Only 11% of AI Citations Overlap Between Platforms. There's No Such Thing as "AI Search Optimization."

Only 11% of AI Citations Overlap Between Platforms. There's No Such Thing as "AI Search Optimization." Speaker 1: A study analyzed 680 million AI citations and the findings that should change how you think about AISearch is this, only 11% of the sources ChatGPT cites also get cited by perplexity. 89% of what works on one platform doesn't work on another. Which means if you're optimizing for AISearch as one thing, one strategy, one playbook, you're probably invisible on 3 out of the 4 platforms. I'm about to show you exactly what each platform trusts, what they ignore, and how to build visibility across all of them. This is the data nobody else is breaking down. This comes from profound citation analysis. 680 million data points across ChatGPT, Perplexity, Google AI overviews, I've also pulled from Princeton's GEO research that tests 9 optimization techniques against 10,000 queries and a separate study that analyzed 1,200 pages across 90 days. Three major studies. I'll connect them all. Now I'm going to show you what each AI platform actually cites, the specific content formats that gets picked up versus ignored, the structural elements that multiply your citation rate and the practical playbook you can start this week. And I'll show you one content format that gets cited 67% of the time and another format that most brands spend all their time on that gets cited just 18%. The gap is brutal. Right now, the entire marketing world is talking about AI search optimization like it's one thing, one channel, one strategy. Optimize your content for AI, they say, and you'll show up. That's like saying optimize for social media without specifying whether you mean TikTok, LinkedIn or YouTube. The platforms are completely different. The audiences are different. The algorithms are different. And now we have the data to prove AI search works the same way. There's no such thing as AI Search Optimization. There's ChatGPT Optimization, Perplexity Optimization, Google AI Optimization and Claude Optimization. Four different games with four different rulebooks. Profound analyzed 680 million citations from August 2024 through June 2025 and found that only 11% of domains overlap between ChatGPT and Perplexity. That means 89% of the sources that one platform trusts are completely ignored by another, and the differences aren't subtle. ChatGPT's number one source is Wikipedia, 47.9% of top citations. Then Reddit at 11.3%, Forbes at 6.8%, and G2 at 6.7%. Perplexity, Reddit dominates at 46.7%. Then YouTube at 13.9%, Gartner at 7%, and Yelp at 4.8%. Wikipedia, which is almost half of ChatGPT's citations, barely shows up on Perplexity. Google AI Overviews calls from Reddit at 21%, YouTube at 18.8%, Quora at 14.3% and LinkedIn. A completely different mix again. And then there's how many sources each platform cites per answer. A separate study, query analysing over a 118,000 AI answers found Perplexity cited 21.87 sources per question. ChatGPT cites 7.92. That's a 2.76x difference. A brand appearing in 10% of Perplexity's responses might show up in only 3.6% of ChatGPT's responses. Not because you're Content is worse, but because the platforms have fundamentally different citation strategies. So if you're running one AI optimization strategy and checking one platform, you have no idea how you're performing on the other three. A Wikipedia presence helps on ChatGPT, but does nothing for Perplexity. Perplexity. A strong Reddit presence helps you on Perplexity and Google AI but barely moves ChatGPT. YouTube content matters for Google AI overviews but not for ChatGPT. One strategy optimizes for one platform and ignores three. Now before I go deeper into platform breakdown, you might be thinking, does AI search traffic even matter? It's still tiny compared to Google, right? Total AI reference traffic is still about 200 times smaller than Google Organic, that's true, but ChatGPT referral traffic converts at 11.4% versus Google's organic at 3.9-3.9%. That's from SimilarWeb Global Ecommerce Report. A separate study across 100 ecommerce stores showed that ChatGPT's converting at 6.7% still meaningfully higher than Google, and the trajectory is what matters. ChatGPT handles over 200 million queries daily. Perplexity exceeded 500 million monthly queries in late 2025. ChatGPT traffic to ecommerce sites has tripled in recent months for some brands. There's an academic study worth knowing about. Kayser and Skolls analyzed 12 months of first-party data from 900 173 ecommerce sites generating $20 billion in annual revenue. They found that conversion rates from ChatGPT referrals increased steadily month over month, even as traffic volume expanded. Revenue per session rose consistently throughout the study period. So this isn't a case where conversion drops as an audience gets broader. The more people use ChatGPT to shop, the better the traffic converts. And here's a number that reframes the urgency. Less than 10% of sources cited by ChatGPT rank in Google's top 10 for the same query. 90% of ChatGPT citations come from URLs at positions 21 or lower on Google. Your Google rankings don't predict your AI visibility. They're almost unrelated. So the channel converts better. It's growing fast, and your current Google position tells you almost nothing about whether you show up. That's why this matters, and why the platform-specific approach I'm about to walk through is worth your time. Now that you understand these platforms are different, let me show you what each one actually rewards, because the tactics that work are specific. I'm going to give you the four rulebooks. What ChatGPT trusts, what Perplexity trusts, what Google AI trusts, and what Claude trusts. These come from the citation studies, not theory and not guesses. So ChatGPT, 87% of its citations come from Bing's top results. Not Google's. If your site isn't indexed properly in Bing, you're invisible to ChatGPT. That's the single most important technical fix for this platform. Beyond that, ChatGPT loves structured content. Comparison tables deliver a 47% citation lift. Pros and cons lists delivers a 38%. Pricing tables delivers a 51%. Wikipedia presence matters because ChatGPT pulls it at 47.9% and you need to make sure your site allows ChatGPT's crawler. It's called OAISearchbot. If you've blocked it in your robot.txt, ChatGPT literally can't see you. Perplexity. This is the community first platform. Reddit accounts for 46.7% of its top citations. YouTube is second at 13.9%. Perplexity cited nearly 22 sources per answer, so there's massive opportunity to get in But it's the highest freshness bias of any platform. 76.4% of its most cited pages were updated within 30 days. Just displaying a prominent last updated date on your page delivers a 47% citation lift. Perplexity also rewards specific facts, exact numbers, measurements, statistics. Vague claims get skipped. Specific data points get cited. Google AI overviews. This is where schema markup is key. FAQ Schema delivers a 221% citation lift. How-to Schema delivers a 184% lift. Product Schema delivers a 118% lift. And if you don't have structured data on your site, Google AI barely knows you exist. Reddit matters here too, 21% of citations. YouTube is 18.8%. Google AI Overview cites 7 unique domains per query on average, which means the competition for those with slots Claude. Claude is debt first. It prefers comprehensive, authoritative content. 3,500 to 5,000 words is the sweet spot. It's more likely to cite a single authoritative source rather than aggregate in multiple. Expert author attribution delivers a 44% citation lift on Claude. Case study citations deliver a 39% and it has less recency bias than the others. Content updated within 6 months still performs well. That's four platforms with four different priorities. So, ChatGPT wants Bing indexing and structured comparison. Perplexity wants Reddit presence and fresh content. Google AI wants schema markup. Claude wants depth and authority. So, a single strategy gets you one of these right at best. A platform-specific approach gets you all four. Now, let me show you the content side because the format you publish in matters as much as where you publish it. A study called Presence AI analyzed over 1,200 pages across ChatGPT, Claude, Perplexity, and Google AI. Overviews over 90 days. Content formats get cited and which get ignored. Comprehensive guides with data get cited 67% of the time. Thought leadership and opinion pieces 18% of the time. A data-rich comparison guide outperformed an opinion piece by 3.4 times. If you're publishing opinion content and wondering why AI doesn't cite you, that's why. So here's the full ranking. Comprehensive guides with data, 67% citation rate. Comparison metrics and reviews, 61%. FAQ heavy content, 58%. Jumping to 71% when you add schema markup. Step-by-step guides, 54%. Industry benchmark reports, 52%. Case studies with data, 48%. Definitions and framework pages, 46%. Tool resource lists, 41%. Best practice checklists, 38%. Trend analysis, 35%. And at the bottom, thought leadership at The structural elements that multiply citation rates are just as specific. So proper H2-H3 heading hierarchy delivers 3.2 times higher citation rates 62% versus 19% for pages with no structure. Comparison tables deliver 2.8 times higher citations. FAQ sections with 10 plus questions and schema deliver a 156% lift. A key takeaway box at the top of the page delivers a 83 to 94% lift. Data visualization delivers an 89 to 100% lift. 103% more citations and freshness. Content updated in the last 30 days gets cited 64%. That's 128% higher than content over 12 months old. The last updated time IMstamp alone is worth a 47% lift. If you published a great guide 6 months ago and haven't touched it, the AI platforms are already moving on. So the formula is specific. You want a 3000-5000 word comparison guide with proper heading structure, comparison tables in the first 500 words, FAQ sections with schema, a key takeaway box at the top, specific data points throughout, and you need to update it monthly. The format gets cited At 67% most brand content on the web doesn't come close. One more piece that ties this together and it connects to why Perplexity and Google AI show such different citations than ChatGPT. Reddit. Reddit accounts for 46.7% of Perplexity's top citations and 21% of Google AI overviews. It appears in 68% of AI search responses across all platforms. Google pays Reddit $60 million a year for real-time data access. A Reddit thread gets indexed and trusted by Google's AI faster than your own blog post. And AI uses upvotes as proxy for verification. Comment with 300 upvotes is treated as a verified fact by a model. Now there's a specific strategy practitioners are using called Citation First. Instead of positioning new Reddit threads, you can find threads AI already cited and engage there. Here's how it works. Step 1. Use AI Visibility Tool to find which Reddit URLs are being cited for your target keywords. Step 2. Look for threads with high time-cited metrics. A thread cited 400 Plus times is a goldmine. Step three, add a high-value structured comment to that specific thread. Structured content on Reddit, like lists or tables, bolded key terms, is 40% more likely to be cited by AI than unstructured text. Now, the three rules that practitioners have tested. First, mirror the query. Start with a clear header that matches what the users ask. Not, our product is great. Something like, the best tools for inventory tracking in 2026 are, followed by a structured breakdown. Second, use lists and bullets. AI models pull from lists more than paragraphs. Third, stay neutral. AI filters out promotional language. Write like a helpful peer. In my experience, based on my tests, Standard industry practice is the moment you sound like a sales pitch that AI skips you. One practitioner on Reddit posted, I shared content with an expectation that AI would reference it for future responses. It appeared as a reply in under a day. That's how fast this channel moves. Now, if you're an e-commerce brand doing three plus million a year and you don't have a Reddit strategy, you're invisible to nearly half of what Perplexity cites and a fifth of what Google AI cites. That's not a nice to Now, here's what all of this comes down to. The Princeton research tested 9 techniques against 10,000 queries, adding citations to authoritative sources, 30-40% visibility lift, adding specific statistics, 30-40% lift, adding expert quotes, 30-40% lift, keyword stuffing, minus 9% that actively hurts you. The playbook that built your Google ranking will damage your AI visibility. 4 platforms, 4 rulebooks. ChatGPT trusts Bing and Wikipedia. Perplexity trusts Reddit and Fresh Content. Google AI trusts schema markups. Claude trusts depth and authority. The brands that build for all 4 become the default answer everywhere, and not just on one AI. And the brands that treat AISearch ...as one strategy will keep wondering why their visibility isn't moving. There's no such thing as AI search optimization. There's only platform-specific AI visibility and now you know the difference. So everything I just walked through, the AI platform differences, the citation data, the content formats, That's the strategy, but you're probably wondering what the results actually look like when you put it into practice. We ran this. We took a beauty brand, built an off-site authority strategy, arguing the exact signals I just described, high authority placements across domains that AI platforms trust. One press release, 581 media outlets picked it up. Yahoo Finance. Business Insider, AP News and in two weeks, ChatGPT brand impressions jumped 13.5%, Gemini went up 12.7% and we haven't even started on schema markups or site optimisation. Just Layer 1 of the strategy. I made a full case study video showing every placement, every AI platform lift and the 3 layer framework for building this for your brand. That video is up here somewhere or you'll find it in the description. If you've just spent 12 minutes understanding the theory, that video is the proof.