Semrush analyzed 325,000 prompts across ChatGPT Search, Google AI Mode, and Perplexity.
They found LinkedIn appearing in 11% of AI responses on average. It ranked as the second most cited domain in the entire dataset.
For context, that puts it ahead of most publishers and reference platforms marketers assume dominate AI answers.
If you’re in B2B - SaaS, consulting, professional services - AI is already sourcing from LinkedIn to answer your buyers’ questions.
The question is whether your content is included.
Here’s what the data says to actually do differently.
Write Articles, Not Just Posts
LinkedIn articles made up 50–66% of cited LinkedIn content. Posts made up 15–28%. The most cited article length was 500–2,000 words.
Marketers usually treat LinkedIn as a feed with short posts only. Articles are considered pointless as engagement is often low.
But this data says AI isn’t prioritizing what performs in the feed. It’s pulling what gives it enough context to answer a user’s question.
A 200-word reaction post usually doesn’t do that. A 1,000-word article that adds depth, expertise, and insights does
Consider a repurposing workflow. Either repurpose posts into LinkedIn articles or website articles into LinkedIn articles.
Stop Resharing. Start Explaining.
95% of cited posts were original. Reshares accounted for roughly 5%.
LinkedIn is full of reaction content. Someone reposts a news story, adds two sentences, and publishes. Unfortunately, it doesn’t get cited in AI search.
Original posts - new ideas is the way to go.
Also, 54–64% of cited posts were educational or advice-led. That’s the format that works here.
Not an opinion on its own. Actual explanation. How something works, what to do, why it’s important. Self-contained posts that don’t rely on outside context to make sense.
That’s what AI can use. Everything else is pretty much ignored.
Consistency Matters More Than Going Viral
Three-quarters of cited post authors had published at least five times in the four weeks before being cited.
Median engagement on those same posts? 15–25 reactions. One comment or fewer.
That’s the most important finding in the report. The content that got cited didn’t look like top-performing content. It looked ordinary.
What those authors had in common was volume and topic focus. Not one standout post. Repeated, consistent coverage of a topical area.
For AI retrieval, covering a topic regularly appears to carry more weight than occasionally nailing a viral high-engagement post.
If AI search optimization is important, stop chasing viral success and build topical depth.
Audience Size Is Less of a Barrier Than Most Assume
Creators with under 500 followers were cited as often as, sometimes more than, those with larger audiences.
That’s a rare thing to happen in this day and age.
It’s an opportunity for consultants, niche operators, and smaller brands that avoid LinkedIn due to competition and underperformance.
Expertise, consistency, and tight topic focus appear to outperform follower count in this dataset. That’s actually a prime opportunity if you execute correctly.
Run Both the Brand Page and the People Behind It
Perplexity cited company pages in 59% of its LinkedIn citations. ChatGPT Search and Google AI Mode cited individual members in 59%.
Company pages have been written off by many marketers as organic posts only reach roughly 1–2% of followers.
But this data suggests company pages still carry citation weight in AI search even when on platform performance is weak.
Use the company page to publish the brand’s expertise on specific topics. Use founder and employee profiles to add named authority and individual perspective.
Both contribute to AI retrieval, just differently.
What to Actually Change
Stop measuring LinkedIn content only by engagement. A post with 20 reactions that covers your product category or services may be doing more for AI search visibility than a post with 4000 likes.
Pick a small number of topics. Cover them consistently. Publish both posts and articles. Make sure the LinkedIn version of your thinking is complete and original, not just a teaser pointing back to your website or a lazy repost of website articles.
One final thing worth knowing:
The Semrush data also found high semantic similarity between cited LinkedIn content and the AI answers that used it. That means your content can directly influence the answer the user is seeing.
Although untrackable from a performance perspective, it’s valuable in the user journey.
Read the full breakdown here.
If you find value in these emails, please share or subscribe if you haven’t.