From Keywords to Conversations: How LLMs Are Rewriting the Rules of SEO

SEO For AI Search

As a result, we’re starting to see a shift in search behaviour as more and more people use LLMs for search. People aren’t just searching anymore, they’re having a conversation. They ask questions, refine, clarify, and keep going, allowing them to gather a huge amount of information much more efficiently than through traditional search.

ChatGPT, and AI Mode have turned queries into ongoing dialogues, and when search becomes conversational, SEO will have to change too. It’s no longer just about ranking for the first query, it’s about staying relevant across that entire conversation. And that’s the biggest opportunity (or challenge!).

 

The Fall of Keywords, the Rise of Semantic Clusters

As keywords become less relevant, and the importance of context grows, there’s a lot of discussion around ‘semantic clusters’: groups of related topics and content that collectively signal depth and authority.

Whilst this all sounds new and fancy in the age of AI, it’s essentially what content SEO has been talking about for years. Content hubs. Hub and spoke, pillar content, and supporting content, whatever you want to call it, it all does the same thing, helping crawlers get a better understanding of your topical relevance.

Both Google and ChatGPT are looking for the same thing and you need to be building out a body of work that proves you deeply understand, and are an authority on the subject.

Writing for Natural Language

With conversation becoming more normal, so does the language we use when we search. This means targeting more natural, question-based structures (who, how, why, what, if) and building supporting articles that anticipate what a user might ask next. When looking at the types of queries that people are searching for that trigger AIOs for example, across our data set, we see roughly ⅓ of all keywords are question based.

Structured data, like FAQ, or how-to schema can also further help provide context, but ultimately it comes down to clarity. LLMs like simple easy to grab information, and the clearer this is laid out the better.