There is no denying that Artificial Intelligence is a revolution in its own right. We already stranded far away from viewing AI only as a quirky tool used to create images on DALL.E, and it has become ingrained in our lives. Indeed, the process of mass adoption of AI into almost every aspect of our life, and on almost every device we have has been happening exponentially. As such it is no surprise that more and more people are turning to these tools for everyday purposes, including but not limited to using AI Tools as Search Engines.
According to a Semrush x Statista report, one in ten U.S internet users are now using Generative AI as their first option for online search, and this number is only projected to rise every year. With the introduction of more and more models dedicated to Generative Search, such as Chat GPT’s recent Search model, Google’s Gemini, and others, it is clear that this is a trend that is there to stay, and companies therefore must adapt to stay relevant in this new AI space.
While standard SEO continues to be extremely important and relevant, and also impacted by AI (you can read more about it in Sven’s article on the topic), it is important to focus on how to be visible and how to be relevant for these new generative search engines. It is a challenge for tomorrow that companies need to take on right now in order to become future proof and not fall behind competition. And it is in our mission here at Semetis to help guide our clients through these exciting and uncertain times. This is why in this article we are going to go over how Generative Search works, and the different strategies a company can use to hone and sharpen their Generative Engine Optimization (GEO).
Understanding how Generative Search Engines Work
Generative Search Engines works by taking the user’s query, reformulating it, passing it through a search engine and summarizing the results into a tailored answer for the user. Therefore the goal is not to appear at the top of a search engine results page but to become more appealing to the AI and more visible in the answer provided.
The end result being different, it is only natural that the strategies to improve the GEO of a website are different than the ones for SEO. It is important to keep in mind that, much like traditional SEO, the effectiveness of these optimizations will depend on the category of the content or website. Out of all the different methods tested to improve the visibility of websites in LLM-based Generative Search Engine, two major categories were highlighted.
Content is the name of the game
Research has shown that including external, verifiable information through external sources will improve a website’s visibility in the responses of Generative Search Engine. Indeed, data-driven evidence and citations are shown to improve the visibility of the content, possibly through enhancing the credibility of the response for the Generative Search Engine, and in turn enhancing the credibility of the sources presented to the end user. Multiple methods are available to introduce more data-based knowledge in the website’s text, and, as mentioned before, the impact of these methods will depend on the category of the website.
An easy first step that proves to have a significant impact on the visibility of your content on Generative Engines is to include clear sources in your articles and content. To have in text, clear, and referenced sources for the information you provide on your website is a massive improvement for the credibility of the content provided to the LLM-based Search Engines. A continuation of this is to not only include the sources but also include statistics to support your statements. This works especially well in pages that are talking about technical topics such as law or policy, but also around topics that are points of debate, with more diverse opinions that can be discussed or argued against. The more substantial sources and data you can add to your point makes it that much more credible to the LLM search engines, and in turn to the user.
A final point of optimization that is in the same vein is to add quotations of authoritative figures that support the claims, arguments, and statements made in the text of the page. All in all these methods of optimizations all have the same goal, to improve the “credibility score” awarded to the specific page from the Generative Search Engine, and in turn improve the likelihood of the page to be used in an answer provided for a relevant query. The LLM will be able to provide content that stands out due to having specific quotations and statistics that support the text, while being correctly referenced as sources.
It’s also about the looks
A different method of Generative Engine Optimization for websites is not only to look at the data and content of the text provided, but also in the way it is written. Indeed it seems that, according to research, the way the content is written is as important for the LLMs in terms of visibility ranking as the credibility in terms of data and sources. One method that scores well with content related to Business, Science, and Health is to have a text with a high fluency, in order to facilitate the reading of more difficult to understand topics. Having text that flows well sentence after sentence, making for an easy read will be favoured compared to dense and technical blocks of text. This will help the LLM-based Search Engines to provide readable content to the users by having easy to summarize content.
Another method is to use an authoritative voice in the text, to make the LLM-based search engines understand that the author knows what they are talking about. The authoritative language impacts the credibility of the source material to the Generative Search Engine, especially on content based around opinions and topics with different points of views, and makes it more visible to the end user. Putting an emphasis on the facts and figures, using assertive language when stating relevant information, will help the content shine through the LLM-based search engine system of recommendation.
What about the classic SEO methods?
Research has shown that traditional methods of standard SEO optimization such as keyword stuffing are not working well with the new Generative Search Engines and their AI models and do not boost performances. In reality it is quite the opposite. When these standard SEO methods are used on websites, the before/after is striking as the pages are in most cases underperforming in terms of visibility on Generative Search Engines compared to their normal versions. As such, and because it is a new playing field, it is important for companies and business owners to adapt to the new rules of the game. However, while these standard SEO methods might not be ideal to implement when working with the new Generative Search Engines, it does not mean that we have to get rid of them, as traditional search engines are of course still standing and still strong.
Conclusion
The optimization of the written content seems to be for now the number one thing to focus on in order to appeal to the AI models’ algorithm for ranking and credibility. It is therefore a question of implementing the new GEO optimization principles while still keeping in mind the standard SEO practices that impact the content created on clients’ and companies’ websites. For now, companies and business owners will have to be able to mix the two different approaches to having optimized content, and balancing that effectively while adapting to the never ending improvements of the AI models will be one major challenge for the foreseeable future. As a digital and innovative agency, Semetis will be there to support its clients in that period of change to find the right balance as efficiently as possible and leverage AI to its fullest potential.
Elio Civalleri
| LinkedinThis email address is being protected from spambots. You need JavaScript enabled to view it.
