Quick answer: A newly reported analysis suggests Anthropic’s Claude leans on Brave Search to decide which web pages to cite — reportedly using Brave’s top results without re-ranking them. If that holds, the most direct way to get cited by Claude is to rank well in Brave Search. The analysis also indicates Claude searches the web on only a minority of prompts and overlaps far more with Google’s results than with ChatGPT’s, which has practical consequences for how you optimize for AI visibility in 2026.
AI assistants are quickly becoming a discovery channel, and each one sources the web differently. A recent analysis covered by Search Engine Land — based on research attributed to Jonathan Clark of Moving Traffic Media — puts a spotlight on how Anthropic’s Claude chooses its sources. Here’s our read on the findings and what they mean for Generative Engine Optimization (GEO).
What the analysis reportedly found
According to the reporting, Claude appears to rely on Brave Search as its primary web index and tends to use Brave’s top-ranked results directly rather than re-ordering them. In other words, where your page sits in Brave’s rankings may translate fairly directly into whether Claude surfaces it. Several other data points stood out:
- Claude searches selectively. The analysis suggests Claude pulls in live web results on roughly a third of prompts (reported around 36.6%), far less often than ChatGPT.
- Search-triggering prompts. Queries with freshness or comparison intent — think “best,” “top,” “near me,” and “X vs. Y” — were the most likely to trigger a web search.
- Low overlap with ChatGPT, high overlap with Google. Claude’s cited sources reportedly overlapped with ChatGPT’s only about 8% of the time, but lined up with Google’s rankings far more often (reported around 64%).
- Consistent behavior. The way Claude expands a query into searches looked highly repeatable across users, making its behavior unusually observable.
These figures come from the cited analysis; we’re reporting and interpreting them, not presenting them as our own measurements. As always with third-party studies, treat the numbers as directional.
Why this matters for GEO
Most GEO advice treats “AI search” as one monolith. This analysis is a useful reminder that each engine has its own plumbing. If Claude really does inherit Brave’s ranking order, then Brave Search becomes a concrete, trackable lever for Claude visibility — something you can monitor, unlike the opaque internals of most chat assistants. And because Claude’s results reportedly track Google closely, the SEO work you’re already doing for Google likely carries over to Claude better than any ChatGPT-specific tactics would.
How to act on it
- Don’t ignore Brave Search. Make sure your site is crawlable by Brave’s bot and work on Brave rankings the way you would Google’s — quality, relevance, and authority still apply.
- Lean into freshness and comparison formats. Current-year signals in titles and “best/top/vs” framing match the prompt types that trigger Claude’s searches.
- Keep investing in Google SEO. Given the high overlap, strong Google rankings are a two-for-one: visibility in Google and a better shot at Claude citations.
- Write answer-first, citable passages. Whatever the engine, a clear, self-contained answer is easier to quote.
Frequently asked questions
Does Claude really use Brave Search?
The reported analysis indicates Claude relies on Brave’s index for live web results. Anthropic hasn’t published a definitive public confirmation of ranking behavior, so treat it as a strong, well-reasoned observation rather than an official statement.
Should I optimize differently for Claude vs. ChatGPT?
Somewhat. For Claude, prioritize Brave (and Google) rankings and freshness/comparison content. ChatGPT draws on its own retrieval, so a broader GEO foundation still matters across both.
How do I check if Claude cites me?
Ask Claude the questions you target and see which sources it references, then compare against your Brave and Google rankings for those queries. (This is exactly the kind of test we run in our GEO Lab.)
The bottom line
If Claude is, in effect, a friendly front-end on Brave Search, then “optimizing for Claude” becomes refreshingly concrete: earn Brave (and Google) rankings, signal freshness, and write content worth quoting. We’ll be putting this to the test ourselves and reporting what we find.
Source
Findings reported by Search Engine Land, based on research attributed to Jonathan Clark (Moving Traffic Media): “Claude visibility and Brave Search rankings.” Data points are paraphrased; figures are approximate and directional.
Leave a Reply