Being Cited by ChatGPT Does Not Make You Visible on Claude

Eleven per cent

Two independent studies arrived at the same number from different directions. An analysis of 680 million AI citations found that only 11% of domains cited by ChatGPT were also cited by Perplexity. A separate study of 118,000 AI-generated answers across ChatGPT, Perplexity, Google AI Mode and Claude found the same figure.

Eleven per cent. These are not variations on one index. They are close to separate ecosystems that happen to answer similar questions.

The practical consequence is blunt: a strategy that gets you cited by ChatGPT has roughly a one-in-nine chance of also getting you cited by Perplexity. Optimising for one and assuming it carries everywhere is, in the words of one analysis, a costly mistake.

The traffic split is moving

If the platforms were static, you could simply pick the biggest. They are not.

ChatGPT still dominates overall AI referral traffic — figures range from around 75% to 87% depending on the measurement window and sector. But in B2B specifically the picture has shifted hard. One tracker found ChatGPT’s share of measurable B2B AI referrals fell from 89% to 62.6% over eight months, while Claude rose to 18.5% and Perplexity to 7.3%.

Read that again if you sell to businesses. Nearly a fifth of B2B AI referral traffic now comes from an assistant that most marketing plans do not mention. Together ChatGPT, Claude, Gemini and Perplexity account for around 99% of measurable AI referrals, and the share that left ChatGPT moved to surfaces with different retrieval logic and citation behaviour.

They cite in completely different ways

The platforms do not merely draw on different sources. They behave differently enough that “being visible” means different things on each.

Perplexity cites heavily. It averages around 21.9 citations per response, more than double ChatGPT’s 10.4, and tied every claim to a specific source in 78% of complex research questions against ChatGPT’s 62%. One study of 34,234 responses found Perplexity cited brands 13.05% of the time against ChatGPT’s 0.59% — a difference of more than forty times.

ChatGPT names rather than links. It mentions brands roughly 3.2 times more often than it cites them with a link. It leans on brand entity strength accumulated over years through third-party mentions, which is slow to build and hard to dislodge once established.

The traffic quality differs too. Visitors from Perplexity view around 13 pages on average against 11.8 from Google, and AI referral sessions generally engage noticeably longer than search traffic.

Why this is not a tidy conclusion

The honest reading is that Perplexity offers the clearer route for a smaller firm — it cites prolifically, favours freshness, and its sources include niche directories you can realistically get into. ChatGPT sends far more traffic but rewards entity strength built over years, which is exactly what a new business does not have.

AI citations barely overlap with Google rankings

Here is the finding that undoes the most common assumption of all.

Only around 12 to 14% of URLs cited by AI engines rank in Google’s top ten for the same query. Ranking first on Google is not a route to being cited, and being cited is not evidence that you rank.

That cuts both ways. A business invisible on Google can be cited by Perplexity on the strength of a well-structured, recent, clearly-sourced page. And a business ranking first for a commercial term can be entirely absent from the assistant answering that same question.

What this means for how you work

Three things follow, and none of them is “do more AI SEO”.

Pick your platform deliberately. Trying to optimise for all four at once with a small budget produces nothing anywhere. Decide based on where your buyers are: B2B technical audiences skew toward Claude and Perplexity far more than general consumers do.

Measure each separately. One AI visibility number is a fiction. Run your category prompts through each assistant individually and track four results.

Do not let one platform’s dominance set your strategy. ChatGPT’s B2B share dropped more than twenty-six points in eight months. Anything built around a single assistant inherits that volatility.

A workable approach

• Write the five questions a buyer asks before choosing someone in your category.
• Run all five through ChatGPT, Claude, Gemini and Perplexity. Record which brands appear where, in four separate lists.
• Note which platform names competitors you consider peers. That is where you are losing ground that is winnable.
• Check whether the sources each assistant cites are places you could realistically appear.
• Pick one platform to prioritise for two quarters, based on where your buyers actually are.
• Keep structured, well-sourced, genuinely dated content — it is the format Perplexity rewards most and it costs nothing extra.
• Build third-party mentions regardless of platform. Entity strength is the one asset that compounds across all four.
• Repeat quarterly. The platform split is still moving fast enough that annual reviews are too slow.

What this does not solve

None of this makes AI search a meaningful traffic channel yet — it is around 1% of sessions for most sites. It will not produce consistent results, because citations are volatile by nature. And it does not replace search: Google still drives vastly more referral traffic than all assistants combined.

What it does is stop you spending a year optimising for one assistant and discovering your buyers were asking a different one.

Related: why chatgpt does not mention your business.

Services referenced in this article: LLM Optimization · AI SEO · Content Marketing · Technical SEO.