30 million sources, five platforms, one question
Where does AI actually get its answers?
Not the training data (that ship sailed months or years ago). The live citations. When ChatGPT, Perplexity, Gemini, or Google's AI Mode responds to a query and links to sources, which websites show up?
Peec AI just published the most comprehensive analysis of this question to date. Their team analyzed 30 million citation sources across five major AI platforms: ChatGPT, Google AI Mode, Gemini, Perplexity, and AI Overviews. The results challenge some assumptions about what "authority" means in AI search.
The top 10 most-cited domains
Here's the overall ranking across all five platforms combined:
- •Reddit.com
- •YouTube.com
- •LinkedIn.com
- •Wikipedia.org
- •Forbes.com
- •G2.com
- •Yelp.com
- •Facebook.com
- •Medium.com
- •Techradar.com
If you've been building your GEO/AEO strategy around getting cited by authoritative editorial publications, this list should give you pause. User-generated content platforms dominate the top four spots. The first traditional media outlet (Forbes) doesn't appear until #5.
This isn't about domain authority in the traditional SEO sense. AI search engines are looking for something different: authentic, specific, experience-based content that answers real questions from real people.
Each AI platform has a different citation diet
The overall ranking hides significant platform-level variation. Here's how each platform's top sources differ:
Google AI Mode: YouTube, Reddit, Facebook, LinkedIn, Yelp
AI Overviews: YouTube, Reddit, Facebook, LinkedIn, Medium
Gemini: Reddit, YouTube, Wikipedia, Medium, PCMag/Forbes
Perplexity: Reddit, YouTube, LinkedIn, Wikipedia, G2
ChatGPT: Wikipedia, Reddit, Forbes, Techradar, LinkedIn
Two things jump out immediately:
Reddit and YouTube appear on every single platform. They're the only two domains that consistently rank in the top 5 across all five AI search engines. If you're ignoring these platforms, you're ignoring the two most reliable citation sources in AI search.
ChatGPT is the outlier. While every other platform leads with Reddit or YouTube, ChatGPT leads with Wikipedia and leans heavily toward editorial sources like Forbes and Techradar. If your strategy is ChatGPT-specific (and for many B2B brands, it is), the playbook looks different than optimizing for Perplexity or Google's AI products.
Why Reddit dominates AI citations
Reddit ranks #1 or #2 on every platform tested. That's not an accident.
AI search engines trust Reddit because it captures something most marketing content doesn't: genuine user experiences. Real people asking real questions and getting real, sometimes brutally honest answers. No press releases, no landing page copy, no carefully optimized blog posts. Just humans talking to each other.
For brands, this creates both an opportunity and a risk:
The opportunity: Active, helpful participation in relevant subreddits can directly influence what AI says about your brand. Not promotional posts (those get downvoted into oblivion) but genuine contributions that demonstrate expertise.
The risk: If people are complaining about your product on Reddit and you're not there to respond, AI search engines will surface those complaints as authoritative answers. You don't control the narrative by default. You earn it.
This aligns with what we've seen in our own AI citation analysis. The sources AI trusts most are the ones that feel least like marketing.
YouTube: The video citation layer
YouTube's #2 position surprised some people. Video content and LLM citations seem like an odd combination. But the data shows AI search engines pull from YouTube in two ways:
Video descriptions. The text you write below your video is crawlable, indexable, and cited by AI platforms. Detailed, keyword-rich descriptions act as standalone content for citation purposes.
Auto-generated transcriptions. AI platforms pull directly from YouTube's auto-generated transcripts. This means every word spoken in your video is potential citation material, whether you intended it or not.
The practical takeaway: if you're publishing video content, review your auto-generated transcription before the video goes live. Errors in the transcript become errors in AI responses that cite your content.
This also means brands with strong YouTube presences have a built-in advantage. All that existing video content is already feeding AI search results, even if you never optimized for it.
Wikipedia: Two levels of influence
Wikipedia's position in the data is interesting because it influences AI in two distinct ways:
Training data influence. Wikipedia was a core training source for most major LLMs. If your brand had a Wikipedia page before the model's knowledge cutoff date, the model didn't just learn about you from that page. It internalized your brand as a known entity. You're not just a source it might cite, you're something it already understands.
Live citation. AI search engines also actively pull from Wikipedia during live queries. This shows up most strongly on ChatGPT and Perplexity, less so on Google's platforms (which makes sense, given Google has its own Knowledge Graph).
The catch: getting a Wikipedia page isn't a quick win. The platform has strict notability guidelines. But for brands that qualify, the compound effect of training data influence plus live citation makes it one of the highest-value GEO/AEO targets available.
What this means for B2B brands
If you sell to businesses, pay attention to which platforms your buyers' AI tools prefer.
G2 only shows up in Perplexity's top sources. That's a narrow but potentially high-value channel. If your target audience uses Perplexity for research (increasingly common among tech buyers), your G2 profile matters more than you think.
LinkedIn ranks in the top 5 for every platform except Gemini. For B2B brands, this validates what many suspected: LinkedIn content isn't just for human networking anymore. Posts, articles, and company page updates from founders and executives are being cited by AI search engines.
Forbes and Techradar matter most for ChatGPT. If earned media and press coverage are part of your strategy, they'll have the most AI citation impact on ChatGPT specifically.
The diversification argument
One of the most practical insights from this data: don't put all your GEO/AEO effort into one platform.
Peec AI's research points out that both Google and ChatGPT have gone through periods of pulling back from Reddit. If Reddit disappeared from AI search results tomorrow, brands with no presence elsewhere would lose a significant portion of their AI visibility overnight.
Every additional platform where your brand has a credible, consistent presence is another opportunity to show up in AI responses. And when AI encounters your brand across multiple independent sources, it references you with more confidence.
This is the core principle behind what we call the CITE framework: building comprehensive AI presence isn't about winning on one platform. It's about being consistently present across the sources AI trusts.
What to do with this data
If you're building or refining a GEO/AEO strategy, here's a practical checklist based on these findings:
Audit your Reddit presence. Search for your brand name and product category on Reddit. What's being said? Are you participating in relevant subreddits? If not, start.
Review your YouTube transcriptions. If you have video content, check the auto-generated transcripts for accuracy. Bad transcripts feed bad AI responses.
Optimize your LinkedIn company page and executive profiles. AI search engines are citing LinkedIn content across four of five platforms tested. This isn't optional anymore.
Check your Wikipedia eligibility. If you're a notable brand without a Wikipedia page, that's a gap. If you have one, make sure it's accurate and current.
Match your strategy to your target AI platform. If your audience uses ChatGPT, prioritize editorial authority and Wikipedia. If they use Perplexity, invest in G2 reviews and Reddit presence. Don't use a one-size-fits-all approach.
Track what's actually getting cited. The data from Peec AI gives you the map. Monitoring tools tell you whether your specific content is showing up.
The brands that win in AI search won't be the ones with the best SEO. They'll be the ones with the most authentic, distributed presence across the platforms AI actually trusts. This data tells you exactly which platforms those are.
Data referenced in this article comes from Peec AI's analysis of 30 million AI search sources, published March 31, 2026.