By Dr. Tamara Patzer
Key Takeaways
- Meltwater’s May 12, 2026, analysis of 9.5 million AI citations ranked LinkedIn second among cited sources, behind YouTube.
- About 75 percent of LinkedIn citations pointed to individual members and 25 percent to company pages. Members with fewer than 10,000 followers accounted for 51 percent.
- The most-cited articles in the sample used lists, clear headings, named companies and hard numbers.
- Google’s guide for site owners, last updated July 10, 2026, says special AI files are not needed and advises writing from firsthand knowledge.
- Meltwater’s later data shows the mix of cited sources shifting from month to month.
Picture an operations director at a regional freight company on a Tuesday afternoon. She asks an AI assistant how to choose dispatch software. The answer arrives in seconds with a short list of criteria and a few links. One goes to a vendor’s website. Another goes to a post by a logistics consultant with a few thousand followers, who walked through five questions to ask before signing a contract.
That second link is the kind of source a new analysis says AI tools reach for more often than many business leaders expect.
Who is being cited, and how often?
Meltwater, a media-monitoring and analytics company, published the research May 12, 2026. Chief Product Officer Chris Hackney wrote up the findings, which cover 9.5 million citations from answers by ChatGPT, Google AI Mode, Google AI Overviews, Gemini, Copilot and Claude. The testing ran four weeks across business-to-business categories.
YouTube led with a 1.52 percent citation share, Meltwater reported. LinkedIn followed at 0.53 percent, ahead of Reddit at 0.44 percent, Capterra at 0.38 percent and Medium at 0.21 percent. The shares look small because the study tracked millions of prompts and AI tools cite a very wide range of sites.
The authorship split stood out. About 75 percent of LinkedIn citations pointed to content from individual members, and 25 percent to company pages. Members with fewer than 10,000 followers accounted for 51 percent of the citations. In a statement released with the report, Hackney said AI models “aren’t looking for the loudest voice in the room.” They are looking, he said, for the most credible one.
What did the cited posts have in common?

Meltwater examined the 24 most-cited LinkedIn articles. All 24 used bullets or numbered lists. Clear headings appeared in 92 percent, named companies or tools in 75 percent, and hard numbers in 67 percent.
By format, “best of” lists appeared in 54 percent of the top-cited content, side-by-side comparisons in 50 percent and “how to choose” guides in 33 percent. Text posts made up 72 percent of cited LinkedIn content. Nearly half of all citations, 48 percent, came from posts less than three months old, and 72 percent of the cited content was original rather than reshared.
How firm are the numbers?
Meltwater sells the tracking tool used in the study, called GenAI Lens, and LinkedIn has promoted the findings in advertising aimed at marketers. Meltwater’s published summary does not list how many prompts it ran in each category or how it calculated citation share. The research covers business-to-business questions over four weeks. Readers can weigh the findings with that in mind.
Meltwater’s own later data shows how quickly the picture moves. Its report for October, published Nov. 13, analyzed about 88.6 million citations across eight AI platforms. LinkedIn’s total fell 7.4 percent between the first and second halves of the month. Its citations in Perplexity dropped about 99 percent, while they held steady in Google AI Mode and grew in Copilot. Meltwater noted that the October sample was far larger than in earlier months, so comparisons with past reports are not valid.
What does Google say?

Google’s guide for website owners, last updated July 10, 2026, says its generative AI features rest on the same core ranking and quality systems as regular Search. The guide advises creating original content based on what the author knows firsthand. It says llms.txt files and other special AI markup will neither help nor harm a site’s visibility, and that content does not need to be broken into tiny pieces.
What can a company leader do with this?
Meltwater’s recommendations are practical. It suggests pairing each question buyers ask with an expert who has hands-on experience, then building posts around direct answers, named tools, criteria, examples and data. It also recommends two to three posts a week per expert.
The study does not prove that any single post earns a citation. It does show what cited posts tend to look like: specific, structured, current and written by someone who plainly knows the subject. For a company with a deep bench of practitioners, that points to the people closest to the work rather than the brand account alone.
Back at the freight company, the consultant’s post worked as a source because it answered the question in the order the buyer asked it, with names and numbers attached. Any company’s own experts can do the same.
Dr. Tamara Patzer is a former newspaper editor who writes Authority Answer, a newsletter that tracks changes in AI search.
Sources: Meltwater, “9.5M AI citations analyzed: How LinkedIn content wins AI search,” May 12, 2026. Google Search Central, “Optimizing your website for generative AI features on Google Search,” updated July 10, 2026.






