Skip to content

What is ‘GEO’ and how does it impact public relations?

Image created in Canva

You are probably already using artificial intelligence (AI) in some context in your PR work. During my time training PR people to use AI, I’ve answered many questions about ‘Generative Engine Optimisation’ (GEO). 

I’ve summarised some of the key questions about GEO in this post.

What is ‘Generative Engine Optimisation’ (GEO)?

GEO, sometimes known as ‘AEO’ (Answer Engine Optimisation), is a term for helping your content stand the best chance of appearing when someone enters a specific, relevant query in the Large Language Model (LLM) that they use (e.g. ChatGPT, Gemini, Perplexity, Claude, etc.).

Its name was inspired by Search Engine Optimisation (SEO), the long-established practice of aiming to understand what the algorithms of search engines like Google or Bing prioritise when surfacing results to users.

The good news for PR people is that earned media (e.g. media coverage) and owned content (your website or LinkedIn articles, for example) are highly regarded by LLMs. PR is so well-positioned to benefit from GEO that Gartner recently observed that PR budgets were likely to increase. 

How does GEO differ from SEO?

Users’ questions on LLMs are far more specific and contextualised, and therefore are likely to address questions further down the sales funnel/customer journey than answers surfaced by search engine results pages (SERPs). (NB – most search engines now have an AI summary box, or ‘no-click’ search).

With GEO, your aim is to earn accurate and relevant citations in AI-generated answers within the most-used LLMs. With SEO, your aim is to rank highly on search engines.

LLMs want detailed, structured content, while search engines reward longer articles and take into account brand mentions. Once upon a time, links were really important, but less so now.

With GEO, you need to measure results not only with visibility and share of voice, but answer accuracy. With SEO, the key is to rank highly and measure traffic and click-through-rates (CTRs). From what I’ve seen and heard, appearing in LLMs is comparatively fickle, and your brand could be in and out of visibility, depending on the context, while it takes longer with SEO for sites to rank, gain authority and rise, then fall if not supported.

Read this research from Profound for more on AI search volatility.

As with SEO, it’s important to factor in branded versus non-branded searches. An easy starter is to ask your LLM for an honest assessment of what it understands about your brand. Factor what comes back into your messaging planning. Next, ask who the key players in your space are and see how your brand compares. You can ask lots of product- or sector-specific questions at this stage. See below for some tools that could help.

What do LLMs want from PR content?

Like search engines, LLMs look for content that is high in ‘EEAT’ content, which stands for Experience, Expertise, Authoritativeness, and Trustworthiness.

You can demonstrate ‘Experience’ with case studies, customer tutorials, and FAQs (like this one). 

‘Expertise’ is more specific and in-depth, so think about how-to guides, whitepapers, step-by-step guides etc. 

What do LLMs class as ‘Authoritative’? Think about third-party endorsements and coverage, so media citations and backlinks to show that other content creators (especially in earned media) rate your brand.

‘Trustworthiness’ is demonstrated with clear sourcing, user reviews, and author biographies. It’s really important to build authority and trust in a specific area, which is why building executive profiles is essential.

Some LLMs have content tie-ins with specific media, so if your goal is to rank on LLMs, rather than solely secure high-profile coverage, consider targeting media that have arrangements with LLMs.

Another thing to consider is that recency matters. Keep your content up to date, and have a solid, useful, LLM-friendly content calendar to stand the best chance of ranking.

What does GEO mean for PR content?

LLMs love structured and contextualised information, so think about how to use listicles in blog posts, FAQs (per above), unique data studies (such as reports and white papers), customer case studies, step-by-step guides, and even press releases are coming into their own, as their structure is exactly what LLMs like.

Deploy a mix of both evergreen content and reactive content for in-the-moment marketing. Over the long term, continue to build a customer case study database for a wide range of service or product features. Build the profiles of your industry experts.

Also, when it comes to copywriting for LLMs, get the important stuff up front and structure pieces to be read by both humans and machines: use subheads, bullets etc.

Where should I start with GEO within my PR agency or in-house team?

It would be a great time to get some basic AI for PR training, which I offer, so everyone knows what they’re dealing with in the world of LLMs. 

I’d also recommend:

  • Start with an audit of branded and unbranded searches on LLMs for 15-20 queries. Save your prompts for later, when you revisit to compare and contrast your progress.
  • Understand prompting techniques. As above, build a database
  • Identify company expert spokespeople
  • Cover the technical basics of your website: data markup, ensure robot.txt files aren’t blocking LLMs, deploy a LLMs.txt file to help LLMs read your site
  • Optimise/structure your content for readability. For example, here, I have used subheads and bullet points with common questions as the heading
  • Create content hubs around relevant content to show your expertise. Create content that’s easy for LLMs to learn from
  • Publish clear, fact-rich, accurate and well-structured information
  • Explore AI tools to help you gauge how and where your brand or your clients appear on LLMs

AI visibility tools include:

Most importantly, be consistent across channels with your messaging. Monitor and iterate your progress and pivot your tactics accordingly. Be aware of the pitfalls of generative AI (they can hallucinate, so always double-check) and be transparent with how you use AI.

This field is fast-evolving, so keep up to date with the last thinking on LLMs. I will update this post when a major change happens. 

Finally, if you or your team needs AI training for PR, please drop me an email.

Disclosure: This piece was written entirely by me, a human! I used Grammarly to check for typos etc.

About me

I’m Chris Lee. I’ve worked in B2B technology PR and journalism since 1998. I have been tracking AI developments in PR since 2018, when I wrote this piece for City AM. Since 2024, I have been training PR teams how to use copywriting in the AI era. 

Feel free to connect with me on LinkedIn here if you are a PR professional.