AI platforms are changing how people find information about products, services and brands – including yours. Here’s what you need to know about GEO and where to start.

By now you’ll have heard the acronym – GEO (generative engine optimisation) is everywhere. The likes of ChatGPT, Claude or Google’s AI Overviews have fast become gatekeepers of information about your sector, your competitors and your organisation. That means the practice of understanding and influencing how AI platforms represent your organisation is non-negotiable for communications professionals.

But in a sea of analysis, tips and future-gazing, where do you start? Here we break down what GEO is, why you need to take note, and how you can get started. 

How people find information has changed 

Not long ago, you typed a query into Google and got a list of links. You clicked on a few and formed your own view. Now, nearly 60% of searches end without a single click. Users ask a question, receive a synthesised answer from an AI system and move on.  

This is not a marginal trend – it reflects a structural change in how people consume information. When that first answer is generated by AI rather than constructed by a human, it matters enormously what the AI has been trained to say, and what it chooses to cite. 

For PR professionals, this creates both a risk and an opportunity. The risk? Information about your organisation may be absent, misrepresented, outdated or described in ways that undermine your positioning. But as PR teams are already producing much of the content that AI systems learn from, they can play a huge role in influencing the outputs.  

What AI systems actually draw on 

To understand GEO, you need a basic picture of how AI answer engines produce their responses. 

When someone asks ChatGPT or Perplexity a question about your organisation – “Is [x] a good employer?”, “Which firms lead on [y]?”, “What do people say about [z]?”  – the AI synthesises an answer from several sources: its training data, live web search results, and the credibility signals attached to different sources. 

Research presented at a recent PRmoment masterclass on GEO found that earned media accounts for around 90% of AI visibility for reputation-style questions. Brands whose coverage is broad, frequent and positive across many sources are significantly more likely to be recommended by AI systems. 

Owned content also plays a role – your website is cited in over 40% of AI responses. But the narrative AI constructs about you tends to be driven by what others have written, not by what you say about yourself.

Other sources that carry weight include Wikipedia and industry databases, LinkedIn articles, sector press and trade media, plus institutional sources such as trade body publications and analyst reports. 

The role of PR 

There is a risk that GEO gets handed to digital or SEO teams because of the perceived technical complexity. But the levers that determine AI visibility are overwhelmingly communications ones – what is written about you and where, your owned content and how it’s structured and how consistent your messaging is across channels.  

PR teams are already shaping the inputs that AI learns from. GEO is the discipline of doing that deliberately, with a clear view of what the outputs look like and what needs to change.  

The first steps 

So where should you start with GEO?  

  1. Prioritise quality earned media coverage. And remember bigger isn’t always better. Trade media often carries more weight than national press for category-level queries.  
  2. Get your owned content in order. A corporate website that is difficult to navigate, light on substance or full of promotional language rather than clear answers will not serve you well in AI responses. AI systems favour content that is structured and directly answers the questions that people actually ask.  
  3. Consistency is key. Make it easy for AI to understand what you do and how you do it by ensuring all your messaging across channels is consistent. If your communications say one thing and your publicly observable behaviour (reviews, announcements, etc) – says another, AI will notice.  
  4. Find out what AI says about your brand by running a structured audit. Test relevant prompts across AI platforms and examine the results for visibility, positioning, sentiment and the sources shaping the answers. From there, the work is familiar – identify where the gaps are, build a communications plan to address them, execute and measure what changes over time.  

firstlight’s signal programme is designed to do all of this and more. So if you want to get your GEO house in order, get in touch.