Beyond backlinks: How AI is changing Digital PR

Over the past decade, digital PR has become a key driver of SEO performance. Backlinks, domain authority, coverage volume and reach have formed the backbone of how success is measured and for good reason. These metrics have historically correlated strongly with organic growth. 

But search is changing in a fundamental way and that shift is beginning to expose the limits of traditional PR measurement. 

What has changed?

Google’s AI Overviews have now been rolled out across more than 200 countries and territories in over 40 languages. Generative AI tools such as ChatGPT and Gemini are no longer experimental interfaces; they are embedded into how people research products, compare brands and explore information. At the same time, research from McKinsey indicates that around half of consumers already use AI-powered tools as part of their decision-making process and that the proportion of AI-centric search interactions continues to grow over time.  

Even for those sceptical about generative tools, there is a deeper point to make here. Google’s traditional search algorithm has not been based on simple keyword matching for years. Modern search uses natural language processing, with models such as BERT forming part of how Google interprets query context and meaning. In this sense, Google has moved from a purely lexical model of search to a semantic model, where understanding context and relationships between words matters as much as literal matches.  

Understanding natural language and context is no longer optional. It is now a core part of how search engines work at scale. 

When both infrastructure and behaviour change at this scale visibility changes with it. It is no longer just about where you rank. It is increasingly about how AI systems interpret and represent your brand. 

From ranking pages to representing meaning

Traditional search engines ranked documents. If you built authority and relevance you climbed the results page. 

Generative systems work differently. Instead of simply ranking content, they interpret patterns across the web and synthesise answers. When a user asks a question the system evaluates how concepts relate to each other, which brands are consistently associated with certain topics and which sources appear in credible contexts. 

This process relies on embeddings. 

An embedding is a numerical representation of meaning. When AI models process text they convert language into vectors in a high-dimensional space where semantic similarity can be calculated. 

While that may sound technical, the strategic implication is simple. When your brand is mentioned online it is not just contributing to a link profile. It is contributing to a semantic representation within AI systems that influences how those systems understand what your brand stands for and where it fits. 

Digital PR, intentionally or not, is now shaping those representations every time it earns coverage. 

Brand presence and visibility in generative search 

There is early but meaningful research supporting this shift. 

Ahrefs analysed 75,000 brands to explore which factors correlate most strongly with visibility in Google AI Overviews. One of their key findings was that branded web mentions have a stronger correlation with visibility in AI Overviews than backlinks. This suggests that broader brand presence across the web plays a significant role in generative visibility.  

This does not mean backlinks are irrelevant. It does suggest that how often a brand is discussed and in what context matters to how generative systems surface brands in responses. 

For PR teams this is both validation and opportunity. It reinforces the idea that shaping the narrative around a brand matters more than ever. It also raises the bar. Securing coverage is no longer the end goal. The depth, framing and consistency of that coverage influence how AI systems categorise and surface your brand. 

Why not all mentions carry the same weight 

Anyone working in PR understands that a brand mentioned once in passing is different from a brand positioned as an expert voice within a feature. 

Historically, differences like these have rarely been visible within measurement frameworks. You might instinctively recognise that one piece of coverage feels stronger than another, but most dashboards have not shown this nuance numerically. 

That is beginning to change as AI systems interpret language structurally. A mention in the headline, a direct quote, and proximity to relevant industry terms all contribute to stronger contextual signals. Mentions at the bottom of an article without relevance or context contribute to weaker signals. 

Over time, the accumulation of these contextual differences shapes your brand’s semantic footprint. This footprint determines how likely you are to be surfaced when AI systems respond to queries related to your category. 

Thinking in associations, prominence and context 

Embeddings are relational. AI models learn meaning through co-occurrence and contextual prominence. This means digital PR strategy should not just ask “where can we land coverage?” but also “is our brand important enough within that coverage to be recognised as relevant to the topic?” 

A weak, incidental reference may technically mention a brand, but if the brand is not central to the narrative it makes little contribution to the embedding that AI systems build. Conversely, consistent coverage where the brand is positioned as relevant to key conversations strengthens semantic associations. 

For example, if a fintech brand is consistently mentioned alongside financial education, consumer behaviour and expert insight, and is positioned as authoritative within that context, it builds a clearer semantic position than mentions scattered across unrelated topics. 

This is where digital PR begins to intersect more closely with data science. The objective is not simply visibility but clarity and centrality of positioning within a measurable semantic space. 

Expanding how we measure PR impact 

The PR industry is beginning to recognise that measurement needs to evolve. Research such as Muck Rack’s State of PR Measurement highlights that while measurement is considered critical many teams still struggle to quantify impact beyond surface metrics. 

AI-driven search raises the stakes further. Not just generative systems but any algorithm that uses natural language understanding interprets context, not just links. 

Backlinks and authority metrics remain valuable for traditional ranking systems. However, they do not fully capture how a brand is interpreted within AI-driven environments. 

To remain competitive brands need to complement traditional KPIs with deeper signals such as: 

• How centrally the brand is positioned within coverage
• Which entities and topics most frequently co-occur
• The consistency of narrative framing
• Alignment with priority themes 

These signals are measurable through natural language processing and embedding analysis, providing a richer picture of how coverage contributes to brand perception in contextual search environments. 

How Connective3 approaches this shift 

At Connective3 we see this as an evolution of digital PR rather than a disruption. 

Our starting point is simple: coverage is language data. Instead of treating placements as static URLs we capture full article text and analyse it using NLP techniques to extract structured signals such as entity salience, sentiment and contextual prominence. 

This forms the basis of our Brand Citation Score, which focuses on understanding how strongly and how centrally a brand is represented within earned media. 

By analysing coverage at this level we move beyond reporting what was achieved and begin evaluating how meaning is constructed around a brand. In a world where AI systems interpret the web through embeddings this layer of analysis becomes increasingly important. 

The bigger picture 

Digital PR is not becoming less relevant in the age of AI. If anything it is becoming more foundational. 

AI systems learn from the web’s language layer. Digital PR shapes that layer every day. The difference now is that the outputs of PR activity feed directly into machine-understood representations of meaning. 

The brands that succeed in this environment will be those that think strategically about the semantic footprint they are building, not just the links they are earning. 

The question is no longer simply where your brand is mentioned. It is how your brand is understood.

Ready to get sarted?

Get the data you need to protect your coverage (and ROI) today.