Last updated July 2026
The objection sounds reasonable: “We already do SEO. AI visibility is just the same thing with a new name.”
It is the most common pushback in the room, and it is wrong. Not slightly off. Wrong in a way that creates a measurable competitive gap for the teams that believe it.
The evidence below is drawn from two of the most-cited datasets on this question. Neither requires a leap of faith. The numbers do the work.
Section 1: The 12% overlap problem
In August 2025, Ahrefs researchers Louise Linehan and Xibeijia Guan published a study testing 15,000 long-tail queries across ChatGPT, Gemini, Microsoft Copilot, and Perplexity. They cross-referenced the URLs cited in each AI response against Google’s top-10 organic results for the same query.
The headline finding: only about 12% of AI-cited URLs also rank in Google’s top 10 for the same prompt.
That average conceals a wide range by platform:
| AI platform | Overlap with Google top 10 |
|---|---|
| Perplexity | ~29% |
| Gemini | ~12% (blended average) |
| ChatGPT | ~8% |
| Microsoft Copilot | ~8% |
| Combined average | ~12% |
Source: Ahrefs, “Only 12% of AI Cited URLs Rank in Google’s Top 10 for the Original Prompt,” August 11, 2025 (15,000 long-tail queries via Ahrefs Brand Radar).
Perplexity sits higher because it is a web-search-first product: it actively retrieves recently indexed pages, creating a tighter relationship with Google’s index. ChatGPT and Copilot rely more heavily on training data and a distinct retrieval layer. That architectural difference produces a gap that a search-only strategy cannot bridge.
The practical implication: if you are in the ~88% of Google-top-10 pages that AI assistants do not cite, your SEO investment has done nothing for your AI visibility. The channels are parallel, not synonymous.
Section 2: The 11% cross-engine overlap problem
The decoupling goes deeper than Google versus AI. It exists inside AI itself.
According to Profound’s analysis of 100,000 prompts run across ChatGPT and Perplexity (published July 2025), only 11% of cited domains appear in both platforms. The other 89% of citations are platform-specific.
That finding has a direct operational implication. If your brand is getting cited by ChatGPT, you cannot assume it is getting cited by Perplexity. If you measure only one engine, you are reading a partial picture. A buyer on Perplexity and a buyer on ChatGPT asking the same question are likely seeing different brands recommended.
Note: Profound is an AI citation-tracking vendor. The study is self-published research, not an independent peer-reviewed paper. The 11% figure is widely cited in secondary sources but should be read with that context in mind.
The takeaway stands regardless: tracking a single AI engine gives you one slice of a fragmented landscape. A full AI visibility program requires measuring each platform separately.
Section 3: Why the signals diverge
Understanding why the signals diverge makes the separation feel less surprising and more actionable.
SEO ranking signals include backlink authority, on-page keyword relevance, technical crawlability, and click-through rate patterns across Google’s index. Google ranks pages against each other using these signals, and a high position reflects competitive strength on those dimensions.
AI citation signals work differently. AI engines select sources based on a combination of:
- Training data that predates any query (content absorbed during model training)
- Retrieval-augmented generation, pulling from a live index that is not identical to Google’s
- The specific phrasing of the prompt, which can activate different retrieval pathways for the same underlying question
- How clearly the content answers the question as structured text
A page can rank in position one on Google because it has the most backlinks in a competitive set and still be invisible in AI answers because it is not described clearly enough for a model to extract a quotable claim from it. A page can be cited constantly by ChatGPT while ranking nowhere near Google’s top 10 because it is an authoritative third-party source the model trusts.
The signals select for different properties of a page. That is the root of the divergence.
Section 4: What a decoupled AI visibility strategy requires
If the two signals are mostly separate, optimizing for one does not automatically optimize for the other. Here is what the decoupling means in practice:
You need dedicated measurement. A Google rank tracker tells you nothing about AI citations. You need to run your target buyer prompts directly against ChatGPT, Perplexity, Gemini, and Google AI Overviews and measure how often your brand appears versus competitors. That is share of voice in AI, and it cannot be inferred from search rankings.
You need to track multiple engines. The 11% cross-engine overlap from Profound’s research means that a citation strategy optimized for ChatGPT may not move the needle on Perplexity at all. Each engine requires its own tracking layer.
Third-party authority matters more than on-site optimization. Studies consistently find that the large majority of AI citations come from third-party sources rather than brand-owned websites. Being described accurately and positively in authoritative earned media is a higher-leverage activity for AI visibility than adding more schema markup to your own pages. An Ahrefs study tracking 1,885 pages that added JSON-LD schema found no statistically meaningful uplift in AI citations.
Prompt phrasing changes outcomes. AI engines do not produce consistent results across rephrasings of the same intent. A monitoring program that tracks only one phrasing per topic is measuring noise, not signal. Effective tracking covers a cluster of variants per buyer intent.
Platforms built specifically for this work include Profound (deep citation intelligence and cross-engine analysis), Otterly.AI (structured prompt tracking with GEO audit guidance), Semrush (which has added AI visibility modules to its broader SEO platform), and Temso (an all-in-one AI SEO platform from $89/mo that tracks share of voice across eight engines and converts monitoring data into a prioritized action plan). These tools are measuring a separate signal from your rank tracker. They are not redundant.
For a full comparison, see /rankings/ai-visibility-tools. Definitions of key terms like share of voice and citation rate are at /glossary.
The decoupling table
| Dimension | Traditional SEO | AI visibility |
|---|---|---|
| What is measured | Position 1–10 in Google’s organic results | Citation rate across AI-generated answers |
| Signal source | Google’s index and ranking algorithm | Training data, retrieval layer, prompt phrasing |
| Consistency | Stable position for a given keyword | Probabilistic: same prompt can produce different answers |
| Key inputs | Backlinks, on-page keywords, technical health | Authoritative earned media, structured content, clear factual claims |
| Cross-platform transfer | Rankings in Google do not transfer to Bing automatically | 11% domain overlap between ChatGPT and Perplexity (Profound, 2025) |
| Google-AI link overlap | High-ranked pages often appear in AI citations | Only ~12% of AI-cited URLs rank in Google’s top 10 (Ahrefs, 2025) |
| Tool overlap | Semrush, Ahrefs, Screaming Frog | Profound, Otterly.AI, Temso, Semrush AI modules |
Sources: Ahrefs study of 15,000 queries (August 2025); Profound analysis of 100,000 prompts (July 2025). See /methodology for editorial standards.
What this means for your next conversation
The “AI visibility is just SEO” objection usually comes from someone who is not wrong about SEO’s value. SEO still matters. Google search still drives significant traffic. The mistake is in assuming the two programs are additive and interchangeable.
They are parallel channels with mostly separate source pools. A brand can dominate Google organic and be invisible in ChatGPT responses. A brand can be consistently cited by Perplexity while ranking on page three of Google. The data makes this concrete: 88% of AI-cited content is not in Google’s top 10. 89% of citations are platform-specific between ChatGPT and Perplexity.
If your team is running both programs, you need to measure both programs separately. If your team is running only one of them, you are optimizing for half the channel mix.
See which tools can give you both signals in one workflow: /rankings/ai-visibility-tools