Google’s AI Optimisation Guide: What it Really Means

The rise of artificial intelligence in search has rocked the digital marketing industry. From AI’s effect on paid advertising to reduced click-through rates, it’s caused quite the stir. Automated answer engines such as ChatGPT and Gemini and search features like AI Overviews have led many marketers to believe that the SEO game has changed forever. 

However, we now know that optimising for AI isn’t much different to standard search engine optimisation practices, and Google’s newly released AI Optimisation Guide tells us exactly that. 

This document does debunk some of the assumptions and theories people have about AI optimisation, but it is important to recognise that Google’s ecosystem, while significant, isn’t the only one. Their documentation offers some helpful insights and relief for a lot of anxious marketers, but it may not be best practice for all agentic search. 

That being said, let's take a look at some of the key points from Google’s guide and discuss what it means for you. 

What Not to Focus On: Common Myths  

A lot of marketers have created overly complicated strategies aimed at altering their technical framework to please AI systems. Google has explicitly said this is not necessary.  

Google has confirmed that:

• You do not need llms.txt files - there’s no secret code or dedicated schema that forces Google’s AI to look at your site 
• You don’t need to overly fragment articles - Google’s systems understand the nuance of a multi-page topic and can extract relevant sections itself 
• You do not need to rewrite content for AI systems - chasing endless long-tail keyword variations is a waste of time. Google understands synonyms, intent, and context 
• Inauthentic brand mentions do not work - fabricating brand mentions on forums is fooling no one

What to Focus On: Non-Commodity Content is Key 

If a lot of the standard practices have remained the same, then what changes should you actually be making? One of the key concepts the guide discusses is non-commodity content.

Commodity content is generic and basic - something pretty much anyone can create. You can usually find hundreds of similar articles and AI can already generate this kind of information for its own training data. Google won’t see much value in linking to websites that only repeat it. 

Because of this, Google’s retrieval systems now favour non-commodity content. This is content that includes original data or research, genuine expert analysis and unique perspectives, and lived experiences that a machine simply can’t replicate. 

The content that an automated tool can draft in mere seconds based on existing web results will be ignored by Google’s AI features. Content that offers something more is more likely to be cited. 

Google Doesn’t Define the Entire Web 

This is where you need to keep in mind that Google isn’t the be all and end all. Their guide does provide a great framework for marketers, but its advice doesn’t extend to the entire internet. 

Google’s system is built on top of its normal search engine and relies heavily on Retrieval-Augmented Generation (RAG). This means it pulls information from websites that already rank highly in search results and traditional SEO factors remain extremely important. 

But, for other agentic systems, the practices Google dismisses could be vital in other areas of the web. 

• llms.txt - Google tells you to disregard these files, but they can be beneficial for helping non-Google agents understand your business quickly 
• Formatting for LLMs - Google has decades worth of indexing infrastructure that allows the company to read websites on the fly. However, many other systems don’t possess this, and instead read websites in real time. For these systems, clear formatting, an organised structured, and straight forward summaries remain crucial
• The ‘zero-click’ problem - many AI tools give users answers directly instead of sending them to the original website, which Google claims helps drive valuable traffic to creators. But other AI systems are often built to summarise information so users never need to click through to a website 

SEO Certainly isn’t Dead 

Google’s documentation proves that the fundamentals of good digital marketing haven’t changed - you still need to build a technically sound website complete with genuinely helpful content. Though investing in original, non-commodity content is becoming increasingly important. 

However, you can’t ignore the bigger picture. As AI agents become more common, brands that ignore new technical standards risk becoming invisible to these automated systems and the users relying on them. 

If you need expert help with search engine optimisation, don’t hesitate to get in touch with us today. 

‍