When Google Says “AI Optimization” Is Just SEO, What Does It Mean?

Nothing has changed, according to Google :) But has it really not?

Google published its official “AI Optimization” guidelines a couple of weeks ago. There’s nothing completely new there (all the guidelines reiterate SEO best practices Google shared before).

Don’t forget that these guidelines are about Google search AI-powered features (AI Mode and AI Overviews). It is not clear if these guidelines apply to Google’s standalone generative AI platform (Gemini). And these definitely don’t refer to other LLM platforms (like ChatGPT or Claude).

—->>> JOIN us to discuss this topic LIVE and bring your questions we will answer in real time

Also, we are hosting another AI visibility AMA, this time on Digital PR. Ask your questions!

SEO is as relevant as ever

The guidelines confirm that SEO is still very relevant because:

• AI answers rely on search results to give more accurate and up-to-date results

• “Fan-out” queries are search queries that Google uses to search before giving an answer.

Google doesn’t call “optimizing for generative AI answers” anything other than “SEO”.

From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.

Best SEO practices still apply to AI answer optimization

Google recommends prioritizing effective SEO strategies over “AEO/GEO hacks”. The guidelines list the same practices that Google recommends for traditional SEO, mainly:

• Create “people-first” content (in these guidelines, it is referred to as “non-commodity”) with a unique/truly original point of view

• Structure your web pages to make them easy to read and use for people (no additional effort to optimize content for agents is required)

• Use high-quality and helpful videos and images. Overall, focus on what your site users would enjoy.

• From the technical standpoint, make sure your page is crawlable by (and is not blocked for) Google’s bots (Search Console is very helpful for verifying that)

• Semantic HTML structure should be created for people (and screen readers). No special optimization for AI agents is required.

• If you have JavaScript, make sure your content is visible with it disabled (Search Console is all you need for that too)

• Make sure your site provides a good user experience when accessed from various devices

• For ecommerce businesses, also ensure your product feeds are submitted in the Merchant Center.

What about the myth-busting?

Google’s guidelines are busting the most common myths in “GEO”:

• LLMs.txt file is not needed (which is not that clear btw)

• Markdown pages are not needed (it was later discovered these actual guidelines existed in markdown :))

• “Chunking” content (i.e., breaking it into short paragraphs for AI retrieval) is not required, unless it is for people ;)

• “Seeking inauthentic” links/mentions may flag your site as spam (just in the traditional SEO). AI systems do evaluate brand and product mentions across the web, but, just like Google, they can tell real mentions from fake ones. No wonder, as this is about Google’s LLMs

• No “special” structured data (schema) is required. For general SEO, it is not a bad idea to use the recommended schema types, as it helps generate rich snippets. This remains questionable as other Google guidelines recommend schema for clarity, not just rich snippets ;)

Overall, the guidelines emphasize two core ideas:

• Google’s AI answers directly rely on organic search

• For Google’s AI answer optimization, no “additional” tactics are required or recommended

BUT AGAIN, This is ONLY About The Search Features

Google talking about search features and saying it is still SEO has two quite contradicting messages, in my opinion:

• It is about search, of course, it is SEO

• Saying nothing has changed is very convenient, since they give us 0 insight into the click-through while everyone is losing clicks over those search features.

Here’s what changing (and problems I have with those guidelines):

• Clicks are going down

• There are no 100 reliable metrics to measure 0-click visibility

• Attribution is weird (as smart brands switch to cross-channel instead of relying on Google, and many of those channels are 0-click too!)

• Many of those myth-busted elements may be useful for other purposes outside of SEO (e.g., LLMs.txt file seems to be adopted quite well by API developers), so the messaging is terribly confusing.

• There seems to be no alignment inside Google (nothing new but still alarming)

• Google is pushing AI features (and an AI-powered search box) without clear guidelines on how businesses should adapt (instead, they just claim nothing is changing)

To be fair, even in the SEO guidelines, Google notes that in the future, AI agents will not only search but also perform various actions on human users’ behalf, like making a purchase or booking a hotel. So it is recommended to get prepared for that future and become familiar with the emerging protocols, like Universal Commerce Protocol. There are also separate recommendations for creating AI-agent-friendly websites, which website owners can start implementing.

This is a mere mention, but it should already be a wake-up call for everyone.

To sum up:

• Everything is changing

• Just SEO best practices are not enough (not that they were actually enough before)

• There’s another level of AI Visibility before LLMs, searching. More on this here. So it is not quite “just SEO”.

If you need help figuring this all out, I am always open to chatting!