Ten years ago, Search Engine Optimisation (SEO) was seen as the key for businesses who wanted to be 'found' on the internet. A whole industry grew up to generate keywords and content that would magically make websites rank higher on Google's organic search.

In fairness, SEO was always more complicated than some of its more optimistic exponents tried to maintain. There was, in fact, no 'magic button', or shortcut that would do all the work for you. The only way to move up the ranking was to produce content that your readers were more likely to find useful, as measured by the time they spent on your site.

This approach has, however, changed significantly since the advent of AI-based internet search. Many searchers now get the information they need without moving beyond the search engine. This page explains more about how AI-based search acquires and presents information—and what this means for those searching and wanting to be found.

What is AI-Based Search?

It is worth explaining what is meant by AI-based internet search and AI search engines.

There is as yet no formally-agreed definition, but those being used in blogs are surprisingly similar.

There are two main aspects to AI-based search:

• The search engine (for example, Google's Gemini is powered by Google Search);

• The engine that presents the information from the search, usually a large language model (and for more about how these work, see our pages on Understanding Large Language Models and Using Large Language Models).

The search engines themselves have not changed very much. Google Search has been evolving for years, as the company sought ways to make search results more useful to users, and this process continues to unfold.

The big difference is in how searches are interpreted, and how results are presented to users.

You may have noticed that when you carry out a Google search, there is a panel at the top headed 'AI overview'. This is Google's AI-based search result. Gemini, Google's AI engine, is used to summarise the findings from the search, and present them in a user-friendly way. It also invites follow-up questions, which can be phrased in a much more conversational way than was possible with conventional search engines.

The Effect of AI-Based Search

The real question is not so much what is AI-based search, but what effect has it had on how we behave?

Crucially, it has changed the way that we interact with search results.

For the majority of people, the AI overview now provides most of the information we need. One study found that a massive 60% of searches do not result in a click through to a website—because all the required information was in the overview.

In other words, Search Engine Optimisation is pretty much dead. There is no point in ranking for clicks, if there are no clicks to be had.

Instead, websites are now looking to be cited in the AI overview, an approach sometimes called Generative Engine Optimisation.

This requires a very different approach to content and visibility online. What's more, the approach also needs to be different for individuals (who wish to be visible for content marketing and thought leadership purposes) and companies (for e-commerce or other business purposes).

AI's Priorities for Search Results

The key to Generative Engine Optimisation is to understand the priorities of AI-based search engines: what do they consider when deciding which sites to cite?

There are, inevitably, many companies trying to work this out, some more scientifically than others. However, a consensus is emerging that seems to be backed by at least some science.

Fundamentally, AI appears to favour information that is credible, trustworthy, relevant and authoritative. The acronym of choice is now EEAT, for Experience, Expertise, Authoritativeness and Trustworthiness, especially for Gemini. This is because this type of information lowers the risk of the search engine being 'wrong'.

But how do AI search engines decide what counts as credible, trustworthy and authoritative?

The quick answer is that it varies by search engine—but there are some commonalities that suggest strategies for improving your chances of being cited.

1. AI Search Engines Favour Credible, Known Sources

   Several of them use some kind of tiered approach. Their preferred sources are peer-reviewed academic journals, industry analysts such as Gartner, Forrester, and McKinsey, government bodies or organisations, and similar authoritative sources. However, in the absence of information from these sources, they will use credible industry-specific journalistic publications or mainstream media sources and press releases, and prefer these to company websites. There is also a ranking of social media, with LinkedIn, Quora and Reddit seen as more authoritative than X and Facebook.

   What does this mean for getting noticed?
   It means you have to earn your coverage. Companies or individuals who are deemed worthy of discussion, media coverage, or publication in peer reviewed journals are more likely to be considered citable sources.

   It is also good to cite your sources, because this verifies the information and builds trust.

2. There Is a Preference for Recent Content

   There is nothing new about this: search engines have always prioritised recent content in preference to older content that is more likely to be out-of-date. However, that now seems to be even more important, and also includes earned coverage in the press.

   What does this mean for getting noticed?
   You have to publish regularly, on the right platforms and in the right ways—and your content needs to attract attention. You also need to keep attracting the attention of industry commentators and analysts, to keep yourself in the search engines' 'eye'.

3. AI Search Engines Like Content That Is Easy to Extract

   AI search engines have to pull together a summary that answers the search question. That means that they favour content that is easy to extract, such as FAQs, bullet points and tables, where the information is clearly laid out. They also seem to favour clear headings, short paragraphs and plain language for the same reason.

   AI search engines also favour content that is still relevant out of context: where you can 'lift' a paragraph without changing its meaning.

   What does this mean for getting noticed?
   You need to produce and publish content that is easy to understand. Each paragraph should be able to stand alone, and a quick glance at the page should immediately show you the main points. There is limited tolerance for waffle or hype.

4. Consistency Is Key

   AI search engines check for consistent patterns across different sources. It is therefore important to be giving the same messages across platforms, types of content, and different forms of media. Unlike traditional search, this means that you can use the same wording without worrying about being penalised.

   What does this mean for getting noticed?
   You need to focus on two things: consistency, and amplifying your content across platforms. Unlike traditional search, you won't be penalised for sharing the same content in different places.

5. Some Industries Experience Tougher Bars Than Others

   Most AI search engines have a hierarchy of industries as well as information sources. It is much harder to be cited in industries that have life-changing consequences, such as healthcare or finance.

   What does this mean for getting noticed?
   Not being cited might not be anything to do with you—and there might not be much you can do about it.

---

A Final Thought

Like traditional search, AI-based search is constantly in a state of flux.

Search engine companies are constantly looking for new ways to ensure that their content is the most useful, to ensure that they become or remain the search engine of choice.

The precise requirements for being cited by AI search engines may therefore change and evolve over time. However, the principles set out on this page are likely to endure because they focus on what is most useful to users—and that has always been the focus for search.

---