All In
FR EN
Back to blog
GEO, visibilité IA & tendances 16 min read

Is LinkedIn Really AI Search's Second Most-Cited Domain?

| By Patrick de Carvalho

Contents

A stat has been circulating on LinkedIn since March: the platform is AI search's second most-cited domain, right behind Reddit. I have seen it in a dozen posts, always with the same phrasing, rarely with a link to the study behind it. I have been on LinkedIn since April 2004, I write about the platform's mechanics for a living, and a claim that convenient gets checked before I repeat it, the same way I checked LinkedIn's own numbers in the piece on its "Seems like AI slop" button. I read the three primary studies it draws on. The exact ranking does not hold up as stated. What does hold up is more specific, and more useful to anyone deciding whether to keep writing LinkedIn articles in 2026.

In short: No study confirms LinkedIn as the flat "second most-cited domain" in AI search. What the primary research shows: LinkedIn is the number one most-cited domain for professional queries specifically, across six AI models (Profound, 1.4 million citations, November 2025 to February 2026). It is cited in 14.3% of ChatGPT Search responses, 13.5% of Google AI Mode responses, and 5.3% of Perplexity responses (Semrush, 325,000 prompts, March 2026). Articles, not feed posts, drive most of that visibility.

Where the "second most-cited" claim actually comes from

Two different studies get merged into one soundbite, and they answer two different questions.

The first is Semrush's "Most-Cited Domains in AI" report, published November 10, 2025, built on more than 230,000 prompts tracked over thirteen weeks, from July 14 to October 12, 2025, and over 100 million AI citations analyzed. It found that LinkedIn and Reddit sat among the top five most-cited domains across ChatGPT, Google AI Mode, and Perplexity, with LinkedIn appearing in close to 15% of Google AI Mode responses during that window. The report lists top domains in this order: Reddit, Wikipedia, LinkedIn, Forbes, Medium. That puts LinkedIn third in that particular list, not second, and Semrush itself does not publish one unified rank that combines all three platforms into a single number.

The second is Profound's professional-query study, published March 9, 2026, covering November 15, 2025 to February 15, 2026 and 1.4 million citations across six models: ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Microsoft Copilot, and Perplexity. This is where the strong claim actually sits, and it is a better one than "second overall": LinkedIn is the number one most-cited domain for professional queries specifically, across every model tested. On ChatGPT alone, LinkedIn's domain rank climbed from 11th to 5th in three months, more than doubling its citation frequency.

Neither study says "second, behind Reddit, full stop." One says "top five, varies by platform." The other says "first, for a defined category of query." Somewhere between a conference talk and a LinkedIn post, that became "second." It is the kind of rounding error that flatters the platform you are posting on, which is exactly why it deserved a second look before I put it in a headline.

The numbers that hold up, and the ones that don't

Strip out the rounding and three figures survive contact with the primary sources.

Metric Value Source
LinkedIn citation rate, ChatGPT Search 14.3% of responses Semrush, March 2026
LinkedIn citation rate, Google AI Mode 13.5% of responses Semrush, March 2026
LinkedIn citation rate, Perplexity 5.3% of responses Semrush, March 2026
Rank for professional queries, six models combined #1 Profound, March 2026
ChatGPT domain rank, three-month change #11 to #5 Profound, March 2026

Semrush's second report, "89K LinkedIn URLs Cited in AI Search," published March 10, 2026, is the one that matters most for content strategy. It examined 325,000 prompts across twelve industry categories and traced 89,000 LinkedIn URLs cited by ChatGPT Search, Google AI Mode, and Perplexity. It also measured something most citation studies skip: fidelity. LinkedIn content scored a semantic similarity of 0.57 to 0.60 against the AI response that cited it, against 0.53 to 0.54 for Reddit and 0.435 for Quora. Semantic similarity, in plain terms, measures how closely the AI's summary matches the original meaning of the source. A higher score means the platform's answer misquotes you less. On that measure, LinkedIn beats both of the communities usually credited with dominating AI search.

The number I will not repeat without a flag is the one about how often people publish. It shows up often enough in commentary that I looked for its source and could not find one I could stand behind with a link. It is dropped here rather than laundered through a secondhand citation.

Why articles carry the weight posts can't

The composition of what gets cited shifted hard between the two study windows. Profound's data shows LinkedIn feed posts moving from 20.9% to 26.0% of citations, long-form articles moving from 6.0% to 8.9%, and the two combined climbing from 26.9% to 34.9% of all LinkedIn citations in three months. Articles are a smaller slice of what LinkedIn hosts and a growing slice of what AI search repeats.

Semrush's URL-level study explains part of why. Articles between 500 and 2,000 words account for 50% to 66% of the LinkedIn content cited across the three platforms it tracked, and 95% of that cited content is original rather than reshared. A comment bolted onto someone else's post carries almost no citation weight of its own.

There is also a split by content type that changes who should care about this most. On Perplexity, company pages account for 59% of LinkedIn citations. On ChatGPT Search and Google AI Mode, it flips: individual members make up 59% of citations on both. If your company page gets more attention than your personal profile in your internal reporting, you are optimizing for the platform that cites you the least. Two of the three major AI search surfaces are pulling from people, not brand pages.

An article is a standalone web page with its own URL, its own title tag, and a structure a crawler can parse cleanly. A feed post is a fragment inside an infinite scroll, harder to isolate and cite as a discrete unit. That difference in shape, not some algorithmic favor, is most of why the numbers land where they do.

What makes a passage citable

Getting cited is not about writing longer. It is about writing sections that survive being lifted out of context, because that is precisely what a generative engine does to them.

An analysis of 15,847 AI Overview results across 63 industries, published by Wellows and sourced to a dataset called AI Mode Boost, found that content scoring above 8.5 out of 10 on a "semantic completeness" measure was 4.2 times more likely to be cited than content scoring below 6.0. Semantic completeness, as the researchers define it, tests whether a passage functions as a complete answer on its own, without requiring the paragraph before or after it to make sense. It was the strongest single predictor they measured, with a correlation of 0.87. The same analysis put the optimal length for an extractable passage at 134 to 167 words, tight enough to be lifted whole, long enough to answer something.

That is a narrower target than the 80-to-120-word paragraph guidance floating around most SEO advice, and it is worth building sections around directly: state the claim, back it with one number and its source, close the thought, stop. If a paragraph needs the one before it to make sense, an AI engine will either skip it or paraphrase it into something that no longer sounds like you.

The GEO article blueprint I use for LinkedIn articles

All In built a seven-block structure for LinkedIn articles, laid out in full in All In's methods, designed around exactly this extraction behavior. It is the same shape whether the subject is prospecting, personal branding, or a stat correction like this one.

The lead, twenty-five to eighty words. Open with a checkable number or a specific claim, not a rhetorical question. The first sentence should answer the query your headline implies.

The "In short" block, twenty-five to sixty words. A dense paragraph that AI engines can lift on its own, sitting right after the lead, before the first section header.

Five to seven body sections. Each one answers a distinct sub-question in 134 to 167 words, per the completeness data above, with at least one figure or named example per section.

One comparison table. Structured data extracts more cleanly than prose. A single well-labeled table beats three paragraphs describing the same comparison.

Definition sentences. Wherever a term carries weight, state it plainly: "X is Y." These sentences get lifted whole into AI answers built around "what is" queries.

Lists with a sentence attached to each item. A bare bullet gets skipped more often than a bullet followed by one explanatory sentence.

A sober close. One line connecting the piece to what you actually do, without breaking the register the rest of the article held.

Where LinkedIn still loses to Reddit

None of this makes LinkedIn the biggest source in AI search overall, and the honest version of this piece says so. Semrush's tracking shows Reddit's citation share on ChatGPT falling from roughly 60% to roughly 10% over its thirteen-week study window, a steep drop, and still a share that dwarfs LinkedIn's low-teens rate on the same platform. Reddit's raw volume of citations, across the open web and every query category, remains larger than LinkedIn's. LinkedIn's advantage is concentration, not size: it wins the professional-query category outright and loses badly outside it. Nobody is asking ChatGPT for movie recommendations and getting a LinkedIn article back. That is not a flaw in the platform. It is the shape of the opportunity, and it means the "second overall" framing overstates the case in the one direction that would matter if it were true.

Publishing reshares instead of original work. Reshared content accounts for roughly 5% of what gets cited. A comment on someone else's post is close to invisible to an AI engine, whatever the comment says.

Writing outside the cited range. Articles under 500 words rarely carry enough substance to answer a query on their own. Semrush's data puts the cited range at 500 to 2,000 words; there is no confirmed narrower sweet spot inside that band, whatever a template tells you.

Making claims without a number attached. "The market is growing" gets paraphrased into nothing. "Grew 41% year over year, per [named source]" gets quoted.

Titling articles for yourself, not for a query. "Some thoughts on visibility" matches nothing anyone types into an AI assistant. "How to check if ChatGPT is citing your LinkedIn articles" matches a real search.

Posting to your company page when you meant to post to yourself. On ChatGPT Search and Google AI Mode, individual profiles draw 59% of citations against company pages. On Perplexity, that flips to 59% for company pages. Know which surface you are actually optimizing before you decide who hits publish.

How to check whether you are actually being cited

There is no single dashboard for this yet, so it takes three habits, not one tool.

Ask the assistants directly. Put the exact questions your articles answer into ChatGPT, Perplexity, and Google AI Mode, once a month, and note whether your content shows up, gets paraphrased without attribution, or is absent entirely.

Watch referral traffic that does not match your usual pattern. A spike in article views with no corresponding spike in likes or comments often means an AI surface sent someone your way, not the LinkedIn feed.

Check semantic similarity when you can. Tools built for this, including the ones Semrush and Frase publish research through, compare your published text against an AI-generated answer on the same topic. A score above roughly 0.55 suggests your wording is shaping the response rather than getting summarized past recognition.

Check How How often
Direct query test Ask ChatGPT, Perplexity, Google AI Mode your own article's question Monthly
Referral traffic pattern Compare article views to engagement metrics Weekly
Semantic similarity Frase, Semrush, or comparable tooling Quarterly

What I don't know

I don't know the exact combined rank across all AI platforms, because no study I found publishes one. Semrush measures per-platform citation rates and a rough top-five list; Profound measures a professional-query subset; neither gives a single number that would let anyone say "LinkedIn is domain number X overall" with a straight face.

I don't know how durable LinkedIn's ChatGPT gain from #11 to #5 will prove to be. Three months of data is a trend, not a law, and Semrush's own tracking shows Reddit's share moving by tens of percentage points inside a single study window. Whatever ranks LinkedIn today can move by the same amount in the other direction. If a study publishes a genuine unified rank across platforms, it gets covered in the All In newsletter, with the same sourcing standard as this piece.

I don't know the real base rate for how often average LinkedIn members publish long-form articles versus feed posts. A widely repeated figure on this exists without a source I could verify, so it is left out of this piece rather than passed along.

What to do Monday morning

Nothing here requires a new tool. Write the article as a standalone answer, not a scroll-stopper: one clear claim in the first eighty words, a number with a named source every 150 words or so, sections tight enough to survive being lifted out on their own, one table, one honest caveat you did not have to include. Post it as yourself, not your company page, if ChatGPT and Google AI Mode are where you actually want to show up. Then check back on it in a month with the same three questions you built it to answer, typed straight into an AI assistant, and see who answers first.

FAQ

No study confirms that exact ranking. Semrush's November 2025 report places LinkedIn among the top five most-cited domains, listed after Reddit and Wikipedia, without a single combined rank across platforms. Profound's March 2026 report confirms LinkedIn as the number one most-cited domain specifically for professional queries, which is a narrower and better-supported claim.

What percentage of ChatGPT responses cite LinkedIn?

Semrush's March 2026 study, covering 325,000 prompts, found LinkedIn cited in 14.3% of ChatGPT Search responses. The same study measured 13.5% for Google AI Mode and 5.3% for Perplexity, meaning citation behavior differs meaningfully by platform.

Do AI search engines cite LinkedIn articles or LinkedIn posts more?

Long-form articles between 500 and 2,000 words account for 50% to 66% of the LinkedIn content cited across ChatGPT Search, Google AI Mode, and Perplexity, according to Semrush's analysis of 89,000 LinkedIn URLs. Feed posts and articles combined grew from 26.9% to 34.9% of all LinkedIn citations between late 2025 and early 2026, per Profound.

Is there an ideal length for a LinkedIn article to get cited by AI search?

Semrush's data confirms 500 to 2,000 words as the range that captures most citations, without a narrower confirmed sweet spot inside that band. Separately, a passage-level analysis of AI Overview results found individual sections of 134 to 167 words were most likely to be extracted whole, which matters more for how you structure sections than for total article length.

How can I check if my LinkedIn content is being cited by ChatGPT or Perplexity?

Ask the assistant the exact question your article answers and see if it appears in the response or the cited sources. Combine that with watching for referral traffic spikes that do not match your usual engagement pattern, and periodic semantic similarity checks using tools built for that comparison.

Does resharing content on LinkedIn help with AI search visibility?

No. Original content accounts for roughly 95% of what AI search engines cite from LinkedIn, according to Semrush's research. Reshares and comments on other people's posts carry close to no independent citation weight.

What is the real difference between LinkedIn and Reddit for AI search citations?

Reddit still carries a larger share of citations across the open web generally, even after a steep drop tracked by Semrush during 2025. LinkedIn's advantage is narrower and sharper: it is the top-cited domain specifically for professional and B2B queries, a category where Reddit does not compete on the same terms.

Sources

  1. Semrush, "The Most-Cited Domains in AI: A 3-Month Study," November 10, 2025: 230,000+ prompts, July 14 to October 12, 2025, over 100 million citations analyzed, top-five domain ranking including LinkedIn and Reddit.
  2. Semrush, "We Analyzed 89K LinkedIn URLs Cited in AI Search," March 10, 2026: 325,000 prompts, 89,000 LinkedIn URLs, citation rates by platform, semantic similarity scores, content-type and length breakdowns.
  3. Profound, "LinkedIn Is the Most-Cited Domain for Professional Queries in AI Search," March 9, 2026: 1.4 million citations across six AI models, November 15, 2025 to February 15, 2026, ChatGPT domain rank change and content-type composition shift.
  4. Gartner, "Gartner Predicts Search Engine Volume Will Drop 25% by 2026, Due to AI Chatbots and Other Virtual Agents," February 19, 2024: original prediction underpinning the shift toward AI-mediated search.
  5. Wellows, "Google AI Overviews Ranking Factors: 2026 Guide to Winning Citations," published December 8, 2025, updated July 29, 2026: semantic completeness scoring, 15,847 AI Overview results across 63 industries, optimal passage length.
  6. ALM Corp, "Google AI Overview Citations and Top-Ranking Pages," 2026: Ahrefs top-10 citation overlap data and BrightEdge overlap figures, February 12, 2026.

All In: being cited by ChatGPT starts with an article worth citing

Chasing a ranking is a distraction from the part that is actually in your control: sections built to answer one question completely, sourced with numbers a reader or a model can check. That is the same standard this piece holds itself to.

All In is the B2B media that decodes LinkedIn, expert blog, weekly podcast and newsletter for SME leaders and sales directors who want to turn LinkedIn into measurable growth. An original creation by Patrick de Carvalho, on LinkedIn since 2004. Motto: "I Never Lose."

Discover All In and get what matters on LinkedIn, every week.