ChatGPT now cites LinkedIn posts almost 5x more than Pulse articles
David Konitzny's URL-type breakdown shows a near-complete inversion across 2026.
David Konitzny broke down every LinkedIn.com citation in ChatGPT answers since January 2026, sorted by URL type. He posted the analysis on LinkedIn on 9 September.
At the start of the year, Pulse articles carried about 78% of LinkedIn citations. By September that had fallen to roughly 12%. Posts moved the other way, from about 13% to about 57%.
Two smaller shifts sit inside that. Konitzny reports company pages peaked near 38% in June and settled around 14%. Personal profiles jumped from about 4% to about 7% around ChatGPT's 6–8 August model switch, per the same dataset.
Konitzny notes the June company-page spike came mainly because Pulse was collapsing faster than company pages were, not because company citations grew in absolute terms. Share can move without volume moving.
How we read it
We think this fits a wider pattern of AI Search leaning on fresher, conversational sources over static editorial. The data here is one practitioner's count. It is worth watching, not yet a settled trend.
Before you act on it
Run the same cut on your own citation logs. Filter ChatGPT referrals or citation exports by LinkedIn URL path and compare January against September. If your brand's LinkedIn presence is weighted toward Pulse articles from prior years, check whether those are still being pulled at all.
The answer is not to abandon one format for another. Posts, articles, company pages and profiles each do different jobs for a reader, and the retriever's current preference is one input. Write the thing that answers the question best in the place your audience already reads you.