First-party research · September 2026
AI engines do not share a citation layer
Most advice about AI visibility treats ChatGPT, Perplexity, Claude and Gemini as one channel. Measured directly, they draw on substantially different sources — and when the same question is asked in another language, ChatGPT returns almost entirely different ones again.
What was measured
Matched buying-intent queries were run against all four engines with web retrieval enabled, and the hostname of every cited source was recorded. Two independent B2B verticals were tested: embedded insurance APIs, an emerging category with a thin content ecosystem, and payments infrastructure, a mature one with years of developer content and comparison sites behind it. Each vertical used 11 queries, each asked three times of each engine.
Finding 1: ChatGPT sits apart from the other three
For each pair of engines, the share of cited domains they had in common, expressed as a percentage of the smaller set:
| Engine pair | Embedded insurance | Payments |
|---|---|---|
| ChatGPT / Claude | 12% | 15% |
| ChatGPT / Gemini | 17% | 12% |
| ChatGPT / Perplexity | 29% | 22% |
| Perplexity / Claude | 44% | 49% |
| Perplexity / Gemini | 42% | 33% |
| Claude / Gemini | 48% | 40% |
The pattern holds in both markets. ChatGPT shares 12–29% of its sources with the others; the remaining three share 33–49% with each other. That it replicates in a thin emerging category and a mature content-dense one suggests it is a property of the engines rather than an artifact of how well a niche is covered.
Finding 2: language separates ChatGPT’s sources almost completely
The same query was asked in English and in a target language, with the request geo-targeted to a market where that language is spoken. Each figure is the mean of six comparisons — two queries, three runs each.
| Language | ChatGPT | Perplexity | Claude | Gemini |
|---|---|---|---|---|
| Turkish | 0% | 15% | 40% | 45% |
| Arabic | 4% | 9% | 36% | 40% |
| Spanish | 6% | 16% | 28% | 25% |
| Japanese | 12% | 4% | 29% | 35% |
| German | 32% | 48% | 52% | 71% |
Turkish returned zero shared domains on every one of six samples. German is the clear exception, which fits a market whose business content sits close to the English-language web.
Per-cell ranges are wide — Claude’s Arabic figure averages 36% across a 6–73% spread — so these are bands rather than point estimates.
Finding 3: the consequence for a single company
Cover Genius, an embedded insurance provider, was measured across 11 buying-intent queries in its category. In one full scan its citation rate varied fourfold depending on which engine was asked.
| Engine | Score | Cited | Mentioned |
|---|---|---|---|
| ChatGPT | 15 | 12% | 21% |
| Perplexity | 29 | 18% | 55% |
| Gemini | 54 | 45% | 73% |
| Claude | 55 | 48% | 71% |
Across three full scans the gap between its best and worst engine ranged from 2.3 to 4 times, and the leading engine changed: Claude in two scans, Gemini in the third.
A second company in a different vertical showed a different failure. In a complete payments scan, Checkout.com was named by all four engines but had its own domain cited by only two — Perplexity, in 18% of answers, and Gemini, in 9% — and never by ChatGPT or Claude: recognition without retrieval. An earlier, partial scan showed it named considerably more often; that scan included a question naming Checkout.com directly, and the drop is consistent with removing it. On five of the eleven payments questions Claude did not search at all, which is covered in Claude answers payments questions from memory.
Finding 4: in payments, brands are named far more than their sites are cited
In the complete payments scan, the most-cited domains were vendors: connectpay.com appeared among the sources of 42 of 115 answers, stripe.com 39, tipalti.com and airwallex.com 36 each, whop.com 33. Independent publishers appeared lower, such as techrepublic.com in 16 and nerdwallet.com in 14.
The gap between being named and being cited is large. Stripe was named in 88 of those 115 answers and its domain cited in 39. An earlier version of this page, based on the partial scan, reported that comparison and review sites led payments citations and drew a contrast with embedded insurance on that basis. The complete scan does not support it, and that contrast has been withdrawn.
What this does not show
Two verticals and five languages, both chosen rather than sampled. One complete scan per vertical for the engine-overlap figures. Across three full scans of the same company, the per-scan score ranged from 32 to 38, while individual engine figures moved more: Cover Genius’s citation rate on Claude was 48%, 41% and 23% in successive scans. Treat per-engine numbers as bands. Nothing here establishes a link between citation and revenue.
Method
ChatGPT was queried through the Responses API with web search forced on; without forcing it the model answers from memory and cites nothing, which reads as absence rather than as a failed search. Perplexity and Gemini retrieve on every call. Claude decides for itself whether to search. Model versions were pinned. Answers returning zero sources, including Claude answers given without searching, were recorded as failed calls and never as a zero. Every recorded answer is retained with its query, engine, run index, cited domains and full text.
Measured for your company
The same measurement runs weekly against any company in a covered category — your score per engine, the competitors named instead of you, and the consensus sources that decide the category. Pricing and access.
The instrument behind these figures is published in full at methodology. The same measurement applied to whole categories: which embedded insurance companies AI engines name and which payments companies AI engines name.
Full scoring method, weights and limitations: methodology. Questions or replication requests: [email protected]