Introduction
The widespread adoption of AI Large Language Models (LLMs) is having a significant impact on how consumers use the internet, making AI Optimisation an increasingly important consideration for brands. In some sectors we have seen a decrease in search engine traffic and a rise in “Zero Click Search”. This is where people have got the answers that they need from Google/AI without the need to visit a website.
At first glance this may seem like a threat, however, early data suggests that AI-referred traffic can be highly qualified, though volumes are currently lower and vary significantly by industry. As a result, we have been putting time and effort into understanding what we need to focus on to succeed in this new environment.
Traditional SEO and AI Optimisation
The more things change the more things stay the same! LLMs use bots to understand the internet, similar to how search engines like Google work. This means that we need to continue to ensure that your website is a friendly environment to bots. Helping them efficiently crawl the website and understand the content on the website.
This is what traditional SEO does. It makes sure the website underpinnings are technically healthy and that the content of each page is focused on its core message (keywords) and structured in a way that is understandable to both Google and AI.
However, this does not mean “business as usual”. There are differences in how the LLM and Google crawl. Search engines crawl continuously, whereas LLMs are trained on a combination of snapshots of the web, licensed and synthetic website data, and user interaction data. This means that LLMs retrieve content at the point of the user making the query and isn’t crawling in real time.
This introduces a risk of not moving fast enough. If changes are only slowly rolled out to a website, we may miss our chance to influence the LLMs until the next round of training.
We need to ensure that we pay attention to the following:
Produce strong content
Content is still king, but we now need to write in an AI-friendly way. This means we continue to produce detailed, context-rich content for the website. With a focus on both keywords and semantically related keywords.
However, the structure of that content may need amending to include “chunking” and make more use of easy to digest headers and bullet points.
Content structure/architecture
When writing content we need to write with “AI first” in mind. This means following a certain structure:
- Addressability – Key claims, statistics, and assertions should be supported by clear, authoritative sources.
- Structure – Content needs clear headers and to-the-point paragraphs. The industry is calling this “chunking”, and it simply means that each paragraph should be able to stand alone and still make sense even when presented out of context. This mirrors how AI retrieval systems extract passages rather than whole pages.
- Interactivity – AI struggles to replace proprietary tools, interactive calculators, live data, and experiences that require user input. By including interactive tools/visuals AI will be more likely to encourage people to visit your website, rather than just steal your content.
Technical SEO
As mentioned above, if a website works for Google, then it will work for AI. The only key difference is the importance of site speed, while Google likes fast websites AI LOVES them.
There are other, unproven, tactics such as an increased use of JSON and LLM.txt but these remain debated and unproven.
Digital PR as a Core Pillar of AI Optimisation
It is in this space where we are seeing the significant changes. In the context of AI Optimisation, this means moving beyond backlinks and focusing on how brands are consistently represented across authoritative third-party sources. We call this “signals”.
In the old world of SEO we focused on building links on third-party websites that pointed back to yours. In the world of AI/LLMs they gather facts and datapoints from around the internet. These facts and datapoints can be from all types of sources and it does not matter if the source is linking to your website or not.
These facts and datapoints are called signals and we need to manipulate these parameters to encourage AI to use your content and cite you as a source. There are several ways that we can do this but before we dive into that we need to talk briefly about the importance of having a strong and clear message about your brand.
LLMs learn through repetition across authoritative sources. This means that we need to repeat the same strong and clear message across the internet. To do this successfully you need to:
- Agree a phrase (ideally 1 short version and 1 long version) that can be used across the internet to encourage repetition.
- Use that phrase across the Digital PR campaign.
- Include it in all:
- Articles
- Wiki pages
- Directories
- Anywhere else your brand is mentioned
Of course, any phrase that is used and deployed must be done in a natural way, we don’t want to make it appear that we are trying to spam or over-manipulate the LLMs!
Current PR work
In this new landscape it is more important than ever for SEO and PR teams to work together. By having an agreed core phrase and an understanding that digital placements are just as important as print media we are able to help shape and influence the “signals” that the LLMs are using to understand your brand and website.
By joining these dots, we will be able to increase performance within AI search.
The types of websites that LLMs will expect to see your brand mentioned are:
- Industry publications (online)
- Wikipedia
- Industry directories
- Established blogs
- Social media
Social signals
It is worth finishing on the importance of social media signals.
LLMs give a significant amount of attention to social media platforms when they are in their training phase. This is because they trust “authentic” discussion over corporate messaging. We can use this to our advantage.
This means using traditional marketing platforms like Instagram, Facebook and TikTok but also including discussion boards such as Reddit.
Text-heavy, discussion-based platforms such as Reddit and forums play a disproportionately large role in AI training and retrieval compared to visual-first social networks. However, engaging with these platforms needs to be done with a view of helping people, not promoting, otherwise Reddit users will talk badly about your brand, which will turn the LLMs against your brand!
Conclusion
The rise of AI-driven search does not signal the end of SEO, but it does mark a shift in how visibility, authority, and demand are created online. Traditional SEO foundations, technical health, crawlability, and well-structured, high-quality content, remain essential. However, success in AI search and the LLMs increasingly depends on clarity of message, consistency across authoritative sources, and content that is easy for retrieval systems to extract, interpret, and trust. Brands that focus only on rankings and clicks risk becoming invisible in the new world of “Zero-Click Search”.
Looking ahead, the brands that perform best in AI search will be those that treat SEO, content, PR, and social as a single connected strategy rather than isolated channels. By reinforcing clear brand positioning across credible third-party sources, supporting key claims with authoritative signals, and creating content and experiences that AI cannot easily replace, businesses can increase both their visibility within AI responses and the quality of traffic that follows. In this new landscape, AI Optimisation is no longer about chasing individual tactics, but about ensuring a brand is clearly understood, consistently represented, and trusted by both humans and machines.
