Yes. LLM SEO still runs through Google rankings, not around them. In self-reported survey data, just 4% of people click through an AI chatbot answer to its source, against 19% from search, per the Reuters Institute's Digital News Report 2026. With so little traffic feedback, an AI's citation call leans on the trust signals ranking already rewards.
Key takeaways
- LLM SEO and classic SEO are not two separate races. Ranking well in Google is the qualifying round for AI citation, not a rival channel competing for the same budget.
- Gartner has modeled that traditional search engine volume could drop 25% by 2026 as AI chatbots take over queries, a scenario projection the firm has since confirmed in direction using publisher traffic data.
- AI answer engines crawl the open web far more than they send traffic back to it, according to Cloudflare, which means your first shot at being read by these systems is the same crawl and index process that already decides your Google ranking.
- Getting cited is the beginning, not the finish. The visitor who does click still lands on a page, and that page still has to earn the reply, the demo request, or the signup.
Why does ranking in Google still matter for LLM SEO?
Ranking in Google still matters for LLM SEO because the same crawling and indexing pipeline feeds both systems. An AI model cannot cite a page it never encountered, and the pages it encounters most often are the ones already crawled, linked, and structured well enough to show up in classic search.
This is not a metaphor. Cloudflare's own network data shows how lopsided the relationship between crawling and referral traffic has become: crawl-to-refer ratios across major AI companies ranged "from Anthropic's 70,900:1 down to Mistral's 0.1:1" during a single tracked week, according to Cloudflare's analysis of AI crawler traffic (the window covered June 19 to 26, 2025, and Cloudflare notes native-app traffic sends no referrer, so the true ratios may run even higher). Some of these systems read thousands of your pages for every visitor they send you. If your site is not already worth crawling and ranking, you are simply not in that read pile, no matter how well the page itself is written.
Can you get cited by an AI engine without ranking in Google first?
You can, occasionally, but it is the exception, not a strategy for LLM SEO. A brand new domain can earn a single citation through a strong third-party mention, a widely shared original dataset, or a quote picked up by a larger outlet the model already trusts. What it cannot do reliably is earn repeat citations across many prompts the way an established, well-ranked domain does.
Picture a two-person GTM team launching a niche B2B tool. They publish one sharp, well-sourced post about their category. A credible newsletter picks it up, and for a few weeks an AI engine might cite that single post for a narrow question.
But without the backlinks, internal link structure, and crawl history that come from sustained classic SEO work, that citation does not compound into a pattern the model reinforces. The team that also does the unglamorous ranking work is the one still showing up six months later.
What actually separates a page that ranks from a page that gets cited?
The two are related but not identical, and treating them as the same thing skips real work. A page can rank without ever being cited, and in narrow cases a page can be cited without ranking well. Here is where the two overlap and where they diverge.
| Signal | Helps you rank in Google | Helps you get cited by AI |
|---|---|---|
| Backlinks and domain trust | Yes, historically the core signal | Yes, indirectly, through crawl priority and perceived credibility |
| Keyword-optimized headings | Yes | Only if the heading also answers the question in plain language |
| Self-contained, quotable sections | Helps readability, minor ranking effect | Directly determines whether a passage gets lifted |
| Freshness and recent updates | Moderate ranking factor | Strongly rewarded, engines prefer current sources |
| Named, sourced statistics | Minor ranking effect | High-leverage, models favor attributable claims |
The table makes the practical point behind this whole approach to LLM SEO: winning classic SEO gets you into the pool AI engines draw from, but extractable formatting is what gets you picked out of it. Neither alone is enough.
How should you split effort between SEO and GEO if you only have one team?
Split it in that order: fix the ranking fundamentals first, then layer in extractable formatting on the same pages, rather than building a separate GEO content track from scratch. A page that already ranks and gets a GEO rewrite compounds faster than a brand-new page built only for AI citation.
In practice that means auditing your best-performing pages for backlinks and technical health first, then rewriting the top sections into direct-answer paragraphs, adding named sources, and building out an FAQ block. This is also where adjusting the page itself for organic and AI-referred visitors pays off twice: the same visitor arriving from a Google result or an AI-cited link should land on a page built around who they are, not a generic template. It is also worth remembering the gap does not close at the citation. We have written before about why the page after the click still loses deals, and a citation from an AI engine sends exactly the kind of pre-qualified visitor that gap wastes most.
Frequently asked questions
Does SEO help LLM ranking?
Yes. Classic SEO work, backlinks, site structure, crawlability, and page speed, feeds the same index that AI models draw from when deciding what to cite. A page that already ranks well in Google has already cleared the trust bar these models were trained to recognize, which makes it a more likely candidate for citation than an unranked page with identical content.
Do LLMs use Google rankings to decide what to cite?
Not directly in every case, but there is a strong practical overlap. Models are trained on crawled web data that correlates heavily with what already ranks, and retrieval-augmented systems often pull from live search results. A page absent from Google's index is largely invisible to both channels at once, so ranking remains the more reliable lever to pull.
What belongs on an LLM SEO checklist?
Start with the fundamentals: crawlable pages, clean site structure, and real backlinks, since none of that changes for AI engines. Add a 40 to 60 word direct answer under every heading, name your sources with a hyperlink, define key terms in plain language, and close with a tightly scoped FAQ section. Update the page whenever the underlying facts change.
Do SEO and GEO work together, or are they separate strategies?
They work together on the same pages rather than as separate strategies. SEO earns the crawl priority, backlinks, and trust signals that get a page into consideration. GEO then decides whether a model actually lifts a passage from it. Treating them as competing budgets usually means neither discipline gets done well.
LLM SEO still starts with the basics
LLM SEO is not a shortcut around Google rankings, it is built on top of them. Rank well first, then rewrite your best pages into direct answers, named sources, and a clean FAQ, because that combination earns repeat citations instead of one lucky mention. Once a visitor does click through, whether from a search result or an AI answer, how Neuropage builds a page around who that visitor is is what keeps the click from going to waste. If you want to see where your own page currently stands, score your current landing page with Neuropage's free audit.