Google says GEO is just SEO. The other engines disagree.

Google published guidance on May 15 arguing that GEO, generative engine optimization, the practice of optimizing for LLM-driven answers, is not a separate discipline. Its position is that AEO, answer engine optimization, optimizing for direct-answer surfaces, and GEO are foundational SEO applied to an AI surface. The claim landed in the same week as I/O 2026, and the timing is not incidental. Google wants the industry to treat AI search as continuous with the search it already understands.

For Google’s own surfaces, the claim mostly holds. AI Overviews and AI Mode draw on Google’s existing ranking systems, so a page that earns strong classic visibility tends to earn AI visibility on Google too. A practitioner optimizing only for Google could accept the guidance and lose little.

The problem is that Google does not run the other engines. An analysis published by GEO consultancy Ayzeo on May 16 made the specific case: Google is right about Google and wrong about everywhere else. ChatGPT, Perplexity, Claude, and Gemini use different retrieval systems, and the gap between them is where the separate-discipline argument earns its keep.

The clearest evidence is overlap. Ayzeo, citing platform-level research, reported that only 12 percent of URLs cited by ChatGPT also rank in Google’s top 10 for the same query. If GEO were simply SEO, that number would sit much higher. A site could rank first on Google for a query and never appear in the ChatGPT answer for it. Ranking optimization and citation optimization are producing different winners.

Several mechanics explain the divergence. Search Engine Land found that 44 percent of LLM citations come from the first 30 percent of a page, which makes answer placement inside a document a retrieval factor that classic ranking never weighted. Authoritas reported that pages with FAQ schema receive roughly 40 percent more citations from ChatGPT, and that 71 percent of ChatGPT-cited pages carry structured data. Perplexity, per the same analysis, cites content updated within the last 30 days at an 82 percent rate, putting a freshness premium on pages that Google would rank fine without a recent edit.

Off-site signals diverge too. AirOps reported that 85 percent of brand mentions in AI-generated answers come from external domains, and that brands with a strong off-site presence are several times more likely to be cited. That is closer to digital PR than to on-page SEO, and it means a brand can optimize its own pages perfectly and still lose the AI mention to how often it is discussed elsewhere.

None of this makes SEO obsolete. Ayzeo’s own framing puts SEO fundamentals at roughly 80 percent of the work across every platform: crawlability, clear structure, genuine authority, and accurate information remain the base layer that every engine rewards. The argument is not that GEO replaces SEO. It is that the remaining 20 percent, the part that decides citations on non-Google engines, follows rules Google’s guidance does not describe.

The risk in accepting Google’s framing wholesale is measurement. A team that treats AI visibility as a single number tied to Google rankings will not track whether it is cited by ChatGPT or Perplexity at all, and will not notice when it loses that surface. The divergence is invisible to anyone who only watches Google.

For the next quarter, treat ranking and citation as two tracked outcomes, not one. Keep investing in SEO fundamentals as the shared 80 percent, then audit citation share on ChatGPT and Perplexity separately, because a strong Google rank no longer predicts either.