New Study Reveals ChatGPT, Claude, and Perplexity Cite Sources Using Completely Different Patterns

Analysis of 83,670 AI citations finds 121x gap in Wikipedia usage between engines; LinkedIn only cited by ChatGPT

AUSTIN, TX, UNITED STATES, January 21, 2026 /EINPresswire.com/ — A new study by Analyze AI, an AI search analytics platform, has uncovered significant differences in how leading AI assistants cite sources when answering questions. The research analyzed 83,670 citations across ChatGPT, Claude, and Perplexity over 54 days and found that each engine relies on fundamentally different source types.

The most striking finding: Wikipedia citation rates differ by over 100x between engines. ChatGPT uses Wikipedia for 12.1% of its citations. Claude uses it for just 0.1%. Perplexity doesn’t cite Wikipedia at all. This means brands optimizing for Wikipedia visibility may only reach audiences using ChatGPT.

LinkedIn showed an even starker divide. ChatGPT cited LinkedIn 900 times (4.1% of citations). Claude and Perplexity cited it zero times. For B2B marketers investing in LinkedIn content, this suggests visibility is engine-specific.

Additional key findings include:
• Third-party sources dominate in B2B: Every time a brand is mentioned, 82.9% of AI citations come from external sources like review sites, news articles, and industry blogs. Only 17.1% come from said brand’s website.
• Content preferences vary widely: Claude favors blog content (43.8% of citations). ChatGPT and Perplexity prefer product pages (60.1% and 54.3%).
• Same brand, different perception: AI engines can rate identical brands up to 79 points apart on sentiment, depending on which sources each engine cites.
• Winner-take-all dynamics: The top 10 brands captured 30% of all AI mentions across 4,980 unique brands tracked.

“Marketers have been obsessing over Reddit lately, but our data shows LinkedIn is actually more interesting—ChatGPT cites it quite a lot lately,” said Ernest Bogore, CEO of Analyze AI. “A Wikipedia strategy that works for ChatGPT will completely miss Claude and Perplexity users. Brands need to track and optimize for each engine separately.”

The study tracked 74,130 brand mentions and 83,670 citations across 15,530 AI responses from November 22, 2025 to January 16, 2026. The full study with detailed breakdowns by engine and content type is available at: tryanalyze.ai/blog/ai-visibility-index-trend-update

Ernest Bio Bogore
Analyze AI
email us here
Visit us on social media:
LinkedIn
X

Legal Disclaimer:

EIN Presswire provides this news content “as is” without warranty of any kind. We do not accept any responsibility or liability
for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this
article. If you have any complaints or copyright issues related to this article, kindly contact the author above.