Earned media for GEO
Earned media is the GEO tier
nobody optimizes.
An AI engine's picture of your brand is assembled from everywhere it has read about you, not just your own website. Across 75,000 brands, how often a brand was mentioned predicted AI visibility far more than its backlinks. Earning that presence is a discipline most teams never work on, which is exactly why it's the largest untapped advantage.
The third tier of AI citation · Last updated
For the marketer who has done the on-page work and is asking what's next.
AI engines decide who to cite partly from your own pages and partly from everywhere else they have read about you: industry press, the community threads buyers trust, video, and reference sources. Across 75,000 brands, how often a brand was mentioned across the web predicted AI visibility far more than its backlinks did. That makes a large part of generative engine optimization an earned-media discipline, not a publishing one. Resonate Labs maps where your brand is and isn't mentioned across the sources engines read, then earns the presence that's missing.
Mentions beat backlinks
Across 75,000 brands, branded web mentions predicted AI visibility far more strongly than backlinks. Domain authority barely predicted it at all.
The model reads everywhere
An engine builds its understanding of you from press, forums, video, and reference sources, not just your domain. Presence there is the lever.
Earned, not bought or built
This is the hardest tier to fake and the slowest to build, which is exactly why the advantage it creates compounds and holds.
Why mentions beat backlinks
The instinct carried over from SEO is that authority and backlinks earn the mentions. For AI citation, the data points somewhere else. In an analysis of 75,000 brands across Google's AI Overviews, Ahrefs measured which signals correlate with AI visibility. The strongest was branded web mentions, at a rank correlation of about 0.664. Branded anchor text was next, at 0.527. Backlink count came in near the bottom at 0.218, and Domain Rating explained almost none of the variance. The signal traditional SEO works hardest barely moved whether AI cited a brand; how often the brand was talked about moved it most.
The reason is mechanical. A model's understanding of your brand is built from everywhere it has read about you, not just from your own domain. The industry press, the podcasts, the community threads, the comparison roundups, the analyst write-ups: all of it goes into the picture the model holds before any single page of yours is even crawled. The more often and more consistently your brand shows up in those places, described the way you want to be described, the more readily a model reaches for you when a buyer asks. Mentions are how you become part of the model's prior.
This is the third tier of what drives citation. The first two, a page an engine can read and content it can quote, are about your own site. This one is about your presence everywhere else, and it is the one almost nobody works on deliberately. That neglect is the opportunity: it is where the largest untapped advantage usually sits.
Where AI reads about you
If presence is the lever, the question is presence where. An engine's view of your brand is assembled from a handful of source types, and they are not interchangeable. These are the categories that matter; which specific outlets and threads matter for a given brand depends on its category and where its buyers research, which is the part worth mapping deliberately.
Editorial
Industry press and analysts
Coverage in trade publications, business press, and analyst write-ups. It carries weight twice over: engines treat established publications as credible, and the coverage describes your brand in specific language a model can lift directly into an answer.
Community
Forums and threads
The places engines treat as firsthand testimony. Reddit alone accounts for roughly a fifth of third-party citations in B2B software answers. You can't fabricate this, but you can earn it by being genuinely present and useful where your buyers already talk.
Video
YouTube and video mentions
Video has quietly become one of the strongest signals associated with AI visibility, across more than one engine. A brand that shows up in demos, talks, and explainer content is building a mention surface most competitors ignore entirely.
Reference
Wikipedia and reference sources
The encyclopedic and reference sources engines lean on heavily. Presence here accelerates everything else, with an important caveat covered next: a large share of reference citations go to sources you will never influence.
The principle is the same across all four: be mentioned often, described the way you want to be described, in places a model trusts. The work is earning that presence on purpose rather than hoping it accumulates.
The reference gate
Reference presence deserves its own note because it behaves like an accelerant. In one client study, brands with a Wikipedia article reached their first ChatGPT citation in about 28 days on average, against about 52 days without one. The sample is small and the timing varies widely by method, so treat it as direction, not a guarantee: presence in the reference sources engines trust tends to get you cited sooner and more often.
Two cautions keep this honest. The first is that reference presence is earned, not declared. A Wikipedia article has to clear a real notability bar and stay well-sourced; a thin or self-serving entry gets removed and can cost you more than it gains. The second is the ceiling. A meaningful share of AI citations, on the order of two-thirds in some analyses of the most-cited pages, go to sources you cannot influence at all, like Wikipedia itself and major reference publications. Don't fight for those. The earned-media work is about the addressable share, the citations your own presence across the web can actually win.
Earning vs. link-building
This reframes a big part of GEO as a public-relations and earned-media discipline rather than a content-publishing one. Getting quoted in a respected industry publication, showing up in the comparison roundups buyers read, being the name founders mention on podcasts, contributing accurate and specific information where the communities a model trusts can see it: these are citation levers now, not just awareness plays. They build the presence that on-page work alone cannot.
It is worth being clear about what this is not. It is not a link-building program with a new name. A link-building program optimizes for the count and authority of pages pointing at you, which the data says barely moves AI citation. Earned media for GEO optimizes for how often and how credibly your brand is described across the sources engines actually read, and it is judged by whether your AI visibility improves, not by how many placements landed. The motions can look like digital PR; the targeting and the scorecard are different.
It is also the hardest tier to fake and the slowest to build, which is exactly why it is defensible. On-page quality can be matched by a competitor in a sprint. A reputation built across the sources engines trust cannot. The brands that treat earned coverage as a citation lever, and not only an awareness play, build an advantage that compounds.
Where Resonate Labs fits
Most brands have no read on this tier at all. They know roughly how their own site is doing and almost nothing about how they are described everywhere else, which is the part that predicts citation most. The first move is a map: where your brand is and isn't mentioned across the sources engines read, and how those mentions describe you.
That map is where Resonate Labs starts. We find the gaps between how you want to be described and how the web actually describes you, then do the earned-media work that closes them, the coverage, the community presence, and the reference and video footprint that make an engine more likely to reach for you. The specific outlets and threads are mapped per brand, because they depend on your category and your buyers. If you want to see where your brand presence stands across the sources AI reads, a AI Visibility Snapshot is the place to start.
Frequently asked questions
Do brand mentions matter more than backlinks for AI citation?
For AI citation, yes. In Ahrefs' analysis of 75,000 brands across Google's AI Overviews, branded web mentions were the strongest predictor of AI visibility, at a rank correlation of about 0.664, while backlink count came in near the bottom at 0.218 and Domain Rating explained almost none of the variance. Backlinks still feed the signals underneath everything, but how often and how specifically your brand is talked about predicts citation far better than how many links point at it.
How do you earn brand mentions that AI engines actually read?
By earning accurate, specific coverage across the sources engines draw on: industry press and analyst write-ups, the community threads engines treat as firsthand testimony, video, and authoritative reference sources. The principle is the same across all of them, be mentioned often, described the way you want to be described, in places a model trusts. Which specific publications and threads matter for a given brand depends on its category and where its buyers research, which is the part that gets mapped per engagement.
Does my brand need a Wikipedia page to get cited by AI?
No, but reference presence accelerates citation. In one client study, brands with a Wikipedia article reached their first ChatGPT citation in about 28 days on average, against about 52 days without one. Treat it as an accelerant, not a requirement, and earn it the legitimate way by being genuinely notable and well-sourced. Fabricated or thin entries get removed and can do more harm than good.
Isn't earned media for GEO just digital PR rebranded?
It overlaps with digital PR, and that's the honest framing. The difference is the target and the scorecard. Traditional PR aims at human impressions and brand awareness; earned media for GEO aims at the specific sources AI engines cite, and it is measured by whether your AI visibility improves, not by reach or coverage counts. The motions look familiar, but the targeting and the measurement are built around how engines assemble answers.
Next step
See how AI describes you, everywhere.
Start with a free AI Visibility Snapshot for a no-commitment read on where you stand. The full Visibility Audit shows where your brand is and isn't mentioned across the sources engines read and how those mentions describe you, scored across every engine:
- Where your brand shows up across the sources AI engines cite, and where it's missing
- How the web describes you today, against how you want to be described
- Which tier, technical, on-page, or earned media, is actually holding your citation back