85.5% of AI citations come from earned media: what 680 million citations reveal

Third, the overlap between top Google rankings and AI-cited sources has collapsed from 70% to under 20%. In 2023, roughly 70% of pages cited in AI answers also ranked in Google's top ten for related queries. By <time datetime="2026-04">April 2026</time> that overlap had fallen below 20%, as AI engines developed citation preferences weighted toward community content and editorial coverage rather than link-built pages.

Why do AI engines prefer earned media over brand websites?#

AI engines cite earned media because retrieval-augmented generation systems weight corroboration over self-description. When a user asks "what project management tool is best for a remote team," the AI looks for consensus across multiple independent sources, not the feature pages of competing tools.

A brand describing its own virtues on its own domain scores low on the corroboration signal. A brand mentioned across industry coverage, user reviews, and forum discussions scores high. The citation logic is closer to academic citation norms than to Google's link graph: frequency and diversity of third-party reference matter more than the quality of the source content.

Lily Ray, VP of SEO Strategy at Amsive, stated at Affiliate Summit 2026: "Brand visibility becomes the new KPI, with ranking influencing RAG citations." The implication is direct: brands that build awareness through press, reviews, and community presence build the citation substrate AI engines draw from. Brands that invest only in on-page SEO build pages AI engines ignore for the 85.5%.

Rand Fishkin's SparkToro research adds a further finding: there is less than a 1-in-100 chance that ChatGPT will recommend the same list of brands twice for the same prompt. Citation pools are volatile, and a brand without broad earned media coverage cannot build the citation frequency needed to appear across repeated queries.

::info:: BrandCited's scan checks citation frequency and consistency across 8 AI engines. A brand with earned media coverage across the top citation domains scores higher on the corroboration signal BrandCited measures. See your score at brandcited.ai.

Which domains capture the most AI citations?#

Reddit, Wikipedia, YouTube, and a specific set of journalism and review sites dominate the AI citation pipeline. The 5WPR index found the top 15 domains absorb 68% of citation volume across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews. For any brand not earning coverage on those platforms, the effective AI citation share approaches zero.

The engine-level breakdown reveals a further problem. A 2026 per-engine audit found that only 11% of domains cited by ChatGPT overlap with domains cited by Perplexity. Each engine maintains a distinct citation pool. A brand appearing in ChatGPT answers has less than a 1-in-9 chance of appearing in Perplexity answers for the same query. Engine coverage gaps compound: a brand visible on one engine is invisible on the others unless it has coverage across different publication types.

AI search visits reached 27.4 billion in Q1 2026, up 42.8% year over year. ChatGPT confirmed 1 billion monthly active users in <time datetime="2026-05">May 2026</time>, the fastest app in history to reach the milestone according to Sensor Tower data cited by Reuters. Perplexity processes an estimated 1.2 to 1.5 billion queries monthly as of mid-2026. The scale of the citation gap for brands without earned media is no longer theoretical — it maps to audience size measured in billions.

Structured data narrows the gap for brand-owned content. BrightEdge's 2026 analysis found that sites with complete FAQ schema see 44% more AI Overview appearances, and 65% of pages cited by Google AI Mode include structured data. Schema maximizes the 14.5% of citations going to brand sites — the 85.5% going to earned media requires a distinct strategy.

What breaks for brands that rely only on their website?#

The earned media finding breaks three assumptions that underpin most brands' current AI visibility plans.

The assumption that Google rankings predict AI citations. With the Google-to-AI citation overlap below 20%, a brand ranking in Google's top three positions on a commercial query has less than a 1-in-5 chance that the same page appears in AI answers for related questions. Rankings and citations have diverged.

The assumption that better owned content is the primary lever. On-page improvement matters within the 14.5% of citations that go to brand websites. For the other 85.5%, the lever is third-party coverage, not content quality on your domain. A brand that has optimized its product pages but has no coverage on Reddit, G2, or trade publications is invisible to AI engines for the majority of citation opportunities.

The assumption that citation monitoring is optional. Brands cited in AI Overviews earn approximately 120% more organic clicks per impression than uncited brands on the same queries, per Seer Interactive's 2026 analysis. Without tracking where AI engines cite your brand — and where they cite competitors instead — brands cannot measure whether their earned media investments are generating AI visibility. The measurement gap has a direct revenue consequence.

Run a free BrandCited scan to see your current citation rate across 8 AI engines. The scan shows which queries your brand answers and which your competitors answer instead, with every gap ranked by impact: brandcited.ai

What should your brand do right now?#

1. 1Map your current earned media footprint. Search your brand name on each major AI engine and note which sources they cite. These are the domains already in your citation pool. More coverage on those same domains builds citation frequency faster than targeting new publications.

1. 1Identify the 10 most-cited publications in your category. For B2B SaaS, that typically includes G2, Capterra, TrustRadius, and the relevant trade press. For consumer brands, it includes Reddit communities, YouTube channels, and review platforms. AI engines cite these domains across all five major platforms — one well-placed review generates more citation potential than extensive owned blog content.

1. 1Add Organization schema and FAQPage schema to your website. Schema maximizes the citation value of the 14.5% of citations that go to brand-owned sites. Use JSON-LD, the format every major AI engine prefers. Here is the Organization schema every brand needs:

1. 1Pursue reviews on the platforms AI engines actually cite. G2, Reddit, and Wikipedia appear in the top citation domains across multiple engines. A single in-depth G2 review citing specific use cases generates more AI citation potential than a hundred owned blog posts.

1. 1Add a named author block to every article. AI engines weight content with named, verifiable authors more heavily than anonymous posts. A named author with a LinkedIn profile and a publishing record builds the entity graph corroboration that raises citation rates. Include sameAs with the author's LinkedIn URL in Article schema.

1. 1Track your citation rate monthly with BrandCited. AI citation pools shift as models update. A score this month is not a score in three months. Tracking the trend is the only way to know whether earned media investments are closing the visibility gap.

AI search news from the past 24 hours#

• Siri AI becomes the 9th AI citation surface: Apple shipped Siri AI at WWDC 2026 on <time datetime="2026-06-08">June 8</time>, backed by a custom Gemini model. iOS 27 ships in September to 1.8 billion active devices. Brands invisible to Gemini will be invisible to Siri from day one. (BrandCited coverage)
• GPT-4.5 retires June 27: OpenAI confirmed GPT-4.5 will be removed from ChatGPT on <time datetime="2026-06-27">June 27</time> following a 30-day sunset. Brands should validate citation scores under GPT-5.5, which has different citation behavior. (OpenAI release notes)
• ChatGPT reaches 1 billion monthly users: ChatGPT confirmed 1 billion monthly active users in <time datetime="2026-05">May 2026</time>, the fastest app in history to reach the milestone per Sensor Tower data cited by Reuters. (Reuters via Quartz)
• Google AI Overviews opt-out deadline June 17: Site owners have until <time datetime="2026-06-17">June 17</time> to opt out of Google AI Overviews. Brands that stay in and get cited earn 120% more clicks per impression than uncited brands on the same queries. (BrandCited coverage)
• Perplexity processes 1.5 billion queries monthly: Perplexity's June 2026 growth estimates put it at 1.2 to 1.5 billion search queries per month, driven by Comet browser adoption and mobile growth. (Perplexity statistics, getpanto.ai)

Run a free BrandCited scan to see your citation rate across all 8 AI engines in 30 seconds: brandcited.ai

Frequently asked questions#

Earned media refers to content published about a brand on third-party domains the brand does not own or control. Press coverage, independent reviews, Reddit discussions, analyst reports, Wikipedia entries, and YouTube mentions all qualify. The 5WPR analysis of 680 million AI citations found that 85.5% of citations reference earned media sources rather than brand websites, making third-party coverage the primary driver of AI citation volume.

Improving on-page SEO and adding structured data helps for the 14.5% of AI citations that originate from brand-owned websites. BrightEdge data shows sites with complete FAQ schema see 44% more AI Overview appearances, and 65% of pages cited by Google AI Mode include structured data. But the 85.5% of citations going to earned media sources requires building coverage on the platforms AI engines already cite — not optimizing your own domain.

Citation pools shift as models update and retrieval systems evolve. The 5WPR research tracked patterns from August 2024 to April 2026 and found the overlap between Google rankings and AI citations dropped from 70% to under 20% during that period. BrandCited monitors citation rates across 8 engines, so changes in which queries your brand answers surface in the score trend month over month.

No. Only 11% of domains cited by ChatGPT overlap with domains cited by Perplexity, per a 2026 per-engine audit. Each engine maintains a distinct citation pool. Brands need coverage across multiple publication types to build citation visibility across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews at the same time.

Start with the platforms AI engines cite most: a review site for your category (G2, Capterra, or equivalent), Reddit communities where your customers discuss their problems, and the trade publications that cover your space. A single well-cited G2 review generates more AI citation potential than extensive owned blog content. BrandCited's free scan shows your current baseline and which competitors are being cited on queries your brand misses.