Being visible on LLM’s has become one of the most contested topics in search marketing: has AI search fundamentally changed the game, or is LLM optimisation mostly repackaged SEO under a new label?
Despite the noise, Google still drives the majority of search traffic, and most AI visibility is built on top of strong traditional foundations. The real question isn’t SEO or LLM, it’s understanding where they overlap, where they differ, and what actually makes an impact.
What the data tells us
Google and traditional search still dominate. Google handles roughly 373 times more searches than ChatGPT according to SparkToro and Datos clickstream data from tens of millions of desktop users. At the time of writing, AI referral traffic accounts for somewhere between 0.15% and 1.08% of total web traffic (depending on the industry and measurement source); and Google’s search market share sits at 90.71%, according to StatCounter. This is an increase of nearly 1.5% since we evaluated Google’s positioning in Search and their at-the-time poor stock performance in March 2025, which has since nearly doubled.
Thus, for the vast majority of businesses, organic search is still where their traffic is. AI search is absolutely worth understanding and monitoring, but the idea that it’s replacing Google any time soon doesn’t hold up against the numbers.
In 2024, Gartner predicted traditional search volume would drop 25% by 2026, but that prediction hasn’t materialised. Google search activity’s share of desktop usage remained flat throughout 2025, whilst SEO interest on Google Trends hit an all-time high in February 2026. This implies the gap between the narrative and the reality is significant, and the industry conversation is moving faster than the actual shift in user behaviour.
It is also worth noting the irony in the fact OpenAI, the company behind ChatGPT, is spending between $410,000 and $600,000 on two SEO-focused hires. ChatGPT itself ranks for over 287,000 keywords and receives an estimated 76.5 million organic visits per month. The company building the supposed replacement for search is investing heavily in traditional SEO to grow.
Organic search still drives AI visibility
When it comes to Google’s AI Overviews within their search engine, traditional search results are still intertwined with the AI summaries produced at the top of the results pages. Ahrefs analysed 1.9 million Google AI Overview citations and found that 76% come from pages already ranking in Google’s top 10 organic results. There is a moderate positive correlation (Spearman 0.347) between organic ranking position and AI Overview citation position. Put simply, if you are ranking well in Google, you are already appearing in their AI Overviews.
Further testing in February 2026 examined subfolders which experienced significant organic visibility declines following a recent Google algorithm update. Every single site also lost AI search citations, with an average decline of 22.5% across all LLM platforms. ChatGPT citations dropped 27.8% on average, making it more sensitive to Google organic changes than Google’s own Gemini product.
Google’s own representatives have been explicit. Danny Sullivan, Google’s Search Liaison up until mid-2025, stated that optimising for AI search “requires no fundamental changes from traditional SEO practices.” His summary, “Good SEO is good GEO.” John Mueller went further, warning that “the higher the urgency, and the stronger the push of new acronyms, the more likely they’re just making spam and scamming.”
Looking beyond Google, Bing has updated its webmaster guidelines in February 2026 to cover, “Bing search experiences, Copilot, and grounding API results.” Fabrice Canel from Bing put it plainly, “In the AI era we understand webpages perfectly, no need for sub-standard. We rank based on what customers see.”
Where AI search genuinely differs
Google AI Overviews and standalone AI assistants like ChatGPT work differently, and the distinction matters for how you approach them.
For Google AI Overviews, the previous section tells the story, 76% of citations come from the organic top 10, and drops in organic visibility cascade directly into citation losses. This is Google’s AI product querying its own index. If you rank well in Google, you show up in AI Overviews, and traditional SEO covers this.
Standalone AI assistants are a different picture. Ahrefs found that only 8-12% of ChatGPT, Gemini, and Copilot citations overlap with Google’s top 10 results, with over 80% coming from pages that don’t rank in the top 10 at all. These platforms pull from different sources, weight signals differently, and in Perplexity’s case, use an entirely separate search index (Brave rather than Google). For these assistants, brand mentions across the web, third-party consensus, and YouTube presence carry more weight than any single Google ranking.
This is where the Generative Engine Optimization (GEO) conversation has some legitimacy. Appearing in ChatGPT’s responses isn’t the same as ranking in Google’s top 10. It’s closer to digital PR and brand building; being mentioned across independent sources, appearing in third-party comparisons, and having a visible presence beyond your own website.
These AI tools use a retrieval layer (typically RAG) that does query search engines, but they also draw on pre-training data where the model has already encountered your brand across the web. There’s no single hidden ranking signal to reverse-engineer, but there is a clear pattern: brands that are visible across multiple independent sources tend to get cited more frequently.
Google’s Information Gain patent, granted in June 2024, reinforces this across both AI Overviews and standalone assistants. The patent describes a system that prioritises documents providing information beyond what other sources already cover. Original data and real expertise are what gets selected over content that repeats the consensus.
The tactics that are already backfiring
Self-promotional listicles, the most popular GEO tactic (publishing “best X” articles on your own blog and ranking yourself first), are now actively causing damage. Lily Ray documented seven SaaS brands with heavy self-promotional listicle strategies that saw organic visibility declines between 29% and 49% in January 2026. The declines cascaded into AI citation losses too. Google appears to be getting better at algorithmically detecting this pattern, and even sites with as few as 10 self-promotional listicles were affected.
Creating separate markdown files for AI crawlers has been dismissed by both Google and Bing. John Mueller went as far as to call the idea “stupid.” Fabrice Canel pointed out it doubles crawl load for no proven benefit, whilst research has shown an analysis of 300,000 domains found zero connection between having an llms.txt file and citation frequency in LLM answers.
Another emerging concern is GEO “citation building” increasingly becoming pay-to-play, with publishers charging brands to be listed on heavily cited pages. This is functionally the same paid link manipulation that Google has been fighting for over 20 years, now wearing a GEO label.
These tactics seem to be following the cyclical cycle tenured SEOs have seen for years: a spammy, grey-area tactic yields results, goes viral on socials and more mainstream, and Google responds by punishing sites spamming these tactics in the next Google Update.
So what should businesses actually do?
The practical path forward doesn’t require abandoning SEO or investing solely in a separate GEO discipline. But there definitely needs to be an awareness and understanding of where your customers are coming from, and there’s certainly value in understanding prompts and topics you’re likely to surface in.
Start with your SEO foundation. Google AI Overviews pull 76% of citations from the top 10. Bing’s guidelines for Copilot are the same as its search guidelines. If you rank well organically, you’re already visible in the highest-volume AI search product.
Earn third-party mentions. For standalone AI assistants like ChatGPT, brand mentions across independent sources carry significant weight. This isn’t a new concept. It’s brand building and digital PR applied to a new surface.
Create content with genuine information gain. Original data, unique testing results, and the specific expertise your competitors lack. This is the content Google’s Information Gain patent rewards, and it’s the content that stands out in AI retrieval where consensus repetition is filtered out.
Monitor your AI traffic. Check GA4 referral sources for ChatGPT, Perplexity, and other AI platforms. Track whether your organic visibility changes correlate with AI citation changes. This monitoring layer is worthwhile, and it’s the one area where AI search does introduce something new to the workflow.
Keep content fresh and relevant. Ahrefs found that 79% of pages cited by ChatGPT were updated in 2025. Regular content refreshes are a low-effort way to maintain both organic and AI visibility.
SEO isn’t going anywhere, and the fundamental work hasn’t changed. Create high-quality, authoritative, well-structured content and build real brand presence across the web. The data only reinforces that position.
If you’re looking for a partner to help strengthen your organic search visibility, and as a result your visibility across AI platforms ensuring your content reaches the audiences that matter, get in touch with our team.
