Digital PR and AI visibility: What 3,500 links tell us about brand visibility in AI

When an LLM processes an article, it condenses the content into an answer or summary, identifying the information, brands and topics most relevant to the story. This means a brand can appear within an article, and even have a backlink, without being important enough to feature in the AI-generated response.

A brand that’s mentioned once towards the bottom of an article, for example, is playing a very different role to one whose research forms the basis of the story or whose spokesperson is quoted throughout. Traditionally, both placements could be reported as one link, despite creating very different signals around the brand.

If earned media is helping LLMs understand brands and the topics they’re associated with, we need to consider what our coverage is communicating, rather than looking at the link alone.

Brand Citation Score is our in-house tool for measuring how prominently and meaningfully a brand features within earned media. Using Natural Language Processing (NLP), it analyses the full text of a placement and gives it a score between 0 and 100.

BCS looks at a combination of signals, including brand prominence, entity salience, contextual relevance and narrative framing. Our research shows that the stronger these signals are, the more likely a brand is to be retained when the content is summarised by AI.

There isn’t one single factor that creates a strong brand citation. Instead, we look at a combination of signals that show how important the brand is to the content and what it’s being associated with.

Brand prominence

Where and how the brand appears within an article matters. Headline mentions, direct quotes and expert contributions give a brand greater prominence than a single linked mention further down the page.

Semantic signal strength

The topics and entities appearing alongside a brand help build a picture of what it’s known for. Looking at these associations across earned media helps us understand whether a brand is consistently being connected with the topics it wants to be known for.

Context and narrative relevance

Research, data and expert commentary can make a brand more central to a story. A good test is to ask whether the story would still work without the brand – if its insight or expertise is fundamental to the piece, that creates a much stronger brand signal.

Links are still an important part of digital PR and SEO, but our research suggests they don’t give us the full picture of how earned media contributes to brand visibility.

Alongside looking at the quantity and quality of coverage, PR teams should consider whether the brand is central to the story, whether it’s being associated with the right topics and whether its expertise is clearly attributed.

This doesn’t mean changing your entire digital PR strategy to optimise for LLMs. Many of the things that create strong brand signals are already fundamental to good digital PR: relevant campaigns, original research, expert commentary and giving journalists a genuine reason to feature the brand.

Our study analysed more than 3,500 verifiable digital PR links across 170 brands, 224 campaigns and more than 2,300 publisher domains. Every link was run through LLMs to test whether the brand was retained during AI summarisation and analysed using our Brand Citation Score to measure the strength of its brand signals.

Our AI Visibility Reporting goes beyond traditional digital PR metrics to understand how earned media is contributing to your visibility across AI and generative search. From brand prominence and semantic signals to LLM citation likelihood and share of voice, we can show you where your brand is visible and where there are opportunities to improve. Contact us today for your free audit.