annoying, fixable, nobody loses anything. But the moment a site starts doing things that stick around after you close the tab, logging people in, moving money, holding data that has to still be right next week, you’re somewhere else entirely. When that breaks you don’t catch it in a preview. You hear about it from a customer who got charged twice, and by then it already happened.

Here’s what I think actually changed, and why this is a now problem and not a forever one. For most of the time I’ve done this, looking done and being done were the same. The only way to make something look finished was to finish it, so the polish told you something true about what was underneath. That’s not the case anymore. You can get the polish without the rest of it, and the polish shows up first. So the moment you feel best about the thing, the demo runs clean, everyone nods, it looks great, is also the moment you know the least about whether it’ll hold.

I think about it as the part you can see versus the part you can’t. The demo is the bit on screen. The part that decides whether this survives a busy Tuesday is the bit that isn’t on screen, and you can’t really judge what isn’t there. The page renders, the button works, the obvious path is smooth, and there’s just no visual tell for the order that double-charges at scale or the data that quietly goes sideways when two people hit it at the same second.

The crossing almost never happens on purpose, which is what makes it sneaky. Nobody decides to turn their website into an application. It goes one sensible little feature at a time. You add accounts, so now you own logins and sessions and the guy who forgot his password. You add payments, so now a bug isn’t a typo, it’s money going the wrong way to actual customers. You start storing things people expect to stay accurate, and “works on my machine” stops meaning anything because the real question is whether it’s still right after a few thousand writes you never watched happen. Any one of those and the rules quietly change underneath you, and “it looks done” stops being worth much, because the thing that can go wrong isn’t the thing you can see.

So if you’re building something right now, the move isn’t to decide the tools are good or bad. They’re fine, I use them. Not for final products but for ideas and brainstorming. The move is to be honest with yourself early about which thing you’re actually making, before there’s real data sitting in it. If it just shows stuff, go nuts, ship it by lunch if you’re feeling it. If it does anything that has consequences, the part you can’t see is the part that matters, and that’s the part I’d slow down on. It’s important to remember that having it generated and actually understanding it when it breaks at 11pm are not the same situation, and you find out which one you’re in at the worst possible time.

That prospect’s a client now. We didn’t rebuild their site, we rebuilt the part that had turned into an application and left the part that was honestly still just a website alone. Figuring out where that line fell was most of the job.

Hopefully this helps answer the important questions early to avoid chaos down the road.

Scrolling feeds optimized by AI.
Reading posts created by AI.
With comments generated by AI.

I know I’m not the only one who gets exhausted by it quickly. AI has given us one of the most powerful toolkits in history, but it’s taken away the mess of being human. Who doesn’t love a good mess?

It’s a tough balance because this AI revolution has deeply impacted my work and what we do at And Zeros. At first, I saw it as a fear and a clock ticking down. It was only time until everyone stopped needing creative humans for their marketing and just prompt away. Fortunately, I found that wasn’t actually the case.

We quickly adapted our business to the new future. We spent the last year and a half learning AI-Search, now commonly known as GEO & AEO. Instead of using AI to do our work, we discovered how to help our clients rank in ChatGPT, Google AI Overviews, and other LLMs. That work has been extremely rewarding and kept us sharp. It’s not like the entire world of SEO flipped, but there’s a lot more involved now that help these rankings and meeting consumers where they are. It felt good not to just prompt and repeat, but to actually learn new skills that optimize for AI-discovery instead of using AI for our creativity.

I’ve been an analog addict my whole life. I started off my music journey recording on a 4-track Tascam Portastudio and cut tape and learned to record music on a Neve 8108 console and tape machines. I’ve always been drawn to the physicality and imperfections of analog music and consumption. It’s honestly what keeps me inspired.

Outside of a constantly growing cassette and vinyl collection, I’ve been collecting VHS tapes and DVDs for years and nothing beats the warmth of a CRT TV. It all feels so human and when art is physical, it does something to your brain.

I also shoot film photography and use actual cameras instead of strictly my phone. Nothing keeps you in the moment more than an actual camera.

Marketing used to be a similar process for me. When I entered the field, I used to spend hours, days, or sometimes even weeks brainstorming campaign ideas and messaging. A go-to-market strategy could take several months to develop. Now… people just talk about the AI agent to create your GTM in one click. I mean sure, it’s technical a GTM, but at what cost?

Humans crave analog experiences.
Humans crave analog connections.

The rise of the real-world experience is here and that’s absolutely hilarious to say. I see it several times a day on LinkedIn now. Thought Leaders talk about how it’s their new revelation! Oh, how industrious.

People who are “investing” in real-world experiences 😂 That’s just hilarious to me because of course you are. You’re investing in being human. We are humans, of course we want real-world experience.

So, while I am a big believer in the tools, just notice how I say tool. It’s the same as getting a new synth or plugin. It’s a palette to use and a color to add in.

The ones prompting with genuine creativity and innovation will still win.

I now take my breaks in my garden. Watering the tomato plants, bird baths, and filling up the bird seed. That’s where the real creativity comes from.

You can’t create if you don’t experience life.

Retrieval-Augmented Generation is the mechanism behind every AI citation. The engine breaks your prompt into narrower sub-queries, retrieves a candidate set of passages, scores each by relevance and authority, then synthesizes the answer and attributes the pieces it used. Teams that understand RAG mechanics optimize systematically. Everyone else optimizes by guessing.

Most people picture an AI engine reading their page the way a person would, top to bottom, weighing the argument. That is not what happens. The engine never sees your page as a page. It sees a pool of text fragments, ranked by math, and it pulls the few that answer the specific question in front of it. If you know how that pool gets built and ranked, you can write for it. If you do not, you are publishing into a process you cannot see.

What is RAG and how does it work?

Retrieval-Augmented Generation is a technique that lets a language model pull in external text before it answers, instead of relying only on what it learned during training. The model first retrieves relevant documents from an index or a live web search, then generates its response grounded in that retrieved material. This is the structure behind grounded answers in ChatGPT, Perplexity, Google’s AI surfaces, and Claude when web access is on.

The reason RAG exists is accuracy. A model working from training data alone has a fixed knowledge cutoff and a tendency to fabricate specifics. By blending generation with a retrieval step, the engine sticks closer to source material and can show its work. That last part matters for you! A system that retrieves before it answers is a system that has to choose sources, and any choice can be reverse-engineered.

There are two retrieval modes worth separating.

Real-time RAG fetches live web pages during the query, which is how Perplexity and ChatGPT’s search mode operate. Training-data answers come from pre-learned knowledge with no live fetch. The architecture decides the behavior. A page can sit at position one in Google and still never get cited, because the engine runs a separate evaluation for whether your text is extractable and trustworthy enough to fold into a written answer.

What happens between your prompt and the citation?

The whole thing runs in the time it takes the answer to start streaming, which is why it feels instant. Underneath, several steps fire in sequence. The engine analyzes intent, decomposes the prompt into sub-queries, runs those searches in parallel, assembles a candidate set of passages, scores them, and synthesizes a response that attributes the fragments it actually used.

The synthesis step is where citation happens. The model is not citing your domain because it respects your brand. It is citing the specific passage it lifted to satisfy one sub-query. Your page can contribute one sentence to an answer that pulls from five other sources, and that single contribution is your citation. This reframes the entire optimization target. You are not trying to win a page. You are trying to own a passage that answers a question cleanly enough to survive the scoring step.

Why does prompt decomposition matter for AEO?

Prompt decomposition, often called query fan-out, is the step where the engine breaks one complex prompt into several narrower sub-queries and searches each independently. A prompt like “best CRM for a small agency” does not run as one search. It fans out into questions about pricing, integrations, ease of use, small-team features, and real user reviews, even though the user typed none of those explicitly.

A single prompt can spawn anywhere from a handful to twenty or more sub-queries depending on complexity. Each one runs its own retrieval. This is the mechanic that breaks keyword-era thinking. You might rank beautifully for the headline phrase and stay invisible to the sub-queries that actually decide the citations, because a competitor answered those adjacent questions better than you did.

The practical move is to map the fan-out before you write.

List the questions a thorough answer would need to resolve, then make sure your content resolves each one in a discrete, self-contained section. Semrush ran a controlled test optimizing four articles specifically against fan-out queries and saw citations of those pieces more than double. The lesson is that owning the supporting questions, not just the headline term, is what gets you pulled into answers.

How does the engine score candidate sources?

Once the candidate passages are assembled, the engine ranks them on two axes that pull in different directions: relevance and authority.

Relevance is semantic match, measured by how close your passage sits to the sub-query in vector space, the mathematical representation of meaning. Authority is the trust layer, the signals that say this source is reliable enough to repeat.

Relevance usually does the heavy lifting. Research on ChatGPT citations across millions of prompts found that within a given retrieval set, freshness and authority matter, but relevance to the fanned-out queries is what determines whether a retrieved page actually gets cited rather than just fetched and ignored. A new page that matches the sub-queries well gets cited. A new page that does not gets retrieved and dropped.

Authority becomes the tie-breaker when relevance is close. In news queries, where many pages match the topic almost identically, engines fall back on page age and source reputation to break the tie. Academic work points the same direction: studies of ChatGPT’s source selection in scholarly contexts found it leans heavily on citation-count signals and well-known journals, amplifying sources that already carry consensus weight. The takeaway for your content is that relevance gets you into the candidate set, and authority decides close calls. You need both, in that order.

Why does answer block placement on your page change citation odds?

The engine retrieves passages, not pages, and it extracts the chunk that most cleanly answers the sub-query. If your answer is buried in the eighth paragraph behind setup and throat-clearing, the extractable unit is harder to isolate and easier to skip in favor of a competitor who stated the answer directly. Structure is not decoration here. It is what makes a passage machine-parsable.

This is the entire case for answer-first writing. Lead a section with a direct, declarative statement that contains the exact phrasing of the question, then support it underneath. A section that opens with its own answer is a clean extraction target. A section that winds up to its answer forces the engine to do work it will often skip. The H2s on your page function like an index of questions; each one should be answerable from the first lines beneath it without the engine needing the rest of the section.

What signals does each engine weight differently?

The engines do not behave alike, and treating them as one target wastes effort. Perplexity cites generously, averaging around 6.6 sources per response, and holds onto those citations longer. ChatGPT cites far fewer, typically three to four domains per response, and churns through them fast. Google’s AI surfaces cite the most, often more than a dozen domains, with a more stable core set per prompt.

Source turnover is the sharpest difference.

SISTRIX tracked source stability across 82,619 prompts over 17 weeks and found ChatGPT replaces roughly 74% of its cited domains every week, while Google replaces around 56%. The median ChatGPT prompt does not hold a single domain across all 17 weeks. Google’s AI Mode, by contrast, tends to keep a stable core of a couple of domains per prompt plus a rotating carousel.

There is also a freshness split. Analysis of millions of citations found ChatGPT skews toward fresher content than Google’s organic results by a wide margin, yet the median age of pages it cites still lands around 500 days, with some cited pages over seven years old. Freshness is a strong signal, especially for news, but it does not override relevance. The strategic read is that platform choice changes your tactics more than your industry does. A piece optimized for Perplexity’s durable, multi-source behavior is a different artifact from one chasing ChatGPT’s fast-rotating, few-source answers.

How can you reverse-engineer which prompts your content can win?

Start from the fan-out, not the keyword. Take a prompt your audience actually types and decompose it yourself into the sub-queries an engine would generate. Free fan-out simulators exist for this, and so does manual work: write down every question a complete answer would have to resolve. That list is your real target set.

Then check which of those sub-queries you already answer cleanly and which you do not. Run the prompt through the engines you care about and read which domains get cited for each facet. The sources winning the sub-queries you are missing are your direct competition for that answer, and the gap between their passages and yours is your edit list. This is more honest than a rank check, because it tells you which specific questions your content can credibly win rather than which phrases you happen to rank for.

What can you change this week based on RAG mechanics?

Three moves, none of which require a rebuild. First, convert your H2s into the actual questions your audience asks, and make the first two sentences under each one a complete, standalone answer. That single change turns vague sections into clean extraction targets.

Second, map the fan-out for your two or three highest-value pages and add sections for the sub-queries you are not yet answering. You are filling the gaps that decide citations, not adding filler. Third, pick your platform priority deliberately. If your audience lives in Perplexity, lean into depth and durable, well-sourced passages that survive its longer citation window. If they live in ChatGPT, accept the fast churn and build a publishing cadence that re-earns citations rather than expecting one piece to hold. The mechanics are knowable. The teams that act on them stop guessing.

FAQ

Is RAG the same as web search inside AI?

Not exactly. Web search is one source RAG can retrieve from, but RAG is the broader pattern of fetching external text and grounding the answer in it. That text can come from a live web crawl, a private document store, or a vector index. When ChatGPT or Perplexity searches the web mid-answer, that is RAG using web search as its retrieval layer. RAG over a company’s internal docs uses no web search at all.

Why do AI engines cite different sources for the same prompt?

Because each engine decomposes the prompt differently, retrieves from different indexes, and weights relevance, authority, and freshness on its own curve. ChatGPT pulls three to four fast-rotating domains, Perplexity pulls around 6.6 and holds them longer, and Google’s AI Mode pulls a dozen or more with a stable core. Same question, different fan-out, different scoring, different citations. Source rotation compounds it: ChatGPT alone swaps roughly 74% of its cited domains week to week.

Does the citation half-life data still apply?

Yes, and it is more useful than a single fabricated number. Research on 3.5 million citation events puts ChatGPT’s citation half-life at about 3.4 weeks, the fastest of the major platforms, with Perplexity nearly 70% longer at 5.8 weeks and Google’s surfaces clustered in the four-to-five-week range. The practical meaning is that a ChatGPT citation needs re-earning roughly monthly, while a Perplexity citation works for you longer. Plan your publishing cadence around the platform you are optimizing for.

ChatGPT reads Reddit constantly and credits it almost never. Ahrefs studied 1.4 million prompts and found Reddit gets cited just 1.93% of the time, yet accounts for 67.8% of every page the model pulls in and never names. That gap between getting read and getting credited is the part of AI search most teams are measuring wrong.

There are two things happening when ChatGPT answers a question, and they look the same from the outside but they are not. The first is retrieval. Your page gets pulled into the model’s working set, read, and used to shape the answer. The second is citation Your brand gets named, with a clickable link the user actually sees.

Most teams chase the first and assume it gets them the second. It does not 🙂

ChatGPT cites only about half the pages it retrieves, and the half it drops is doing real work in the answer without ever showing up. Nothing demonstrates that better than Reddit.

What the study found

Ahrefs ran 1.4 million ChatGPT prompts and tracked which retrieved pages made it into the final answer as citations. Reddit landed at the bottom… a 1.93% citation rate against more than 16 million retrieval events, the second-highest volume of any source. Put differently, 67.8% of every page ChatGPT pulls in and then refuses to credit comes from Reddit.

The model is mining Reddit for what people actually think, then handing the citation to a more presentable source. Ahrefs put it plainly: ChatGPT uses Reddit to understand topics and gauge consensus, but almost never gives it credit. It reads the room, forms a view, then footnotes an institution.

This makes sense when you look at what people ask AI. A lot of prompts are not fact lookups. They are recommendation and comparison questions; which tool is best for a small team, how does this service actually feel to use. Reddit has the raw language for those answers. ChatGPT borrows it, then cites someone in a suit.

Where citations actually come from

The same study found 88% of ChatGPT’s citations flow through one channel; ordinary web search. Not its Reddit feed, not YouTube, not academic sources. The general search index, the same one you have been trying to rank in for years.

So the entry ticket to being cited is still ranking. That is not the old SEO game wearing a new label, but it is not a different game either. If you do not show up in conventional organic results for a query, the channel that produces nearly nine in ten citations is closed to you before the model reads a word of your page.

Once you are in that pool, one small thing moves the needle more than it should. Pages with plain, readable URLs got cited 89.78% of the time. Pages with opaque ones, 81.11%. ChatGPT screens the title and the URL before it ever opens the page, so a descriptive slug is a nearly nine-point swing that costs nothing but a habit at publish time.

There is a freshness wrinkle worth knowing. Within any single answer, the pages that got cited skewed older, around 500 days on average, while the freshest pages got read and discarded. ChatGPT likes fresh content in general, but inside a given answer it leans on established pages that match the question well. Relevance carries the weight. Freshness only breaks ties.

The one idea to take from this

Stop treating AI visibility as a single number. Retrieval and citation are two different funnels with two different jobs, and optimizing one does almost nothing for the other.

Retrieval is being in the conversation. You earn it with breadth, ie: showing up where the model gathers context, Reddit included. Citation is being named in the verdict. You earn it with precision; ranking in search, writing titles that match the specific question being asked, and keeping your URLs clean.

The diagnostic that matters is the ratio between the two. How often are you retrieved, and what share of that turns into a citation? A page that gets read constantly but cited rarely is doing free work for the model and getting nothing back. That is the Reddit problem, and it can be your own content’s problem too. A blended dashboard hides it. Splitting the two metrics is the only way to see where your credit is leaking.

If retrieval is the signal entering the board, citation is the moment it crosses into the mix loud enough to hear. Plenty of tracks get recorded. Fewer make the master.

What to do this week

Pick the pages you most want cited and check whether they rank organically for their target query. If they do not, citation is effectively closed until they do, so ranking comes first.

Then take each page’s keyword and break it into the narrower questions a user is really asking underneath it. Rewrite the title to answer the sharpest one, not the broad term. ChatGPT splits every prompt into sub-questions before it retrieves, and it cites the pages that answer those, not the ones that vaguely cover the territory.

Fix your URLs going forward. Descriptive slug, every time. Nine points for free.

And separate your measurement! Track retrieval and citation as two columns and watch the ratio. That number tells you exactly which pages to rewrite.

One caveat, because the data deserves it. This study ran on ChatGPT 5.2 in February 2025, and Ahrefs flags that newer versions have already shifted citation behavior. The specific percentages will drift. The structure will not. Retrieval and citation are separate funnels with separate rules, and that is the part worth building around.

FAQ

Should I stop investing in Reddit?

No. A low citation rate is not low influence. ChatGPT reads Reddit to form the opinions it then attributes to other sources, so your presence there shapes what the model says about you even without a visible link. Treat Reddit as a sentiment input, not a citation output. The mistake is expecting links, not getting them, and walking away from the job Reddit actually does.

Does this hold for Claude and Perplexity?

The Reddit numbers are ChatGPT-specific, but the retrieve-more-than-you-cite pattern shows up across every engine. Each one favors different sources and applies its own selection logic, so measure them separately rather than assuming one model describes all of them. The gap is universal, but the source preferences are not.

How do I measure my own ratio?

Use a tool that separates retrieved pages from cited pages instead of reporting one blended score. Compare how often your URLs enter the working set against how often they get named, and track it per page over time. Several platforms now expose this directly and let you filter for prompts where a competitor is cited and you are not, which hands you a concrete list of pages to fix.