Why SEO is still important in the age of AI Search

The discourse around AI and LLM search is rife in the marketing world with many attempting to get to grips with just how this will impact website traffic and existing marketing strategies. In an ever-evolving landscape we will break down why traditional SEO is still a fundamental element in your marketing strategy, as well as what activities you should incorporate and build-upon to help propel visibility in LLMs and AI models.

The Emergence of AI Overviews and LLM Search

There have been many changes in the way search works over the years, however, the functionality has remained fairly similar as an indexing system matching user queries with relevant sources. 

The shift we are currently seeing in the search landscape could be the biggest yet with AI systems such as AI Overviews and ChatGPT changing how search works.

Generative AI

Generative AI is powered by Large Language Models (LLMs), these are machine learning models that are trained on massive data sets of code and text, such as ChatGPT and Gemini. The core functionality of LLMs is to produce natural language summaries (answers) by processing users' search queries. AI search is largely driven by these LLMs using them to analyse content and develop an understanding of what queries are related so that they can provide relevant and conversational responses.

However it is not as simple as just keyword matching, it functions based on ‘reasoning’ and product or brand attributes. 

LLMs work like an advanced version of predictive text on your phone, they have been trained on large amounts of text so they can recognise patterns in how words and ideas fit together. When a user types their query in, the model breaks it down into smaller pieces (called ‘tokens’). The model looks at these ‘tokens’, matches them to its internal dictionary, and uses its training to predict what word (or ‘token’) should come next. Because of this, AI tools care more about the meaning of your words rather than how many times you repeat or optimise for a keyword. That’s why semantic consistency is key when writing content.

AI Overviews

AI Overviews are a SERP feature that started rolling out in August 2024. They work by actively retrieving fresh information from websites to construct an answer to the users query and deliver those answers at the top of the Google Search Results Pages (SERPs).

It could be seen that AI Overviews works in a similar way to the ‘People Also Asked’ snippet, as it allows users to explore topics and discover relevant websites in a quicker and more easily digestible way. 

AI Overviews works off of Google's ‘query fan-out’ process, which expands a single user query into lots of mini-queries and fires them all at the same time to deliver the response. The purpose of this process is to collate a wider and more diverse set of results that a single query often wouldn't be able to do.

The process works by breaking down the original query to the core intent and semantic components, to gain as much meaning and context. This allows AI Overviews to issue multiple related searches across subtopics and data sources which creates the response, identifying more supporting webpages by accumulating a diverse set of helpful links and answers associated with your query.

The SEO Impact - What we know (and don’t know)

It’s important to state that this is an ever-evolving landscape and what may be true today can be eclipsed by new changes tomorrow. However patterns are beginning to emerge as a result of the introduction of AI Search systems.

AI Overviews & Zero-Click Searches

For brands appearing in AI Overviews, a common result is often a growth in Impressions as your website becomes more visible in the SERPs.

Usually we would expect to see increased visibility in the SERPs equate to more clicks to our websites. However, because the user is now receiving the answers they need directly in the search results page, often no click is required to source further information and the user ends their journey here with ‘zero clicks’.

The term ‘zero-click’ search has been around for years with the impact of rich snippets - such as ‘People Also Ask’ - providing users with the information they need. However, with the implementation of AI Overviews ‘zero-click’ searches are becoming more prevalent.

The ‘Crocodile Effect’

The combination of a speedy growth in impressions but not clicks is creating what marketing professionals are calling the ‘Crocodile Effect’ or ‘the great decoupling’ between the metrics.

Image Source: Marta Haddock

A recent study by Semrush, highlighted that when AI Overviews appeared within search results, organic traffic clicks dropped by 64% and the click-through rate for the top organic position dropped from 24.7% to 15.8%.

So, while it may be too early to know what the long term implications of AI Search features like AI Overviews will have, we can already see results beginning to change across multiple industries.

Impact of LLMs on Organic Traffic

Unlike AI Overviews whose impact can be seen in our reports, the effect of LLM searches on Organic Traffic and our results is harder to discern with the difference coming down to visibility and measurability.

As use of LLMs increases the impact of LLM searches on Organic Search is harder to quantify at this stage due to a limited accessibility of relevant data. Though it is suggested that - as with AI Overviews - because answers are being found directly within the LLM platforms less people are likely to click, increasing those ‘zero-click’ results.

Why SEO is still essential in the AI era

What is clear is that both AI Overviews and LLM searches are changing the search landscape which naturally raises the question of where traditional search and SEO fits in this new era.

Some marketers suggest that AI search may eventually replace traditional search and SEO, and new approaches such as GEO (Generative Engine Optimisation) are beginning to emerge. While it is worth paying attention to these developments and considering how strategies might evolve, the foundations of search visibility continue to rest on well-established SEO principles.

Well-Structured Quality Content

High quality content remains important not only for search engines but also for large language models. Google’s E-E-A-T guidelines, which emphasise originality and expertise, appear to support visibility in both traditional and AI driven search contexts. Similarly, large language models often respond well to content that directly addresses user questions, is easy to digest, and written in natural, accessible language. Approaches such as FAQs and question-based headings may also help improve the likelihood of content appearing in AI Overviews or rich snippets.

Site Performance

Technical SEO and user experience factors are also key areas to maintain focus on. Since AI Overviews draw from crawlable, accessible sources, practices such as clear site architecture, logical header structures, optimised page speed, and semantic HTML all contribute to a stronger foundation. These improvements not only support search engines but may also enhance visibility within AI search features.

Brand Visibility & Building Authority

The weight and authority your voice carries has always been an important factor in ranking well in traditional SERPs and this is no different for AI Search.

Digital PR continues to become an effective way to expand your website’s visibility by earning mentions in third-party publications - either through having your content quoted as a source of expertise or securing links to your content. This often requires targeted outreach, by identifying relevant outlets and providing genuinely useful content for their audience. Websites have seen meaningful momentum with this strategy, however, there are limitations. Gaining connections with niche industry publications are within reach, but high-authority outlets are much tougher to break into. Hiring a link development agency or professional PR agency may help to lighten the load on internal teams and allow you to access those hard-to-reach publications that can really matter.

Structured Data

Structured data is another important area to consider. Detailed schema implementation can provide additional context and meaning to content, making it more understandable to both search engines and AI models. Some research indicates that high levels of property completion within schema markup may correlate with stronger performance in AI results.

The balanced path forward

While there is certainly more to consider, we know that SEO fundamentals such as those listed above are propelling websites in AI Search and - especially in a time of rapid change - maintaining and strengthening SEO fundamentals continues to be a practical way to support both traditional search performance and emerging AI-driven visibility.

AI-driven search is still in its infancy, and many aspects of long-term impact remain uncertain. What we do know is that change is already underway, and it would be short-sighted to assume search will remain as it’s always been. Search has always evolved and our practices with it.

While the shape of Search may change - through AI Overviews, LLM-powered answers, or new formats yet to emerge - we believe SEO will continue to play a vital role and have a seat at the table.