Why AI Search Engines Ignore Your Website (And Cite Competitors Instead)
AI search engines show a measured bias toward earned media over brand-owned content in the studies tested. Here's why your website may be skipped, what the research says, and what to do about it.
AI search engines systematically cite earned media over brand-owned content. Multiple independent studies confirm this at scale: when ChatGPT, Perplexity, Gemini, and Claude synthesize answers, they pull from independent publications and earned coverage — not your website. In Ahrefs' and Moz's data, most citations come from pages that do not rank in the Google top 10 for the exact query, so SEO rank alone does not secure a place in the answers now mediating how buyers research vendors.
This article explains why the bias may exist, what the data shows, and where the evidence says to put effort.
Research summary: earned media dominance across AI engines
Every major study measuring AI citation sources reaches the same conclusion. Earned media is the dominant source category across all AI search engines tested.
| Study | Sample | Key Finding |
|---|---|---|
| arXiv 2509.08919 — unrefereed preprint (Sep 2025) | Multi-vertical; Google vs. ChatGPT search | In its Canadian automotive, consumer-electronics and software tests, earned sources were 69-78% of ChatGPT-search sources vs. 32-54% for Google |
| Chen, Wang, Chen & Koudas, "Navigating the Shift" (EDBT/ICDT 2026 workshops, University of Toronto) | Large-scale source typology | Claude highest at 65% earned; Gemini 46%; social nearly absent |
| Stacker/Scrunch (Mar 2026) | 87 stories, 30 brands, 2,600+ prompts | 239% median lift in AI citations from earned distribution |
| Ahrefs (May 2025) | 75,000 brands screened (about 74%, those with AI Overview mentions, analyzed) | Branded web mentions correlate about 3x more strongly with AI Overview brand visibility than backlinks (Spearman 0.664 vs. 0.218) |
| Muck Rack, What Is AI Reading? (May 2026) | 25M+ links cited by ChatGPT, Claude, and Gemini across 17 industries | 84% of cited links came from earned media, the sources brands neither own nor pay for — journalism, third-party corporate content, Wikipedia, Reddit, PubMed, government and academic sources combined; journalism alone 25-27% |
| Fullintel-UConn (IPRRC, Mar 2026) | 400 prompts, 10 personas, one engine, one topic | 47% of citations to third-party news and informational sources; 48% to corporate, university and health-network sites |
| Moz (2026) | 40,000 queries | Moz's 2026 analysis of 40,000 queries found that 88% of Google AI Mode citations are not in the organic top 10 for the same query. The figure is the overlap between those two result sets. |
What the research found: AI engines have a structural bias toward earned media
In September 2025, an unrefereed arXiv preprint published a comparative analysis of AI search engines and traditional web search. Its authors ran controlled experiments across multiple verticals, languages, and query types to measure where AI engines source their citations versus Google.
The finding was unambiguous: AI search engines exhibit a systematic and overwhelming bias toward earned media over brand-owned and social content — a stark contrast to Google's more balanced mix.
The vertical-level data:
| Vertical | Google earned share | AI search earned share |
|---|---|---|
| Software products | 45.4% | 72.7% |
| Consumer electronics | 54.1% | 77.6% - 92.1% |
| Automotive | 40.6% | 69.1% - 81.9% |
The spread narrows for transactional queries, but for the informational and consideration queries that drive B2B research and vendor selection, earned media is what AI engines reach for first.
A second study — Chen, Wang, Chen and Koudas's "Navigating the Shift", from the University of Toronto and presented at the EDBT/ICDT 2026 workshops — measured source typology across AI engines. Claude concentrated most heavily on earned sources at 65%, followed by GPT-4o at 57%, with Gemini at 46%. Perplexity and Gemini were more balanced on consideration queries. Social content was nearly absent from AI search results entirely.
The pattern is consistent: AI engines were trained on the open web. The open web's most credible content is in editorial publications, not brand sites. That training bias is plausibly reflected in citation behavior, though the studies above measure which sources get cited, not whether on-page optimization can shift them.
How earned media distribution multiplies AI citations
The earned-media bias raises a follow-on question: does distribution across multiple publications compound the signal?
Stacker's March 2026 study found a 239% median lift in AI citations when content moved through earned distribution channels versus brand-owned content alone. The study, conducted with Scrunch AI analytics, analyzed 87 stories across 30 brands, queried 2,600+ prompts across 8 AI platforms over 30 days, and was reported by GlobeNewswire (March 16, 2026).
Key numbers from the Stacker/Scrunch study:
- 239% median increase in AI citations from earned distribution vs. owned content
- Cross-platform AI coverage increased from 5.4% to 17.9% at the median — nearly tripling coverage breadth
- 97% of distributed stories earned at least one AI citation vs. 82% for owned-content-only stories
Noah Greenberg, CEO of Stacker: "AI search isn't a single ranking position; it's a long tail played across platforms, prompt variations, and answer formats. Our data shows that coverage breadth is the new authority signal."
Why AI engines were built to discount brand content
The bias is structural, not accidental. It follows from how large language models learn to attribute credibility.
AI engines were trained on the open web, where independent editorial publications represented the highest-credibility sources. During training, the models observed that authoritative third-party sources — Forbes, TechCrunch, Reuters, Harvard Business Review — were the category humans trusted, cited, and linked to. Those training signals became citation signals.
When an AI engine synthesizes an answer, it draws on learned associations between source types and credibility:
- Independent publications with editorial standards = high credibility
- Brand sites = self-advocacy
- Social content = nearly excluded
The September 2025 preprint also flags a "big brand bias" in how AI search treats well-known and niche brands, and says niche brands must over-invest in building tangible, verifiable authority to overcome it. For a brand without third-party coverage, that means the signal AI engines can find is weak or absent in AI-generated answers.
These studies find that most citations come from outside the top ten and that earned sources outweigh brand-owned ones in the verticals and engines tested; they do not test schema markup, blog volume or SEO agencies, and they do not show that SEO work has no effect. Ahrefs and Moz both find that ranking helps. What the evidence does show is that rank alone does not secure a citation, and that where citations concentrate is third-party editorial coverage, which has to be earned.
Brand mentions track AI Overview visibility three times more closely than backlinks
Ahrefs studied 75,000 brands (May 2025) and found that brand web mentions correlate three times more strongly with AI Overview visibility than backlinks — 0.664 for mentions versus 0.218 for backlinks.
The distribution was stark:
- Top 25% of brands by web mentions averaged 169 AI Overview mentions, over 10x the 14 averaged by the next quartile
- Bottom 50% for web mentions barely registered in AI Overviews, averaging just 0 to 3 mentions
Tim Soulo, CMO at Ahrefs: "You just need to see where your competitors are mentioned, where you are mentioned, where your industry is mentioned. And you have to get mentions there — because then if the AI chatbot would do a search and find those pages and create their answer based on what they see on those pages, you will be mentioned."
This inverts standard SEO logic. Traditional search optimization was about engineering specific pages for specific queries. AI citation optimization is about building a web of earned mentions across trusted third-party sources. The currency changed from links to mentions. The source type changed from any indexed page to editorially independent publications.
Which brands are winning in AI search and why
The same brands appear consistently across AI-generated answers in competitive categories, and they also have earned coverage in the publications AI engines draw from. The data below is consistent with that, though it does not establish cause:
- Muck Rack's What Is AI Reading? study (May 2026, 25M+ cited links): 84% of cited links came from earned media, the sources brands neither own nor pay for — a category spanning journalism, third-party corporate content, Wikipedia, Reddit, government, and academic sources, with journalism alone at 25-27%
- Fullintel-University of Connecticut study (IPR Research Conference, March 2026): 47% of citations came from third-party news and informational sources, with another 48% from corporate, university and health-network sites, across 400 prompts run on a single engine and a single topic
- WorldCom PR Group (160 independent PR agencies): research shows up to 90% of citations driving brand visibility in LLMs come from earned media
The brands winning in AI search built earned coverage in the publications AI engines index as authoritative sources before the measurement studies arrived. The studies above associate their presence in AI answers with that coverage; they do not isolate its effect from on-page optimization, content volume, or domain authority in the traditional SEO sense.
The Gartner CMO gap: 65% expect disruption, 32% are changing skills
Gartner's February 2026 survey of 402 senior marketing leaders quantified the readiness gap: 65% of CMOs expect AI to dramatically change marketing within two years, but only 32% believe significant skill changes are needed.
That gap is the problem in data form. CMOs who know the rules are changing are not updating their playbooks at the same pace. The implicit assumption is that traditional brand authority investments transfer to AI search. The overlap studies above show that rank alone does not carry over: most AI-cited pages sit outside the top ten.
Gartner's analysts: "CMOs must build the literacy to prioritize high-impact use cases, validate outputs and manage risk. Otherwise, AI becomes something happening around them, not led by them."
The brands that recognized the shift early are building what Machine Relations calls Earned Authority — Layer 1 of the Machine Relations stack. Earned media placements in Tier 1 publications that AI engines already trust. Without that foundation, the layers above it (entity clarity, citation architecture, distribution across AI surfaces) have nothing authoritative to build on.
The owned content trap: why publishing more owned content may not move AI citations
There is a paradox marketers running content programs need to understand. The studies above do not show that publishing more owned content improves AI citation rates; they find that earned sources outweigh brand-owned ones.
As brands recognize that "content is how you show up in AI search," owned content volume is increasing. But the studies above find earned sources outweigh brand-owned ones in what AI engines cite, and none of them shows that adding owned-content volume changes that balance.
The September 2025 preprint reports that social content contributes very little to AI-search answers, and that earned sources outweigh brand-owned ones in the verticals it tested. On this evidence, producing more of the underweighted content type is not the lever to rely on.
The correct correction is a category switch: from owned content that AI engines discount to earned coverage that AI engines trust. The distinction is not about writing better content. It is about where that content lives and who is vouching for it.
A piece of analysis on a brand blog carries the authority of the brand that wrote it. The same analysis, covered by Forbes or placed as a contributed article in Harvard Business Review, carries the authority of the publication that independently decided to run it. AI engines read those signals differently. The studies above suggest where the content lives is a large part of how AI engines weigh it.
Coverage breadth: why single-engine optimization fails
Coverage breadth — the percentage of relevant AI platforms where a brand surfaces consistently — is the metric that explains cross-platform AI visibility.
AI visibility is not a single-engine problem. ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews draw from overlapping but distinct source pools. The Yext analysis of 17.2 million distinct AI citations across six platforms (January 2026) found that Gemini favors first-party sites while Claude cites user-generated content at 2-4x higher rates. No single optimization strategy works uniformly.
| AI Engine | Source Bias | Coverage Strategy |
|---|---|---|
| ChatGPT (GPT-4o) | 57% earned, strong editorial preference | Forbes, TechCrunch, industry publications |
| Claude | 65% earned, highest editorial concentration | Broad editorial + high-authority domains |
| Perplexity | More balanced, but earned-dominant on consideration queries | Wikipedia, editorial, niche publications |
| Gemini | Favors first-party sites more than peers | Combination of owned + earned coverage |
| Google AI Overviews | About 62% of cited URLs NOT in the organic top 10 (Ahrefs) | Independent authority sources |
Coverage breadth solves this. A brand with earned coverage in multiple high-authority publications has placed its authority signal across the range of source pools AI engines pull from. Moz's 2026 analysis of 40,000 queries found that 88% of Google AI Mode citations are not in the organic top 10 for the same query. The figure is the overlap between those two result sets. Cited pages and top-10 pages overlap far less than traditional SEO assumes.
What PR practitioners and CMOs are saying from inside the shift
The earned media bias is no longer only visible in academic research. It is showing up in practitioner conversations from the comms and PR side simultaneously.
Brian Olson, brand PR lead at Hormel Foods (January 2026): "By the end of 2026, appearing in LLM responses will stand shoulder-to-shoulder with impressions, which continue to lose relevance as a primary KPI."
Gab Ferree, founder of Off the Record, at an Axios HQ webinar (February 2026), reported by Stacker: "Media relations are becoming machine relations. It's on the comms professionals to learn the patterns of AI and then take action on them."
These are not AI tool companies with a product to sell. These are PR practitioners and comms leaders describing a shift they are seeing from inside their work, arriving at the same conclusion the research data reached: earned media in trusted publications is what AI engines cite, and that makes earned media the infrastructure of AI-era brand visibility.
How Machine Relations systematizes the earned media advantage
Machine Relations — the discipline coined by Jaxon Parrott, founder of AuthorityTech, in 2024 — names and systematizes the pattern the research describes.
Machine Relations is the discipline of earning AI citations and recommendations for a brand by making that brand legible, retrievable, and credible inside AI-driven discovery. The stack has four layers:
| Layer | Name | Function | Why It Depends on Earned Media |
|---|---|---|---|
| 1 | Earned Authority | Trusted third-party coverage in publications AI systems recognize | Foundation — without earned coverage, AI engines have no credible signal to cite |
| 2 | Entity Clarity | Unambiguous brand identification in AI systems | Strongest validation signals come from earned coverage, not owned properties |
| 3 | Citation Architecture | Content structured for AI extraction | On-site structure helps extraction, but in the studies above earned sources still account for more cited sources than brand-owned ones |
| 4 | Distribution (GEO/AEO) | Amplification across AI surfaces | Amplifies a signal that either exists or doesn't — if earned authority is absent, distribution pushes nothing |
The brands invisible in AI-generated answers are not failing at Layer 3 or Layer 4. They are failing at Layer 1. Their SEO investment is strong. Their content program is running. Their entity signals are clean. Those assets help, but on the evidence above they are not what most citations rest on: most cited pages come from outside the top 10, and earned third-party coverage is where much citation concentrates.
As Jaxon Parrott wrote in his Machine Relations breakdown on Medium: "PR's original mechanism — earned media, direct editorial relationships, third-party credibility from real publications — is the exact mechanism AI engines use to decide what to cite. The publications haven't changed. What changed is the reader."
What to do about it
The research points to a specific correction. The correction is earned media coverage in publications AI engines trust, distributed across enough publications to build coverage breadth across AI platforms.
That work requires relationships, not just content. A placement in Forbes or TechCrunch requires a direct relationship with an editor or journalist who trusts the source. Those relationships take time to build — or they can be accessed through an agency that already has them.
The evidence does not point to more owned-website content as the lever: in the studies above, earned sources outweigh brand-owned ones in what AI engines cite, which is consistent with a structural preference for them.
Ahrefs' analysis of the 1,000 pages ChatGPT cited most in September 2025 found that, among the cited pages that also rank in Google, 65.3% sit on domains with Domain Rating 81 or higher (median DR 90). 28% of those 1,000 pages had no organic visibility. Ahrefs measured the domain-rating distribution of pages already in that citation inventory. The study does not establish that Domain Rating causes citation, that Domain Rating 81 is a citation requirement, or that changing Domain Rating changes which sources are selected. The GEO paper (Aggarwal et al., KDD 2024) found, on its GEO-bench benchmark, that adding statistics, quotations or cited sources lifted a position-adjusted visibility score by 30-40% for its top methods, on the authors' own test engine rather than a live answer engine. That benchmark measures what happens when content changes; the Ahrefs figure above measures where ChatGPT's citations land, and domains rated 81 or higher skew toward large, established sites rather than long-tail content sites.
FAQ
Why do AI search engines prefer earned media over brand content?
AI engines were trained on the open web, where independent editorial publications represented the highest-credibility sources humans cited and returned to. That training created a structural preference for earned media over brand-owned content. According to an unrefereed September 2025 arXiv preprint, earned sources were 69-78% of ChatGPT-search sources in its Canadian automotive, consumer-electronics and software tests, compared with 32-54% for Google. The bias is a learned pattern from the training corpus reflecting how editorial authority was distributed across the open web.
Does SEO performance affect how AI engines cite a brand?
Ranking helps, but it is not the main route to a citation. Moz's 2026 analysis of 40,000 queries found that 88% of Google AI Mode citations are not in the organic top 10 for the same query. The figure is the overlap between those two result sets. Ahrefs' study of 15,000 prompts found that, on average, only about 12% of links cited by the AI assistants it tested appear in Google's top 10 for the same prompt (Perplexity is the outlier, at nearly 1 in 3), and concluded that "ranking well does help", though not necessarily for the queries you might assume. Ahrefs' 2026 AI Overviews update found that about 38% of cited AI Overview pages still rank in the organic top 10. And in Ahrefs' study of 75,000 brands, branded web mentions correlate about three times more strongly with AI Overview brand visibility than backlinks (0.664 vs. 0.218; a correlation, not proof of cause). The overlap figures describe how much cited pages and top-10 pages coincide, not the effect of SEO performance. Together they suggest that earned authority in trusted publications, not technical SEO optimization alone, is where much AI citation concentrates.
Who coined Machine Relations?
Jaxon Parrott, founder of AuthorityTech, coined the term Machine Relations in 2024 to name the discipline of earning AI citations and recommendations for brands. AuthorityTech is The World's First Machine Relations Agency, and its clients include 20+ unicorn startups.
How much of AI citation comes from earned media?
Multiple studies quantify this: Muck Rack's May 2026 What Is AI Reading? analysis of more than 25 million cited links found 84% came from earned media, the sources brands neither own nor pay for — a category that spans journalism, third-party corporate content, Wikipedia, Reddit, government and academic sources, with journalism alone at 25-27%. WorldCom PR Group cited research showing up to 90% of AI brand citations from earned media. The Stacker/Scrunch study found a 239% median lift from earned distribution.
What is coverage breadth and why does it matter?
Coverage breadth measures the percentage of AI platforms where a brand surfaces consistently across prompt variations. Different AI engines (ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews) draw from distinct source pools. Stacker's March 2026 research found earned distribution increased cross-platform coverage from 5.4% to 17.9% at the median. A brand with coverage across multiple high-authority publications surfaces across more of those pools.
What is the difference between earned media and owned content for AI visibility?
Earned media is coverage in third-party editorial publications the brand did not pay for or control editorially. Owned content is anything on a brand's website, blog, or social channels. AI engines treat these categories differently: third-party editorial coverage carries independent vouching, while owned content carries self-assertion. In the unrefereed September 2025 arXiv preprint, earned sources were 69-78% of ChatGPT-search sources in its Canadian automotive, consumer-electronics and software tests, compared with 32-54% for Google.