Search is changing - is your website keeping up?
The unavoidable topic right now is AI. Much speculation has been written about AI "taking our jobs" or "changing the world", even destroying it. They might all be right. 🤷🏻♂️
Something that is not speculation is the extent to which search has already changed, especially since the introduction of AI summaries in Google search results.
When we talk about the impact of "AI" and search, the direct impact of AI tools such as ChatGPT, Perplexity, Claude, Gemini, and Copilot has been quite modest. The change in how Google search results work has had the biggest impact.
Zero-click search
AI summaries have had a massive impact on the relationship between Google Search and websites. We’ve entered an era of "zero-click" searches. This is where a search engine query is answered directly in the search results page, removing the need for the searcher to visit the source.
Further reading: Google users are less likely to click on links when an AI summary appears in the results and Click-through rates (CTR) for top-ranking pages have dropped dramatically
Zero-click searches have been growing for a while. It can be said that this evolution started originally with the introduction by Google of the Knowledge graph in 2012, and took off with the introduction of “Featured snippets” in 2014. Zero-click searches by the end of the 2010s had risen to a level of around 25%.
Google expanded their use of the “Knowledge panel”, providing even more information directly in search results. Zero-click climbs further, up to over 50% during the first half of the 2020s.
Further reading: Zero-Click Search Is Evolving Into Zero-Search Discovery: Here’s Why and Zero-Click Searches And How They Impact Traffic
AI-overviews
After the introduction of AI-overviews, zero-click has reached the heights of around 65% of searches (depending on query type and topic area) and as high as around 80% when an AI-overview is shown. Zero-click is now firmly established as the norm, not the exception.
The next biggest impact has been on the queries themselves. The words that people are placing into search engines. These phrases have become more conversational and increased in detail.
The short, fragmented, and abbreviated queries of times gone by have shifted into natural language questions.
With that shift, expectations have also shifted. People are no longer just adjusting and revising their queries to get better top results; they are phrasing their initial questions in order to get complete answers instantly.
We are heading into an era where AI-agents are doing the searching for us and summarising the answers, and where these agents will also take the next step of visiting a site and completing an action - making a booking, a purchase, or submitting a form.
What to do? Understand the nuances
Currently, the “New World” of SEO can appear to be a confusing bucket of alphabet soup.
AIO. AEO. GEO. LLMO. AAIO. 😱
These are all acronyms that have appeared in recent times. The exact definitions vary, which is understandable. Things are still fluid because of technology developing at such speed.
Grouped together, they describe a layered “stack” of subtly different solutions that require subtly different approaches. We are pivoting from focusing on search engines to information ecosystems.
AIO: AI optimisation.
The overarching strategy to optimise content for artificial intelligence systems. This is the foundation of the stack where SEO in the traditional sense is still what underpins it all - Discoverability is still key.
AEO: Answer Engine Optimisation.
This is where we optimise content so that it gets featured or cited in specific instant-answer functions. Such places like Featured snippets, overviews, and voice queries. Here well-structured, accessible content will help make this easier to happen.
Further reading: What Is Answer Engine Optimization? Answers, Tools, & Proven Tips
GEO: Generative Engine Optimisation.
Focusing on AI-powered search engines to help get your content to be the basis for generated responses. Not just included as a source, but being an influence on the text generated. Here we are talking about not just AI summaries in Google search, but responses in AI tools like Google AI mode, Perplexity, ChatGPT.
Further reading: Generative Engine Optimization (GEO): The Future Of Search Is Here
LLMO: Large language model optimisation.
Here is where content is optimised for the way in which large language models “understand” your content and aim to get it included as a deeper part of the (generated) conversation and long-form content in tools like Gemini, Claude, and ChatGPT. Potentially even as part of training data, making this the "long game" and part of achieving the equivalent of "top of mind" status in the age of AI.
Further reading: The LLMO White Paper: Optimizing Brand Discoverability in Models like ChatGPT, Claude, and Perplexity
AAIO: Agentic AI optimisation.
This is the optimisation of your content, interfaces, and systems so that AI agents can successfully find, understand, and use them. This involves APIs and a clean technical structure to enable AI "agents" to execute decisions and perform tasks without necessarily needing human interaction.
Further reading: What is Agentic AI Optimization?
Next steps
After getting a feeling for the nuances of the various AI search solutions, the next step is to identify your weak spots and define your tactics in relation to each of them that are relevant to you and your organisation.
Over all, good, useful content is still good, useful content. AI solutions haven’t changed that - but your website will often form the ingredients that are baked together and eaten elsewhere, rather than consumed directly on your website.
Visitors will be increasingly AI agents and not humans. This is where clearly structured information and accessibility will be essential. A fully accessible website is not only easier for humans to use, including those using assistive technology, but also machines.
The more accessible your website, digital products, and services are, the more likely an AI agent will be able to successfully complete tasks on behalf of a human.
Quality content, structured information, APIs, and high levels of accessibility will be the pillars of success.