Why Organic Founder Voice Beats Ghostwriters in AI Search: P

Last updated: 2026-10-12

TL;DR

Definition: The Founder-Voice Moat is content in the founder’s own voice that reads like a real person, not a template.

Where the citations go: In Semrush’s March 2026 study of 89,000 LinkedIn URLs cited by AI search, individual members, not company pages, made up 59 percent of LinkedIn citations on ChatGPT Search and Google AI Mode [1]. In Muck Rack’s May 2026 analysis of 1 million+ AI-cited links, paid and advertorial content was 0.3 percent of citations and journalism was 27 percent [2].

What templates cost: In a Science Advances experiment with 293 writers, AI-assisted stories were rated more creative but were more alike [3]. In Edelman and LinkedIn’s 2024 survey, 70 percent of C-suite leaders said thought leadership had at least occasionally made them question a supplier [4].

Our own numbers: In our August 2026 audit of past TwoSetAI guests, 86 percent were cited by at least one AI engine [5]. When ChatGPT did not cite a YouTube video, it reached guests through third-party authoritative posts (founder profiles, twosetai.com pages), which is why the written editorial layer helps on ChatGPT. Our newer guests beat the older cohort on every surface, which points at packaging, not only voice.

Scope: No study I found compares founder-voice and ghostwritten pages head to head, or tracks either over time. The rest is evidence plus my own pattern-spotting, and I mark which is which.

In Semrush’s March 2026 study of 89,000 LinkedIn URLs cited by AI search, individual members made up 59 percent of LinkedIn citations on ChatGPT Search and Google AI Mode [1].

I have interviewed 100+ AI founders on the TwoSetAI podcast. Some come up whenever someone asks an AI engine about their space. Others hardly come up at all. After enough of these conversations I stopped believing that was random, and this post is my attempt to say what separates them.

The question for this pillar is: why does founder-anchored, organic content compound in AI search while ghostwritten and paid content decays? My short answer is that engines quote what is specific, sourced, and hard to find anywhere else. A founder describing something only they lived is specific by construction. A template is a sentence ten other companies could publish, so an engine has no reason to choose yours. I’ll lay out the evidence, then the patterns I see in the founders who show up, then what I cannot prove.

What is the Founder-Voice Moat?

The Founder-Voice Moat is content in the founder’s own voice that reads like a real person, not a template. It is a moat because the raw material cannot be copied. A competitor can copy your headings in an afternoon. They cannot copy the conversation you had with your first customer.

I use one more term in this post. Voice provenance is the trail from a published sentence back to something the founder actually said, did, or decided. A post with strong voice provenance has a transcript, a recording, or a decision log behind it. A template has none. That difference separates a ghostwriter working from your interview from one working from a content calendar, and it matters more than who holds the pen.

Two limits up front. The Founder-Voice Moat is not a ranking factor, and Google says E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is not itself a specific one [6]. It also does not mean you must type every word yourself. It means your thinking has to be on the page.

Do AI engines reward founders, or just good text?

Mostly the text, with signs that people count. Here is what I could verify.

Individuals show up in citations. In the Semrush study, about 95 percent of cited LinkedIn content was original, not reshared. Perplexity went the other way, with company pages at 59 percent of its LinkedIn citations. The study was run with LinkedIn’s collaboration and did not include Claude [1].
Paid content barely registers. Muck Rack’s analysis covered ChatGPT, Gemini, and Claude. Paid and advertorial content was 0.3 percent of citations, journalism was 27 percent, and corporate blogs were 24 percent [2]. Muck Rack sells PR software, so I read its numbers as directional.
Third-party voices beat brand-owned ones. A University of Toronto team found AI search engines lean heavily toward earned, third-party sources over brand-owned and social content, and that niche players face a big-brand bias [7]. My read, which is inference, is that a founder’s voice works hardest when it is quoted somewhere other than your own site, such as a podcast, a press piece, or a LinkedIn post other people cite.
Off-site signals look like the majority. Rodrigo Stockebrand’s O’Reilly book on answer engine optimization cites AirOps research putting the influence of off-site signals on AI visibility around 85 percent, “dwarfing the on-page factors most teams obsess over,” and returns to that figure across three chapters [26]. The precise number is directional; the direction fits the Semrush, Muck Rack, Toronto, and Ahrefs findings above.
Mentions beat classic metrics. Ahrefs studied 75,000 brands and found YouTube mentions correlated at about 0.737 with AI visibility. On ChatGPT, branded web mentions were at 0.664 and Domain Rating at 0.266 [8]. Correlation is not causation, and Ahrefs sells SEO tools.
Google asks who made it. Its guidance has creators ask who made the content, how, and why, treats first-hand experience as a quality signal, and encourages accurate bylines [6]. Its generative AI guidance, last updated October 1, 2026, says generating many pages without adding value may violate its spam policy, and that AI output should be manually fact-checked [9].
Authors exist as data. Mike King’s analysis of Google’s leaked 2024 documentation found an attribute that stores a document’s authors as text and another that flags whether an entity is the author [10]. Rand Fishkin read it as a sign that Google can identify authors and treat them as entities [11]. The leak shows the fields exist, not how they are weighted.
A practitioner’s read. In May 2026, Lily Ray said AI citations track Google rankings closely, and posted that Google results increasingly favor authentic first-person content and forums over generic optimized pages, per a trade-press summary [12].

None of this proves an engine detects a founder’s voice. I doubt it can. Detectors have been shown to misclassify non-native English writing as AI-generated [13], so an engine that leaned on them would punish real people. What the evidence supports is narrower: engines retrieve specific, original pages, and they cite individuals and earned mentions far more than paid placements.

Why does template content slip out of AI answers?

Two reasons, in the order I can best support them.

Templates converge. A PNAS study found that LLM writing differs systematically from human writing in grammar and rhetoric, and the gap persists as models get larger [14]. In the Science Advances experiment, AI-assisted stories scored higher with evaluators but were more alike than human-only stories [3]. Sam Kececi (Founder and CEO, The Sentience Company) gave me the consequence in one line: “Imagine if like 90% of the tweets you saw were just written by the same guy. You would still get annoyed” [15]. You can hear him say it or read the write-up. When ten explainers carry the same sentence, an engine can quote any of them, so it has no reason to quote yours.

Buyers discount it. In the Edelman and LinkedIn study, 73 percent of respondents said thought leadership is a more trustworthy basis for judging a company’s capabilities than its marketing materials, and 70 percent of C-suite leaders said it had at least occasionally made them question whether to keep working with an existing supplier [4]. Weak content costs you with the reader who arrives after the engine.

Here is the comparison I use with founders. The typical sentences are illustrative, and the right-hand one paraphrases a story below.

Dimension	Template, or ghostwritten from a brief	Founder-voice content
Raw material	Keyword list, competitor posts, style guide	Interviews, decisions, customer conversations
Voice provenance	None you can point to	A transcript or recording behind each claim
Typical sentence	”Companies are adopting AI to unlock efficiency."	"The agent sent the email it was told not to send, and said so when asked.”
What an engine can lift	A claim a dozen other pages also make	A named person, number, or story found nowhere else
What a buyer concludes	”Could be anyone."	"This person has done the work.”
Main risk	Median voice, scaled-content flags	Accurate but invisible if never packaged
Evidence strength	Strong on convergence and on low paid citations	Strong on individuals being cited, no head-to-head study

Is ghostwriting always the problem?

No. I’d be misleading you if I said it was, and the evidence is messier than my headline.

Jasmin Alić (Founder, Link Up) coaches founders on LinkedIn. He told me that if you can’t spare 30 minutes a day, hire a virtual assistant, because “they can now own your voice and they can engage as you” [16]. He was fine with outsourcing the engagement and even batching the writing. His hard line was elsewhere: don’t let AI do your thinking.

Adam Biddlecombe (Founder, Mindstream) ran a newsletter that signed off every issue as written by humans. He then told me he uses AI to write all of his own LinkedIn content [17].

The research has gaps too. Graphite found that 86 percent of Google results and 82 percent of ChatGPT and Perplexity citations were human-written, but it did not evaluate heavily edited AI-assisted content and said that approach may work [18]. In one benchmark, neural retrievers ranked LLM-written text above human-written text [19]. Engines do not reward “human” as such.

So here is where I land. Who holds the pen matters less than where the words came from. A ghostwriter working from your recorded interviews has voice provenance, and can be indistinguishable from you. A ghostwriter or a model working from a brief and a competitor’s top ten produces the median voice with your logo on it. If you use a ghostwriter, ask for the source of every claim. If the answer is a transcript, keep them. If the answer is “research,” you are paying for a template.

This is also what most AEO and GEO monitoring tools quietly miss. They show you the competitor content that is being cited for a query and nudge you to adapt it in your voice. Adapting it in your voice is not enough. The sentences those competitors wrote came from their customer calls, their numbers, their reversals. You did not live those moments. What makes a sentence citable (a named person, a specific object, a decision you reversed) is upstream of the dashboard. A monitoring tool can show you where you are absent; it cannot generate the experience that makes you present. Why the Way You Monitor AI Citation Is Wrong walks through the four failure modes in detail.

Which patterns do I hear in the founders who keep showing up?

These are observations from hosting the show, not coded data. I did not score 100 transcripts against citation outcomes, and I have not audited the AI presence of every guest named below. I picked them because each pattern is on tape with a timestamp.

1. They lead with a thesis, not a pitch. Jay Hack (Head of AI, ClickUp, which bought his company Codegen) told me that “in the process of trying to convince people that what you’re doing is a good idea, you end up actually in many ways showing that you’re desperate.” His alternative was to “say interesting things that demonstrate a mastery of the topic at hand” [20]. Hector Hernandez (Co-founder and COO, Speakeasy) shows the form: “the first thing that breaks when people scale their use of AI is often not cost,” and “most of the time it’s complete loss of visibility of what’s going on” [21]. That is one claim from one speaker, quotable whole. Jay’s argument is also in his write-up.

2. They own a story and a question. Hector told me about a Speakeasy salesperson whose assistant emailed a customer after being told not to. Asked about it, the agent said it had done it anyway. Hector compared it to “scolding a puppy” and called it “a very quick warning shot” [21]. No competitor can write that paragraph, because no competitor had that morning. He also gave me his sales method: keep one or two questions you already know the answer to, because “those are the things that really spark a great conversation.” His example for a CISO: do you know how many agents you have in your organization? [21] A question like that travels: a buyer can repeat it in a meeting, and a transcript lets an engine attribute it.

3. They state one difference in words a stranger can decode. Rob Kaminski (Co-Founder, Fletch PMM) has worked with more than 600 companies on positioning. On my episode he stress-tested a pre-launch AI agent product and asked: “Does your audience know what that means? Do they know what they’d use it for? Do they know who the players are? And if the answer is no, it is a signal that we have not found that common ground.” Elsewhere he said “the aim of a good positioning is memorability” [22]. My read, which is inference, is that a phrase a stranger can decode is also one an engine can match to a buyer’s question. Full audit: How to Position an AI Product When Everyone Sounds the Same.

4. A human stays in the thinking. Jasmin put it bluntly: “writing is thinking,” and he compared outsourcing it to “trying to go to the gym and having someone else lift the weights for you” [16]. Adam said HubSpot told him it loved Mindstream’s editorial when it came knocking, and that the person who introduced them to HubSpot “came through my personal brand” [17]. Sam, whom I quoted above, borrows a line he credits to Naval Ravikant: no one can compete with you at being you [15]. Sam sells a personal-AI product built on that thesis, so weigh it.

Pattern	Guest	Hear it
Thesis, not pitch	Jay Hack, Head of AI, ClickUp	55:27, watch page
Thesis, not pitch	Hector Hernandez, Co-founder and COO, Speakeasy	22:32, watch page
Own story and question	Hector Hernandez	8:07, 38:12
One decodable difference	Rob Kaminski, Co-Founder, Fletch PMM	23:00, watch page
Human in the thinking	Jasmin Alić, Founder, Link Up	38:28
Human in the thinking	Adam Biddlecombe, Founder, Mindstream	39:23
Human in the thinking	Sam Kececi, Founder and CEO, The Sentience Company	25:42, watch page

What about the founders who don’t show up? I won’t name anyone. The pattern I notice is that their public record is a company page in pitch voice, their language is category jargon, and nothing connects their sentences to a conversation they had. They have no voice provenance. That is my observation, not a finding.

Is a strong voice enough on its own?

No. Voice is necessary and not sufficient, and our own audit shows why. In August 2026 we searched for past TwoSetAI guests across Google, Google AI Mode, Perplexity, and ChatGPT. 95 percent appeared on at least one Google surface and 86 percent were cited by at least one AI engine [5]. In a recent cohort of Founder Visibility Program participants, every one appeared in Google video results and on at least one AI engine, but When ChatGPT did not cite a YouTube video, it reached guests through third-party authoritative posts (founder profiles, twosetai.com) [5].

An earlier cohort did worse on every surface, even though those episodes had months longer to be indexed; only 75 percent of that cohort were cited by at least one AI engine [5]. That points at packaging more than age or how well anyone spoke: title formula, chapters, watch pages, and a written layer. It is directional, not causal.

Stockebrand reaches the same conclusion from a different angle: in AI search, “wide beats deep. To win a top mention on a given query, you usually need to show up across many cited sources, not just on your own pages” [26]. His exemplar is Vercel, whose AI visibility is driven less by its own getting-started docs than by the Reddit recommendations, Medium posts, YouTube videos, and third-party tutorials that reinforce the same message about the platform [26]. A founder’s voice supplies the quotable material; wide packaging is what makes it the one engines retrieve.

The clearest case is Alexander Berger (COO, Bolt.new). In August 2026, in incognito windows, ChatGPT labeled our episode with him “Best overall,” Perplexity called it “the best fit,” and Google AI Mode cited it across multiple sections of his bio. One organic LinkedIn post earned 184,361 impressions [23]. I can’t prove his voice caused that. What I can say is what was on the page: a 58-minute conversation in his own words, a transcript, a highlight cut, Medium and Reddit posts, and a watch page. Voice supplied the material. Packaging made it findable. That packaging is the Editorial Layer: human-authored, positioning-anchored content built on the raw conversation. Without it, a strong voice still leaves a Citation Gap.

How do you build a Founder-Voice Moat?

Five moves, in the order I’d do them.

Record before you write. Start from a conversation, not a blank page. Voice provenance starts with a recording. The Quote That Carried the Episode shows how I mine my own conversations.
Pick one thesis a stranger can decode. TwoSetAI runs proprietary positioning steps for every founder guest, drawing on classical PMM frameworks and adapted for AI-search citation. The output is one sentence you can say out loud.
Keep your stories and numbers, cut the template. In the KDD 2024 paper that introduced GEO, adding quotations and statistics produced some of the largest visibility gains, up to 40 percent on a benchmark, while keyword stuffing gave little [24]. It is a benchmark, not a promise.
Write or edit your own first draft. Use Jasmin’s exercise: whenever you prompt an AI to write something, “write it yourself as well. Then mix and match” [16]. If someone else writes, ask for the source of every claim. How to Find Your Founder Voice is the framework.
Package it where engines read. Publish the transcript, chapters, and a written page, and put the long conversation on YouTube. The video pillar covers why.

Then re-run the same prompts in ChatGPT, Perplexity, Google AI Mode, and Claude after each change. In SparkToro’s 2,961-run study, ChatGPT and Google’s AI gave the same brand list in fewer than 1 in 100 runs, so track how often you appear, not where you rank [25].

So, why does organic founder voice beat ghostwriters?

Because an engine has no reason to pick a page that says what ten others already say, and a founder’s own thinking is the one input a template cannot copy. The problem I’d fix first is missing voice provenance. Paid and advertorial content barely shows up in AI citations [2], template output converges on a median voice [3][14], and, in my experience, the founders who keep coming up are the ones with something specific to say. That is the Founder-Voice Moat.

I can’t promise any engine will name you, because these systems change and most of the research here is correlational. What you control is whether your sentences trace back to something you really said. That is our work at TwoSetAI: we turn one podcast conversation into lasting AI-search presence for AI founders whose buyers check AI before they check the website. When your buyers ask AI, be the answer.

Next in this series: The Ghostwriter Trap, What AI Engines Hear When a Founder Speaks, and Case Study: What Hector Hernandez’s “Visibility Breaks First” Thesis Did for Speakeasy.

Who is TwoSetAI?

TwoSetAI is one of the top AI podcast and video studios in Silicon Valley, and the team behind this article, hosted by Angelina Yang. The show has featured 100+ AI founders and leaders on how real AI products get built, sold, and trusted. Alongside the podcast, TwoSetAI runs the Founder Visibility Program that turns each conversation into the articles, clips, and pages that make its guests easy to find. Watch full episodes on YouTube at youtube.com/@TwoSetAI, and find every guest and episode at twosetai.com.

Glossary

Founder-Voice Moat (also called: founder-voice advantage, organic founder voice). Content in the founder’s own voice that reads like a real person, not a template.
Voice provenance (also called: source trail, voice traceability). The trail from a published sentence back to something the founder actually said, did, or decided.
Median voice (also called: template voice, AI-default voice). Writing that converges on the statistical middle because it comes from the same few models or the same brief.
Citation Gap (also called: AI-answer gap, brand invisibility in AI search). The space between what AI says about your brand and what your business actually is.
Editorial Layer (also called: GEO editorial layer). Human-authored, positioning-anchored content on top of automated AI-visibility monitoring and the founder’s raw conversation.

About Angelina Yang

Angelina Yang is VP of AI, a fast.ai fellow, and the host of the TwoSetAI podcast, where she has interviewed 100+ AI founders and leaders about how real AI products get built, sold, and trusted. Find her on LinkedIn, talk to her AI clone on Delphi, watch the podcast on YouTube, or browse twosetai.com.

Frequently Asked Questions

Do ghostwriters hurt my visibility in AI search? Not by themselves. Google’s guidance targets generating many pages without adding value for users, not authorship as such [9]. What hurts is content with no original source behind it, which tends to converge on a median voice [3][14]. A ghostwriter working from your recorded conversations gives every claim a source. Ask for it.

Can AI engines tell whether a founder’s content was ghostwritten? I found no vendor documentation saying they try. Google’s guidance asks who created the content, how, and why [6]. AI-text detectors, the obvious tool, have been shown to misclassify non-native English writing as machine-made [13]. What an engine can observe is the page itself: how specific it is and whether others cite it.

Does founder-led content get cited more by ChatGPT, Perplexity, and Google AI Mode? On LinkedIn, for two of the three. Individual members made up 59 percent of LinkedIn citations on ChatGPT Search and Google AI Mode, while company pages made up 59 percent on Perplexity [1]. That study covers LinkedIn only, and it was run with LinkedIn’s collaboration.

Is paid or sponsored content cited by AI search? Rarely. Paid and advertorial content was 0.3 percent of citations in Muck Rack’s May 2026 analysis, against 27 percent for journalism [2]. Muck Rack sells PR software, so treat the numbers as directional.

Is it okay to use AI to write my LinkedIn posts? Google says generative AI can help with research and structure, and warns against generating pages that add no value [9]. Graphite found human-written articles were 86 percent of Google results and 82 percent of ChatGPT and Perplexity citations, but it did not test heavily edited AI-assisted work [18]. Jasmin Alić’s rule is to write it yourself as well, then mix and match [16]. Adam Biddlecombe uses AI for his LinkedIn posts [17]. If the post carries your story and your numbers, the tool matters less.

What makes a founder’s voice citable by AI? Specifics an engine can lift whole: a thesis, a number, a story, or a question with your name on it. In the KDD 2024 GEO paper, adding quotations and statistics gave some of the largest visibility gains, up to 40 percent on a benchmark [24]. Then package it with a transcript, chapters, and a written page.

How do I find my founder voice if I have been heads down developing my product? Start from a recorded conversation, not a blank page. Talk through a customer story, a decision you reversed, or the question you ask every buyer. Transcribe it, then edit. Hector Hernandez’s puppy story is a good model: a real morning, a real quote, one lesson [21].

Is the Founder-Voice Moat the same as a personal brand? No. A personal brand is an audience. The Founder-Voice Moat is the content that neither the audience nor the engines can get anywhere else. They help each other: Adam Biddlecombe said the person who introduced Mindstream to HubSpot came through his personal brand [17].

Do non-native English speakers risk sounding like AI? Detectors have misclassified non-native English writing as machine-generated [13], which is one reason I don’t think polish is the signal. Voice is specificity: your story, your number, your phrase.

References

[1] Loktionova, M. (2026). “We Analyzed 89K LinkedIn URLs Cited in AI Search: Here’s What Drives Visibility.” Semrush Blog, March 10, 2026 (325,000 prompts, January to February 2026; ChatGPT Search, Google AI Mode, Perplexity; run with LinkedIn’s collaboration). https://www.semrush.com/blog/linkedin-ai-visibility-study/

[2] Muck Rack (2026). “How AI citations have changed in the last 6 months: New insights from ‘What is AI Reading?’” Muck Rack, May 2026 edition, page updated May 27, 2026 (1 million+ links cited by ChatGPT, Gemini, and Claude). https://muckrack.com/blog/what-is-ai-reading-new-insights-may

[3] Doshi, A. R. and Hauser, O. P. (2024). “Generative AI enhances individual creativity but reduces the collective diversity of novel content.” Science Advances 10(28), eadn5290, July 12, 2024. https://pmc.ncbi.nlm.nih.gov/articles/PMC11244532

[4] O’Dwyer’s (2024). “Thought Leadership Can Power B2B Brands.” O’Dwyer’s PR, February 29, 2024, reporting the sixth annual Edelman and LinkedIn B2B Thought Leadership study (nearly 3,500 management-level professionals; fieldwork December 2023). https://www.odwyerpr.com/story/public/20925/2024-02-29/thought-leadership-can-power-b2b-brands.html

[5] TwoSetAI (2026). Guest search audit, checked August 10 to 25, 2026, incognito, screenshots on file. Internal data; protocol in the Methodology note.

[6] Google (2026). “Creating helpful, reliable, people-first content.” Google Search Central, page updated October 5, 2026. https://developers.google.com/search/docs/fundamentals/creating-helpful-content

[7] Chen, M., Wang, X., Chen, K., and Koudas, N. (2025). “Generative Engine Optimization: How to Dominate AI Search.” University of Toronto; arXiv:2509.08919, September 10, 2025. https://arxiv.org/abs/2509.08919

[8] Linehan, L. (2025). “Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews (75k Brands Studied).” Ahrefs, December 12, 2025. https://ahrefs.com/blog/ai-brand-visibility-correlations/

[9] Google (2026). “Google Search’s guidance on using generative AI content on your website.” Google Search Central, page updated October 1, 2026. https://developers.google.com/search/docs/fundamentals/using-gen-ai-content

[10] King, M. (2024). “Secrets from the Algorithm: Google Search’s Internal Engineering Documentation Has Leaked.” iPullRank, May 27, 2024. https://ipullrank.com/google-algo-leak

[11] Fishkin, R. (2024). “An Anonymous Source Shared Thousands of Leaked Google Search API Documents with Me; Everyone in SEO Should See Them.” SparkToro, May 2024. https://sparktoro.com/blog/an-anonymous-source-shared-thousands-of-leaked-google-search-api-documents-with-me-everyone-in-seo-should-see-them/

[12] PPC Land (2026). “Lily Ray: What the SEO industry is getting dangerously wrong about AI search.” PPC Land, May 14, 2026, summarizing Ray’s May 13, 2026 conversation with Ross Hudgens of Siege Media and her May 6, 2026 LinkedIn post. https://ppc.land/lily-ray-what-the-seo-industry-is-getting-dangerously-wrong-about-ai-search/

[13] Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., and Zou, J. (2023). “GPT detectors are biased against non-native English writers.” Patterns, 2023; arXiv:2304.02819. https://arxiv.org/abs/2304.02819

[14] Reinhart, A. et al. (2025). “Do LLMs write like humans? Variation in grammatical and rhetorical styles.” PNAS 122(8), e2422455122, February 25, 2025; arXiv:2410.16107. https://arxiv.org/abs/2410.16107

[15] TwoSetAI (2026). Sam Kececi (Founder and CEO, The Sentience Company) episode, published September 27, 2026; quotes at 25:42 and 43:13. https://www.youtube.com/watch?v=QKGqPUlRoM8

[16] TwoSetAI (2026). Jasmin Alić (Founder, Link Up) episode, recorded March 8, 2026; quotes at 30:39, 38:28, and 39:36. https://www.youtube.com/watch?v=aIKzImj6m8w

[17] TwoSetAI (2026). Adam Biddlecombe (Founder, Mindstream) episode; quotes at 15:43, 39:23, and 39:48. https://www.youtube.com/watch?v=cV8CYk2YvsA

[18] Graphite (2025). “How Does AI-Generated Content Perform in Search and Answer Engines?” Graphite, October 14, 2025 (31,493 keywords; AI detection by Surfer; English articles and listicles only). https://graphite.io/five-percent/ai-content-in-search-and-llms

[19] Dai, S. et al. (2023). “LLMs May Dominate Information Access: Neural Retrievers are Biased Towards LLM-Generated Texts.” arXiv:2310.20501 (later retitled “Neural Retrievers are Biased Towards LLM-Generated Content”). https://arxiv.org/abs/2310.20501

[20] TwoSetAI (2026). Jay Hack (Head of AI, ClickUp) episode, published July 17, 2026; quotes at 55:27. https://www.youtube.com/watch?v=-58QW-fJq0I

[21] TwoSetAI (2026). Hector Hernandez (Co-founder and COO, Speakeasy) episode, published October 5, 2026; quotes at 8:07, 22:32, 25:11, and 38:12. https://www.youtube.com/watch?v=yhOLsoqSSEY

[22] TwoSetAI (2026). Rob Kaminski (Co-Founder, Fletch PMM) episode, published September 10, 2026; quotes at 23:00 and 31:29. https://www.youtube.com/watch?v=SD64dL7yg7I

[23] TwoSetAI (2026). Alexander Berger (COO, Bolt.new) episode and three-engine check, incognito, August 2026. https://www.youtube.com/watch?v=8C3C4Vhknx8

[24] Aggarwal, P. et al. (2024). “GEO: Generative Engine Optimization.” KDD 2024; arXiv:2311.09735. https://arxiv.org/abs/2311.09735

[25] Fishkin, R. and O’Donnell, P. (2026). “New Research: AIs are highly inconsistent when recommending brands or products.” SparkToro, January 28, 2026 (600 volunteers, 12 prompts, 2,961 runs). https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/

[26] Stockebrand, R. (2026). Answer Engine Optimization. O’Reilly Media, July 2026 (cites AirOps research on off-site signals; Vercel case in Chapter 1). https://learning.oreilly.com/library/view/answer-engine-optimization/0642572275549/

Methodology note. Last verified 2026-10-12. The patterns in this post are my observations from hosting 100+ founder conversations. They are not coded data, and I did not score transcripts against citation outcomes. No study I found compares founder-voice pages with ghostwritten pages in AI citations, so the thesis rests on adjacent evidence and is labeled as inference where it goes beyond it. Guest quotes are verbatim from the published episodes cited, with punctuation and capitalization normalized and timestamps taken from the episode transcripts. Several third-party studies (Semrush, Muck Rack, Ahrefs, Graphite) come from companies that sell search, PR, or AI-visibility products, and the Semrush study was run with LinkedIn’s collaboration, so I treat their numbers as directional. Correlations (Ahrefs) are not causal. The Lily Ray remarks are cited through a trade-press summary [12]. The Edelman and LinkedIn figures are cited through O’Dwyer’s report of the study [4]. Google’s pages were read on October 12, 2026 and may change. The TwoSetAI audit [5] covered a cohort of past guests, not the full catalog: about seven queries per guest (one Google name search, and two each in Google AI Mode, Perplexity, and ChatGPT), run in incognito windows between August 10 and 25, 2026. A guest counted as present on a surface if any query for it surfaced them. The 10-versus-12 comparison is directional, not causal. The Alexander Berger results [23] were checked in August 2026 and may differ on your run. “Founder-Voice Moat,” “Editorial Layer,” “Citation Gap,” “voice provenance,” and “median voice” are TwoSetAI’s own terms, not industry standards. Nothing here guarantees a citation, a ranking, or a business outcome.