Finding the articles where unlinked brand mentions pay off most in AI visibility

Most B2B marketers who start working with AI search engines begin with their own website. That makes sense, because that is where they have control. But the biggest lever for your AI visibility is not on your own domain: it sits in the external articles, comparisons and analyses that AI models consult when someone asks a question in your category. Get named there and the model cites you. Stay unnamed there and you simply do not exist for ChatGPT or Perplexity. That is where unlinked brand mentions do their real work.

The problem is that you cannot be mentioned everywhere at once. There are thousands of pages that are theoretically relevant, but only a fraction of them genuinely influence what AI says about your brand. In this article I show how to find exactly those high-value sources, how to prioritise them and how to get into them through digital PR, outreach and editorial contributions. No buckshot, but a targeted approach that sends your limited time and budget to the places where the difference is made.

Test your AI visibility: score your page with our GEO check.

Why unlinked brand mentions outweigh your own content

Language models do not learn from a link graph the way classic search engines do, but from raw text. When independent sources write about your brand consistently, in editorial articles, analyst reports, comparisons and review platforms, the model learns that your brand is a known, credible player worth naming. That is a fundamentally different mechanism from placing a backlink.

The numbers back that up. Ahrefs analysed 75,000 brands and found that brand mentions on the web show the strongest correlation with visibility in AI Overviews: a coefficient of 0.664, against just 0.218 for backlinks. Brands in the top quartile for web mentions averaged 169 AI mentions, against 14 for the quartile below. In other words: being talked about weighs three times as heavily as being linked to. That is exactly why hunting for the right articles pays off so well. I work this principle out more broadly in brand mentions over backlinks.

Step 1: identify the sources AI cites in your category

Before you approach anyone, you need to know which pages actually matter. AI models do not rely on a random list, but on a fairly fixed set of sources per category. There are three complementary ways to map them.

First: ask the models themselves. Put the buying questions your customers ask to ChatGPT, Perplexity and Google AI (“best [category] supplier for B2B”, “compare [product]”, “alternatives to [competitor]”) and note which sources get cited. Perplexity and Google AI show their sources explicitly, which makes this easy.

Second: analyse your competitors’ citations. Which articles get named when an AI recommends your best-performing competitor? Those pages are your target list. I cover this method in detail in citation mining.

Third: recognise the archetypes models structurally prefer. Research shows clear patterns in which types of sources get cited:

Those patterns are no coincidence. On ChatGPT, Wikipedia accounts for 47.9% of citations, while on Perplexity, Reddit dominates with 46.7% of top citations. Important: only 11% of domains are cited by both ChatGPT and Perplexity, so your target list differs per platform.

Step 2: prioritise your target list

A long list is not yet a plan. You have to rank the pages by expected impact, because not every mention is worth the same. A placement in an authoritative trade title or on a comparison platform that AI structurally cites outweighs ten mentions on weak sites. AI models implicitly apply a notion of authority that resembles E-E-A-T: a publication in a respected outlet carries more weight than an anonymous forum post.

I score every source on three axes:

Sources that score high on all three are your quick wins: that is where you invest first. Sources with high impact but low feasibility (think of a Gartner report) go on a longer horizon. Pages that are barely cited, you leave alone for now, however tempting an easy mention may be. The Princeton and Georgia Tech research into GEO also showed that the content itself matters: content with statistics and quotes from credible sources is demonstrably cited more often. Keep that in mind when assessing a target page, because a well-structured article you fit into performs better than a thin list.

Step 3: get in through digital PR and outreach

Now that you know which pages count, everything hinges on getting in. This is where GEO differs fundamentally from classic link building: the goal is an editorial mention, not a bought link. According to an analysis by Muck Rack, earned media is said to be responsible for the vast majority of all AI citations, which underlines that journalistic coverage is the shortest route to AI visibility.

The most effective tactics for B2B:

• Digital PR with data and expert insights. Offer journalists at your target outlets original research, numbers from your own practice or a strong opinion. This is the approach I work out in digital PR and link building.
• Journalist platforms. Services such as the HARO successors and Qwoted let you respond quickly to journalist queries at authoritative publishers. Low effort, high return when a placement lands.
• Contributions to comparison and review pages. Make sure your brand appears fully and correctly on G2, Capterra and sector comparison sites. Collect reviews actively, because those pages are exactly the sources AI consults.
• Guest contributions in trade press. A well-argued expert article in a sector title makes you visible not only to readers, but also to the models that cite that title.

The common thread: become the source instead of waiting for someone to name you. How to package your own expertise so media and models pick you up, you can read in becoming a source for AI.

Step 4: measure, learn and repeat

GEO is not a one-off campaign but a cycle. After every placement, check whether your AI visibility actually improves: put your standard questions to the models again and see whether you are now named more often, and in better context. Remember that freshness counts: research shows a large share of AI citations goes to content from the past year, so older mentions lose force. How to assess that context objectively, whether a model frames you positively, neutrally or negatively, you can read in measuring the sentiment of AI brand mentions.

Work in quarters. Every quarter you revise your target list, because the sources AI cites shift as models update. A source that dominates today may weigh less in six months. By comparing your baseline with your current position, you see in black and white which placements paid off and where you need to invest again.

Frequently asked questions

Need help?

Want to translate this into execution? See how we approach this with AI visibility.