As large language models (LLMs) and generative AI become more integrated into how people find information, the rules for earning attention and building trust are being rewritten.
This shift was the focus of a recent Medianet webinar (check out the recording here) hosted by Managing Director Amrita Sidhu. The session featured AI and PR expert Celia Harding, Founder of LEOPRD, a language engine optimisation (LEO) PR advisory, alongside Mercedes Carrín, Medianet’s Head of Marketing.
Public relations has long been a foundational discipline in communications. Now, thanks to AI models prioritising editorial content over paid media, PR is reclaiming its role as a powerful visibility engine for brands. This moment could mark a turning point, with PR leveraging technology to deliver reliable communication and elevate its standing in the media landscape. Yet it also comes with risks, as misinformation can surface in LLM responses. PR teams must remain vigilant to ensure brand accuracy and protect against being misrepresented.
For decades, search engine optimisation (SEO) has driven brand visibility strategies. Celia Harding, with over 20 years of PR experience, launched LEOPRD to address a critical new challenge: language engine optimisation (LEO).
While SEO focuses on ranking in Google search results, a new reality is emerging as people increasingly rely on curated, AI-generated responses for information.
“Imagine a brand is mentioned in a single AI-generated response and it’s trusted by millions, or it’s effectively erased from the conversation,” Harding explains. “That’s the era that we are facing now.”
LEO or Generative Engine Optimisation (GEO) ensures brands appear in these concise, AI-powered summaries. If your brand is missing from the answer, you’re invisible to the user.
Many wonder if their SEO investments have become obsolete. The experts agree SEO isn’t dead; it must evolve alongside LEO/GEO. Core principles still apply, as large language models often trust the same high-authority websites that rank well in search.
“Search engine optimisation is not dead,” Harding says. “It’s just evolving to meet the needs of these new language engines and how they prioritise content.”
The key shift lies in the goal: from optimising backlinks to securing mentions and presence in AI responses.
Mercedes Carrín highlights this overlap, especially in how content is structured. “There’s a huge overlap between SEO and visibility in LLMs,” she says. “It’s about structuring your content properly, using semantically rich headings and very descriptive language.”
She adds a crucial insight on why impressions increasingly fail to translate into clicks, describing a pattern that's now common across sites:
“We’ve seen at Medianet, and I’ve seen it across the board, the crocodile effect where your impressions are going up, but your clicks are not going up like you would normally see on the traffic.”
That “crocodile effect”, i.e. a widening gap where visibility rises even as click-through rates fall, is becoming a defining feature of AI-powered search landscapes.
“Our content doesn’t always have to deliver clicks,” Carrín adds. “It needs to deliver value, build brand awareness, and drive business outcomes.”
LLMs operate on trust signals; valuing authority, credibility, and relevance. According to Harding’s March 2025 research, earned media such as editorial coverage, awards, and reviews make up a large share of citations these models use.
“These large language models, they are in the business of signals and they’re looking for that authority,” Harding says. “They’re looking for trust.”
She emphasises that offline opportunities, like speaking engagements and podcast interviews, remain vital for building credibility. “The report... shows that it’s not only what you’re doing online, it’s offline that counts,” she adds.
Social media had minimal impact on corporate visibility in the March 2025 study, but personal brands, especially on LinkedIn, proved far more influential. Paid content may gain temporary traction, as LLMs rarely distinguish native advertising from editorial unless specifically prompted.
A key mental shift for PR professionals is treating LLMs as another stakeholder, Harding says. “Just as we build media relationships, we now need to build a presence in LLMs.” This means understanding how each model works, because they’re not all the same.
Harding explains each model has unique priorities:
Both speakers stressed moving beyond vanity metrics like “reach” and “advertising value equivalent (AVE).” Harding recommends mixing traditional SEO tools with AI-specific platforms such as Profound, Scrunch, Trakkr, and Peec to measure visibility. She also suggests analytics to “separate the traffic from large language models versus organic and paid search.”
Optimising content for these “personalities” means tailoring structure, data, and language to match each model’s strengths. As Mercedes Carrín explained, this is no different from how PR teams pitch to journalists.
"Just as we build media relationships, we now need to build presence in LLMs," she said. "That means understanding how each model works, because they're not all the same."
The conversation ended on an optimistic note. Mercedes Carrín believes marketers’ adaptability will serve them well, while Celia Harding calls this the start of the “biggest 10 years ahead” for PR.
Both urged PR teams to start small, but start now.
“If you’re not already asking ChatGPT what it knows about your brand, that’s your first step,” Harding said. “And if you’re not seeing LLM traffic in your analytics, chances are you’re not showing up.”
Carrín added, “Pick one problem to solve. Give your team permission to experiment. Get curious. The worst thing you can do is wait.”
The message is clear: instead of fearing AI, PR professionals should embrace it as a new stakeholder. By sharpening critical thinking, honing storytelling, and prioritising trust-building content, PR leaders can reassert their relevance and drive real results in the age of generative AI.
To support PR teams in this shift, Medianet has rolled out and is launching several AI-powered tools designed for visibility, efficiency, and impact measurement:
A proprietary impact score as part of our media monitoring service, measuring credibility and impact accurately; replacing inflated, inaccurate metrics like reach and ad value equivalent.
📎📎📎
How to get your media release cited by AI tools like ChatGPT, Gemini, and Copilot? Click to read.
Ready to rethink your media monitoring? Talk to our team today. Book a demo here.