top of page
Proactive International PR

What are AI platforms reading in 2026?

Writer: Proactive PR
Proactive PR
Nov 9
4 min read

As organisations race to improve their visibility in Large Language Models (LLMs) such as ChatGPT, Gemini, Claude and Perplexity, an important question is emerging – what content types do these AI platforms actually trust?


Thanks to a recent report from Muck Rack (published October 2026) and our own insights into the world of Generative Engine Optimisation (GEO), we’re closer to an answer, and for communications teams, the findings should be reassuring.


Is AI disrupting traditional media relations?

Yes – but credibility, authority and third-party validation remains central to how AI platforms are generating their recommendations and answers.


The report's findings have challenged several assumptions about AI visibility and reveal where organisations should focus their efforts if they want to be surfaced, cited and recommended more often. For example, one of the most striking findings is that earned media continues to dominate AI citations:


  • 84% of citations across AI-generated responses originate from earned sources, while around 99% come from non-advertorial content.

  • Journalism alone still accounts for roughly 22% of citations.


What’s clear is that AI systems remain heavily reliant on independent validation, and this means key industry publications, analysts and research institutions still have a role to play in shaping how companies appear in AI-generated answers. However, if your organisation is absent in their content, then platforms will struggle to view you as an authoritative source.


Do all AI platforms use the same sources?

They do not. According to Muck Rack – who provided more than 1,000 prompts to ChatGPT, Gemini, Claude and Perplexity – only 212 domains appeared across all four platforms.


Furthermore, each LLM demonstrated favouritism to different sourcing preferences. For example, ChatGPT surfaced more government and NGO sources than the other platforms, whereas Gemini favoured owned media from relevant companies on the searched topic. On the other hand, both Claude and Perplexity were more reliant on third-party corporate content.


The organisations most likely to succeed are those that build a balanced mix of:


  • Owned content and credible product information

  • Earned media coverage and third-party validations

  • Industry commentary, thought leadership and expert analysis

  • Research-driven content and practical guidance

  • Newsroom and corporate content that keeps AI platforms supplied with current, accurate and accessible information about your organisation.


In other words, AI visibility is becoming an ecosystem challenge rather than a channel challenge.


Does it matter which publications my company features in?

Absolutely, but perhaps not in the way most businesses expect. Many organisations assume AI platforms primarily rely on major national publications when generating answers. The reality is much more nuanced.


Presently, the median monthly audience of the outlets cited by ChatGPT is around 6.5 million visitors, while for Claude, Gemini and Perplexity the median cited outlet reaches fewer than one million monthly visitors. This suggests that AI platforms are often drawing information from specialist trade titles, niche industry publications and regional media outlets, not just household-name news brands. The report specifically highlights smaller trade and regional publications as a valuable opportunity that many PR professionals overlook.


For communications teams, this represents an important shift in thinking. While national coverage remains valuable, trusted trade and specialist publications can have an equal, or even greater, influence on AI-generated recommendations and citations.


For B2B technology companies, that's music to the ears, as niche industry titles are often the most rewarding outlets for reaching both customers and AI platforms alike.


Do AI platforms care which content type my business uses?

Another misconception surrounding AI visibility is that there is a universal content format that performs best. Muck Rack’s research actually found the opposite to be true.


The type of question being asked has a significant impact on what gets cited:


  • How-to queries overwhelmingly favour practical guides, accounting for 68% of citations.

  • Best-of prompts most commonly cite buyer's guides and review content.

  • Questions relating to ‘industry trends’ favour expert analysis, research and thought leadership

  • Comparison queries rely heavily on long-standing reference content.


Organisations need to stop asking ‘what content should we publish?’ and focus on the questions their target audiences are asking – and the answer should determine the format used. For example, a technology company trying to shape industry perception may benefit most from expert commentary and research-backed thought leadership. At the same time, an organisation seeking visibility around practical implementation challenges may achieve better results with educational guides and instructional content.


Do I really need a GEO strategy?

Yes, because AI search is evolving too quickly to leave visibility to chance.


One of the clearest findings from Muck Rack’s report is just how rapidly AI platforms can change their sourcing behaviour. The report found that when the same question was asked twice on the same day, source overlap ranged from just 41% for ChatGPT to 94% for Perplexity, meaning the sources behind an answer can vary significantly even when the prompt remains identical.


It also showed that throughout 2026, ChatGPT substantially reduced its reliance on journalistic sources while increasing citations from government, NGO and owned-media content. For businesses, this means AI visibility is no longer something that can be measured once and forgotten.


A GEO strategy helps organisations monitor how AI platforms represent their brand, identify shifts in citation behaviour and adapt content and PR activity accordingly. This is exactly why we developed OSCAR: to provide ongoing visibility into how AI platforms view your business, benchmark performance against competitors, work out which outlets you should prioritise and identify the actions most likely to improve your chances of being surfaced, cited and recommended. For more information on this managed service, visit https://oscar.proactivepr.com/.

 
 
bottom of page