Why earned media matters more than ever in the age of AI
Hard Numbers’ Paul Stollery on how AI is changing the communications industry - the Socially Mobile autumn 2026 alumni lecture.
The autumn Socially Mobile alumni lecture with Paul Stollery, co-founder of data-driven PR agency Hard Numbers, was a fascinating take on where we are with AI in the communications and public relations industry - and what’s to come next.
Paul gave an overview of how AI is changing PR and communications, a shift I personally find equally exciting and terrifying.
The rise of Generative Engine Optimisation
The biggest shift already underway is Generative Engine Optimisation (GEO) - how brands are working to influence what AI tools say about them, and why this is becoming as important as traditional media and search.
For Hard Numbers, GEO has been their biggest growth driver over the past two years, and it’s easy to see why. AI is transforming brand visibility and reputation.
What the research shows
A research study commissioned by Hard Numbers, Reputation in the Age of AI, found that investing in PR and editorial media continues to shape reputation, and is perhaps becoming more important than ever.
Key takeaways:
Earned media accounted for 61% of the sources cited by ChatGPT
Owned media accounted for 44%
Social media accounted for less than 1%
A second study, Category Recommendations in the Age of AI, found a direct link between earned media performance and brand recommendations in LLM-driven platforms such as ChatGPT and Gemini.
Key takeaways:
Brands that LLMs perceive to have more influence (through earned media citations, for example) are recommended up to 80% of the time compared with less than 50% for those with lower media influence.
Brands with higher positive media sentiment are three times more likely to win head-to-head comparisons against competitors.
The findings make a compelling argument for investing in GEO.
Gullible engines
However, Hard Numbers’ most recent research report, published in June 2026, raises some serious concerns. Gullible Engine Optimisation shows how brands can manipulate AI results to recommend themselves - and even, potentially, shape the answers relating to their competitors - highlighting how gullible LLMs can be.
For a case in point, see Hard Numbers’ “best-dressed UK PR practitioner award” press release which successfully tricked ChatGPT, Gemini and Claude. A bit of fun, yes, but also evidence of how easily AI engines can be manipulated.
“Because AI engines repeat what's published with no human oversight, an unverified page can become a trusted source - and the same method could plant false claims about a company's reputation, a product's safety or a political issue,” said Paul.
AI in agency workflows
Paul then explored what's still to come: how AI is being deployed across agency workflows, from research and media relations to creative and measurement.
As practitioners, we all have a responsibility to continue learning about AI to keep pace with the rapid advancements in technology. AI is already being rolled out across agencies to help with various tasks: from research and analysis to content creation and copywriting.
However, when AI is used for day-to-day press release and pitch writing, there is an obvious reputational risk. In an age when misinformation and ‘fake news’ is prevalent, especially on social media, factual accuracy is a huge concern. Many journalists are quick to call out ‘AI slop’, so it’s essential that human oversight remains at the core of media relations.
In the UK, there is also a trust problem. Only 36% of the UK population trust AI, compared with 87% in China, according to the Edelman Trust Barometer flash poll on AI. Adoption tends to follow trust.
Where to start
So, what should agencies and practitioners be doing as a starting point when it comes to GEO?
Paul recommended auditing your brand regularly across LLMs to benchmark GEO visibility, test prompts, analyse citations and then using these findings to refine your approach.
The message for Socially Mobile alumni was clear. Keep learning, keep checking, and keep people at the centre of the work.
Further reading: AI for Public Relations: A How-To Guide for Implementation and Management by Stephen Waddington and Ben Verinder. All author proceeds go to Socially Mobile.
Paul Stollery is co-founder and creative director of Hard Numbers, a data-driven public relations agency, where he leads the agency's work on Generative Engine Optimisation and the use of AI in communications.