Hey there 👋

Last week we split citations from recommendations, and one point that kept coming up is that the sources behind an AI answer often aren't your own website. For professional and B2B questions, LinkedIn is one of the biggest of those sources.

It's also growing fast. Otterly tracked 1.3 million AI citations of LinkedIn content in the first half of this year, and LinkedIn's monthly citation count rose 49.9% between January and May.

This week, we look at what the newest data says gets cited, where the studies disagree, and a publishing system you can use.

Let's go! 🚀

Google AI Overviews jump from 26% to 80% of branded queries. Google AI Overviews appeared on more than 80% of tracked branded queries by late September, according to new DemandSphere data. That was up from about 26% at the start of the month. The increase accelerated in the final days of September. AI Overview presence peaked at 90.48% on Sept. 27. [Source: Search Engine Land] 

We Analyzed 89K LinkedIn URLs Cited in AI Search. AI search engines like ChatGPT are actively shaping how your brand gets discovered, described, and evaluated, and LinkedIn is one of their primary sources. [Source: Semrush] 

LinkedIn is the most-cited domain for professional queries in AI search. Between November 2025 and February 2026, the platform surged from outside the top 20 to rank among the most‑cited sources on ChatGPT and is now the #1 most‑cited domain for professional queries across all AI Search platforms. That’s powerful data for both companies and individuals looking to grow their influence. [Source: Profound]

Profound saw LinkedIn climb from around #11 to around #5 among the domains ChatGPT cited in real user conversations between November 2025 and February 2026. When it ran a designed basket of professional prompts across six AI platforms, LinkedIn came out as the #1 cited domain, based on 1.4 million citations.

The growth carried into this year. Otterly found LinkedIn's monthly AI citations rose 49.9% between January and May 2026, and its share of social media citations went from 7.8% to 11.7%. Scrunch puts the volume at roughly 8 million LinkedIn citations a week in the US for industry and commercial prompts.

There are two limits worth knowing.

This is a professional-topics story. Across all queries in Google AI Overviews, Ahrefs ranks LinkedIn #19, with 0.6% of citations among the top 50 domains. For B2B, it's much stronger: Meltwater found LinkedIn in the top five cited sources for 14 B2B categories, #1 for marketing and advertising and #2 for financial services and fintech.

It isn't every engine, either. In Otterly's dataset, Gemini cited LinkedIn 38 times in five months, against 245,352 times for ChatGPT.

AI engines have moved from citing who you are to citing what you write.

Profound saw the shift first. Between November 2025 and February 2026, profiles fell from 33.9% to 14.5% of ChatGPT's LinkedIn citations, while posts and long-form articles grew from 26.9% to 34.9%.

Otterly's larger dataset, covering January to June 2026, shows how far that went. Once job listings, company pages, and other non-content pages were stripped out, here's how each engine's LinkedIn citations split:

Source: Otterly, January to June 2026

Articles also earned more per piece, at 8.5 citations per article URL against 5.9 per post. A polished profile with an empty activity feed is now close to invisible to AI.

Profound and Otterly counted differently, so they aren't one time series, but Semrush (articles at 50% to 66% of cited content) and LinkedIn's own internal data (roughly 60%) point the same way.

The exception is Meltwater, whose 9.5 million citation study found text posts made up 72% of cited LinkedIn content and articles 12%.

My read is that articles hold most of the citations in most datasets, and posts keep you in the stream between them. So make sure you do both!

Named individuals earned 91.7% of LinkedIn citations in Otterly's data, and Meltwater put the split at 75% members to 25% Company Pages.

Semrush's per-engine breakdown adds a wrinkle. Individual members account for 59% of LinkedIn citations on ChatGPT Search and Google AI Mode, while Company Pages account for 59% on Perplexity.

Semrush measured how closely AI answers matched the meaning of their LinkedIn sources and got semantic similarity scores of 0.57 to 0.60 on a 0-to-1 scale. In the authors' earlier studies, Reddit sources scored 0.53 to 0.54 and Quora 0.435.

It's still the only study measuring this, and the data is from January and February. But it suggests LinkedIn may also shape how your category and point of view get explained to buyers.

Engagement barely predicts whether AI cites you, and the newer studies are clearer on it than the early ones.

Otterly found a -0.06 correlation between likes and citations, and -0.04 for comments. Scrunch analyzed 12,000 ChatGPT observations of LinkedIn posts and found reactions had near-zero predictive power once content was controlled for.

Semrush found the median cited post had 15 to 25 reactions and no more than one comment, and members with fewer than 500 followers were just as likely to be cited as bigger accounts. Meltwater found 51% of cited creators had fewer than 10,000 followers.

What did stand out was frequency. In Semrush's data, around three-quarters of cited post authors had published more than five posts in the previous four weeks, and roughly 60% of cited article authors were frequent posters too.

AI visibility looks more like a stream than a stockpile. Profound's 11-day half-life applies across the web, and the pages that lasted were the ones cited by several engines. That's a gap for LinkedIn content, because Otterly found 87.4% of cited LinkedIn URLs showed up on only one platform.

Freshness matters on the way in, too. Meltwater found content from the previous three months made up 48% of LinkedIn citations, and content older than a year just 12%.

None of this proves that posting more gets you cited. Taken together, though, it makes a strong case for treating LinkedIn as an ongoing publishing channel rather than a one-off campaign.

LinkedIn also offers something a new or small website often can't, and that’s an established domain that AI engines already cite heavily, where your ideas get attributed back to you.

However, it doesn't replace publishing on your own site, which still matters for AI and for search.

Five moves, in the order I'd set them up. The first three decide what you publish, and the last two decide who publishes it and how you'll know it's working.

Build posts and articles around real prompts, then lead with the answer.

List 25 questions your buyers type into ChatGPT, pulled from sales calls, support tickets, and a tracker like LLMRefs. Turn each into a post or article that states the answer in the first line or two, adds a couple of supporting points, and closes with an example or number from your own experience.

LinkedIn's own AI search guide suggests opening each major section of an article with a 30- to 80-word answer. Same idea, article scale.

Write articles for depth and posts for rhythm.

Otterly found the median cited article ran 1,021 words and the median cited post 185 words. LinkedIn recommends 800 to 1,200 words for articles and 200 to 300 for posts. Length isn't the lever, though. Otterly measured a 0.03 correlation between word count and citations, so pick the length that fits the answer.

Structure matters more. Meltwater looked at the 24 most-cited LinkedIn articles: all of them used bulleted or numbered lists, 92% had clear H2 or H3 headings, 75% named specific companies or tools, and 67% included hard numbers.

Keep it plain text. Scrunch found unicode bold and italic cut ChatGPT citations by 58%, although Perplexity showed no significant penalty.

Original, specific, educational content wins.

In Semrush's data, knowledge and advice made up well over half of cited content, and almost two-thirds on Google AI Mode. About 95% of cited posts were original, and reshares made up about 5%.

Scrunch put numbers on what "specific" means. Posts with technical details were 77% more likely to be cited, with no measurable effect on reactions. Naming real companies, people, or products lifted citations 33%, and topic specificity lifted them 18%. Those are ChatGPT-only results from an observational model, but they line up with everything else here.

Your strongest material is proprietary: client data, first-hand lessons, specific results.

Your experts carry most of the citations, but the Company Page still earns its keep on Perplexity.

Give a handful of experts a defined topic each, support them with drafts or templates, and have them post their own angle on research your Company Page publishes. Keep category and product terms identical across every author, so AI has one clear story to repeat.

Track whether you show up in answers, and not just in the feed.

Run the same target prompts through ChatGPT, AI Mode, and Perplexity on a fixed schedule, or use a tool like LLMRefs, to see which LinkedIn URLs appear, disappear, or gain ground.

Give it at least 30 days before you judge anything. In Profound data cited in LinkedIn's guide, the median page took 6.81 days to earn its first citation, but one in 10 took longer than 37 days. Then double down on what keeps showing up, and expect even your winners to fade within weeks.

The Scenario: You know you should be posting on LinkedIn for AI visibility, but you're staring at a blank page. Use this prompt to turn your expertise into a month of citation-ready posts and articles built around the questions your buyers are actually asking AI.

The Prompt:

Pro tip: Before you publish, paste each target question into ChatGPT and Perplexity and note who gets cited today. That's your baseline, and your competition.

These are the most popular AI tools we use at Rise Up Media. If you're not using them already, they're worth a look.

Full disclosure: some links above are affiliate links. If you sign up, we'll earn a small commission at no extra cost to you.

Profound's newest data shows citations behave less like search rankings and more like a news cycle, with the median page losing half its share within 11 days of peaking. That makes a channel you can publish to every week more useful than it used to be.

AI engines already pull heavily from LinkedIn for professional topics, and most of those citations go to articles from named people rather than viral posts. The problem is that most companies still use LinkedIn for announcements, company updates, and the occasional CEO reshare.

None of this means posting more will automatically get you cited. But the pattern holds across every study: the LinkedIn content that gets picked up is original, specific, written by named people, and published often enough to stay in the stream.

Until next time, keep exploring the horizon. 🌅

Alex Lielacher

P.S. If you want your brand to show up in Google AI Mode, ChatGPT, and Perplexity, reach out to my agency, Rise Up Media. That's what we do!