LinkedIn’s AI citations: 75% come from personal profiles

If you want AI tools to cite your LinkedIn content, post as a person, not a company.

New research from Meltwater and LinkedIn analysed 9.5 million AI citations across the major AI models, including Perplexity, ChatGPT, Copilot and Claude. Of the citations that came from LinkedIn, 75% came from individual member profiles. Only 25% came from company pages.

The takeaway is that AI prefers human context. It favours real expertise, direct experience and credible answers over polished corporate messaging. We suspect originality plays a part too. A company page post often repeats a press release or blog, while a personal post is more likely to say something new.

The study also looked at the 24 most-clicked LinkedIn URLs and found five types of content AI loves: “best X” listicles, how-to-choose guides, side-by-side comparisons, educational explainers, and thought leadership backed by data. Structure matters as well. Bullet points, numbered lists and clear H2 and H3 headings make content easier for AI to pull from.

So what should you do? Stop asking employees to reshare generic company announcements. Instead, help your experts and leaders write structured, original pieces in their own voice, with marketing acting as editorial support. Articles and newsletters are especially strong.

We’re acting on this ourselves. After three months of debate, this report convinced us to move our monthly social media update newsletter from the Avocado Social page to Alison’s personal LinkedIn profile. There’s no way to transfer subscribers, so we’re starting from scratch. We’ll report back on how it goes.