Generative Engine Optimisation for Press Releases

We all know Generative AI is changing the world around us. It’s changing how we work and live. One of the most important transformations AI tools have unleashed is in changing how we find information. This makes Generative Engine Optimisation (GEO) more important than ever.

The impact of AI on search tools

According to a recent webinar from PRmoment, Jamie Brader states that around 18-24 months ago only about 1% of search activity was happening through AI tools but predictions for the near future forecast that most search activity will happen within Large Language Models (LLMs).

AI is impacting search activity in other ways too. In the US Google searches per U.S. user are down nearly 20%, with the data suggesting Google isn’t losing users instead it’s losing repeat searches. This means users are now becoming less likely to “refine” their searches.

In Google's AI search mode 93% of queries exhibit zero-click behaviour meaning users are finding what they want without clicking a website link. Your ranking number in traditional search results no longer guarantees you visibility. If AI doesn’t cite you, for many people you no longer exist.      

Ensuring visibility

We’ve established that GEO is important, but this leads us to a question - how do we improve the chances that press releases are discovered and cited by AI tools?

Firstly, lets start with the good news - LLMs love press releases. LLMs love structured, factual and clearly attributed information. They love to know the details of the who, the what, the why, the where and the when. The predictability of press releases is something that LLMs can get behind.

One tip is that retrievability is more important than reach. If you want your press release to be discoverable then you need to know when your press release is visible or invisible to AI. Paywalled content, PDFs and changing URLs are all invisible to AI, so focus on open-access content, newswires, structured crawlable pages and consistent entity mentions.

LinkedIn is a great example as LinkedIn articles have stable URLs. The emphasis here is on LinkedIn articles not LinkedIn posts, as posts don’t have stable URLs. Targeting the right media for your brand is not enough, you need to target the right media for LLMs.

A third tip is that consistency matters. Consistent messaging should be repeated across owned, earned and social media channels. Sometimes AI tools can have long memories. The most cited URLs can spread back across the past and this is especially common when asking more detailed questions.

In contrast, recent research from Gartner has shown that when AI searches imply recency almost half of the citations are news articles. This demonstrates the need for proactive earned media efforts and ongoing consistent messaging. A multichannel approach is important and being active on social channels is a must.

As mentioned, LinkedIn articles are a great way to achieve the social aspect of your multichannel approach. Keeping your storytelling consistent across these multiple channels can ensure your press releases are more readily discovered by LLMs.

The human and LLM reader are more similar than you might think

Something else to keep in mind is that you shouldn’t be writing one version of the press release for a human audience and one for LLMs. You often have multiple audiences in mind when writing a press release and optimising your press release for LLMs is just another one to add to the list.

Also, optimising your press release for an LLM audience is very similar to making it more engaging for a human audience too. As you need to engage with a human audience to get your press release cited in the first place, optimising your content for LLMs can achieve two goals at once.

Some more tips to structure your press release for LLM’s include:

• Concise, factual headlines that help to increase visibility by LLMs.

• Bullet points that can be important not just for your AI audience but also for a human audience by highlighting key information and data.

• Bolding can be helpful too. For example, bolded section headers can be a great way to guide a LLM (or a human) through your content.

• Including pull quotes is another important way to highlight the main points of your press release in a way that makes it more readable for LLMs and humans.
  

A great example of an effective summary or key highlights section comes from this Reuters article.

Reuters have ensured that their summary section is bullet pointed and provides context, both great for LLM readability.

AI might be changing the way we consume information, but by adjusting your approach you can ensure that your press releases are seen by LLMs and humans. The fundamentals of a great press release haven’t changed, and it turns out that AI tools reward them too.

Thinking carefully about retrievability, consistency and clear formatting can help to future-proof your communications strategy for our brave new AI world. With GEO you don’t have to start from scratch, you just need to enhance what you already know how to do.

With thanks to PRmoment and Cision for inspiration taken from their recent webinars.

Tom Dockings is a PR Intern at Fair Communications. For more information visit: faircommunications.co.uk