Strategy
Risto RehemägiThe days of the “ten blue links” are over.
Large language models (LLMs) are the new intermediaries between users and the content they seek. It’s clear that this is the direction Google is heading with the launch of AI Mode in the US.
Google still leads the market and that is unlikely to change without a drastic platform shift (top browsers or devices changing default search engines). Google saw more than 5 trillion searches in 2024, or about 14 billion per day, giving it a 93.57% market share. ChatGPT has estimated 37.5 million search-like prompts per day, giving it a 0.25% market share. That’s less than Microsoft Bing (4.10%), Yahoo (1.35%), and DuckDuckGo (0.73%) but it is growing quickly.
What does this mean for SEO?
This evolution poses a significant challenge for publishers. Since AI-driven answers don’t direct nearly as many clicks back to websites like traditional search did (AI Overviews Reduce Clicks by 34.5%), there’s less incentive for content creation.
SEO returns are dropping fast, so the investment must also be decreased. The old model of hiring in-house writers and strategists (or even cheaper freelancers) just for SEO doesn’t make sense anymore because it is too slow and too expensive.
LLMs do create an opportunity for smaller players though. If they strategically tie LLMs into their content creation process, they can produce content at similar scale and quality to much bigger companies. Even if they don’t win the market in terms of traffic, it can still outperform other marketing channels for ROI.
How to get mentioned in LLMs?
The techniques to get mentioned in LLM answers are still very similar to traditional SEO. In essence you need to increase your digital footprint on the web. The more content you produce and the more people talk about you, the more surface area LLMs have to reference.
One thing changes slightly - link-building is no longer just about getting a backlink; it’s about getting your brand discussed and referenced across the web (they dont have to link). Because LLM don’t assign authority to their sources as well as Google search does, you can abuse this mechanic by writing mentions yourself on prominent domains (Reddit, Medium, Linkedin articles).
Analytics?
There are very few tools for us to track anything yet. It is not clear whether or not the big AI companies will release analytics tools that would let us see what’s happening inside.
Perhaps most notable tool we do have is Cloudflare AI audit, which allows us to see how LLM crawlers interact with our website.
There are also some more qualitative LLM answer tracking tools popping up, that show us whether or not LLMs reference us for specific queries (like https://trakkr.ai/).