Stop quoting the research. Start influencing the results. 

For the last 18 months, the PR industry has been enthusiastically repeating a set of statistics about AI search and earned media. 

At Hard Numbers, our own research identified back in October 2024 that earned media was cited in 61% of ChatGPT responses relating to the reputation of the world’s 100 biggest brands. Nine months later, research from Muck Rack produced an even larger number, albeit using a broader definition of “earned media” that included sources like “corporate blogs”, “social media/user generated content” and aggregator sites. 

It was genuinely exciting to see our research shared so widely across the industry. The idea that PR and earned media influence AI-generated search results clearly struck a nerve. 

Likewise, Gartner’s widely shared prediction that “by 2027, mass adoption of LLMs as a replacement for traditional search will drive a two times increase in PR budgets” became catnip for conference presentations, LinkedIn posts and thought pieces. 

If I had a pound for every time I’ve seen that Gartner slide quoted, I’d almost have enough money to buy a large pack of Lurpak from Tesco Metro. Which, admittedly, is now roughly equivalent to the latest funding round received by ChatGPT. 

The problem is that too much of the industry is still dining out on these statistics rather than doing the actual work required to influence, track and measure performance within AI search platforms — or generative engine optimisation (GEO), as it’s known. 

Frankly, I’m sceptical about the Gartner prediction. There’s currently limited evidence that LLMs will replace traditional search behaviour at the scale implied, certainly within the next two years. Google processes around five trillion searches annually from more than four billion users. That’s not a market position likely to disappear overnight. That said, the Gartner data is directionally correct. 

Earned media does influence AI search results. Reputation shapes visibility inside generative platforms. And PR professionals are uniquely positioned to influence the sources AI systems trust, cite and surface. The opportunity now is to move beyond awareness and into execution. 

That means identifying where your organisation is – and isn’t – appearing within AI-generated responses around your sector, products and services. It means improving on-site content, securing authoritative off-site citations, and understanding which publications influence AI visibility. Most importantly, it means measurement. 

If you want to build the case for greater investment in PR, you need evidence. Data from Google Analytics, Google Search Console, SEMrush, Peec AI and other tools demonstrating changes in visibility, citations, referral traffic and competitive positioning. 

At Hard Numbers, we’ve been helping clients do that for the last two years. We’ve trained hundreds of in-house professionals on the fundamentals of generative engine optimisation following our widely shared GEO for PR Primer. The opportunity for PR is real. The time for simply repeating the statistics has passed. Now it’s time to influence the results and show it with, ahem, hard numbers.