Google Just Updated Its AI Search Guidance, But What About ChatGPT, Perplexity, Claude and other AI platforms?

If you've been keeping an eye on the SEO world recently, you may have seen Google release new guidance around how websites can perform well in its AI-powered search features, things like AI Overviews and AI Mode. It's a useful document, and we'll come to the key takeaways in a moment. But there's something worth flagging before we dive in.

The guidance is written entirely from Google's perspective, about Google's own products. ChatGPT, Perplexity, Claude, Gemini none of them get a mention. And that matters, because a growing number of people are now starting their searches on those platforms instead of Google.

So let's look at what Google is actually saying, and then ask the bigger question: does any of it apply beyond Google Search?

What Google's guidance actually says

The headline message is reassuring for anyone who has been doing SEO properly: nothing fundamental has changed.

Google's AI features are built on top of the same core ranking systems that have always existed. That means the basics: good technical SEO, a well-structured site, and content that genuinely helps people are still the foundation. If you've been doing those things well, you're already in a reasonable position for AI search too.

The more significant shift is around content quality. Google is drawing a clear line between two types of content. On one side, there's what they call commodity content generic, could've been written by anyone, adds nothing new. The kind of article that recycles what's already out there without bringing any fresh perspective or genuine experience to it.

On the other side, there's non-commodity content. Firsthand insight, expert opinion, real experience. Content that could only have been written by someone who actually knows what they're talking about. Google is being quite direct: the second type is what their AI systems are looking for. The first type is increasingly invisible. For businesses investing in content, this is an important signal. The bar has risen, and volume without quality is unlikely to get you very far.

What Google says you can ignore

Equally useful is what Google explicitly tells you not to bother with. There's been a lot of noise recently about various AI SEO tactics, and Google has taken the unusual step of calling several of them out directly. You don't need to create special AI text files. You don't need to break your content into small chunks for AI to process. You don't need to rewrite your content in a specific way for AI systems. And chasing artificial mentions across the web isn't going to help you.

This is worth knowing, both for your own strategy and for filtering out the more questionable advice that tends to circulate whenever something new emerges in search.

But here's the gap in this guidance

Google's document is clear and helpful within its own scope. The problem is that scope is narrower than many people might assume when they read it.

Search behaviour is shifting. A meaningful and growing number of people are now turning to ChatGPT, Perplexity, and similar platforms as their first port of call when they want to find something, research a purchase, or get a recommendation. These platforms are not Google. They work differently, they pull information from different places, and the signals that influence what they surface are not identical to Google's ranking systems.

So when Google says that seeking mentions across the web isn't particularly helpful, that might be true within Google Search. But on platforms like Perplexity, which actively cites sources and surfaces content based on what's being referenced and discussed, visibility across reputable websites and publications may carry significantly more weight.

I've seen this play out recently with a local, direct-to-consumer business we work with. After implementing schema markup consistently across their site, we not only saw an improvement in rich snippets within Google Search which was the expected outcome but within four weeks the business started appearing and being linked to directly within both ChatGPT and Perplexity. Before the schema work, they had no presence on either platform at all. That's one example, and it would be premature to draw sweeping conclusions from it. But it does suggest that structured data may be doing more work behind the scenes on AI platforms than Google's own guidance would lead you to believe. It's something we're watching closely.

Understanding how AI platforms actually retrieve information

To make sense of all this, it helps to understand a term you may have started seeing more frequently: RAG, or RetrievalAugmented Generation. In plain English, RAG is the process an AI uses to go and fetch current, realworld information before generating a response. Rather than relying purely on what it was trained on which has a knowledge cutoff and can go out of date a RAGenabled AI reaches out to live web sources, retrieves relevant content, and uses that to inform its answer. Think of it as the difference between asking someone a question from memory versus asking them to quickly look something up before they answer you.

Google's AI features use RAG. So do ChatGPT, Perplexity, and Claude but they each do it very differently. And those differences have real implications for how your content gets found and cited.

Google uses RAG in a way that's deeply tied to its existing search index. When someone asks a question, it retrieves relevant pages using the same core ranking signals it has always used, then generates a response based on that content. This is why traditional SEO remains so directly relevant to Google AI visibility. If your content ranks well organically, it is already in the pool Google's AI is drawing from.

Perplexity takes the most aggressive approach to retrieval. It searches the web for almost every query, citing sources in nearly 98% of responses and averaging close to ten citations per answer. This makes it the platform where being mentioned and referenced across reputable external websites carries the most direct value. If your business or content is being talked about in authoritative places, Perplexity is most likely to surface and link to it.

ChatGPTtriggers web searches selectively, depending on the type of query. Discovery and recommendation queries "what's the best tool for X" or "who provides Y in my area" are much more likely to prompt a search than straightforward informational questions. This means the way your content is framed and the questions it answers can influence whether ChatGPT reaches for it at all.

Claude is the most conservative of the major platforms when it comes to retrieval, and its behaviour is publicly documented through Anthropic's own technical materials. Rather than searching the web by default, Claude first assesses whether its existing training knowledge is sufficient to answer a question. Only when it identifies a genuine gap does it reach out for live information. The practical result is that Claude cites external sources in fewer than four out of ten queries, averaging just 2.1 citations per response compared to Perplexity's 9.8.

There's also something most people don't know about Claude: when it does search the web, it doesn't use Google or Bing. Research has found an 86.7% overlap between Claude's search results and Brave Search, an independent search engine with its own separate index. This means your Google rankings don't automatically translate into Claude visibility; your content needs to be discoverable within Brave's ecosystem too.

Claude also respects your robots.txt file, checking it before retrieving any content from your site. If your site is blocking AI crawlers which many publishers started doing in 2024 to prevent training data collection there's a good chance you're also inadvertently blocking Claude's retrieval bots, removing yourself from AI search citations entirely. Research suggests this is happening to the majority of publishers who've implemented any kind of AI bot blocking. It's worth checking.

What we actually know works across all platforms

It's important to be honest here. None of the major AI platforms outside Google have published official guidance on how to optimise for their search features. What follows is based on available research, industry observation, and our own experience not confirmed platform documentation.

That said, a clearer picture is starting to emerge. Traditional SEO remains the strongest foundation you can build on. The signals Google has refined over decades: authoritative content, quality backlinks, clear site structure, technical health are not irrelevant to AI platforms. In fact, platforms like ChatGPT and Perplexity are known to draw on web content that ranks well in traditional search. If your SEO foundations are strong, you're already better positioned than most.

Being cited and referenced across reputable external websites appears to carry significant weight, particularly on Perplexity. If your business is being talked about in industry publications, well-regarded blogs, or local press, that visibility seems to translate into AI platform presence more directly than many people realise.

Content clarity and structure also appears to matter considerably. AI platforms are essentially reading your content and deciding whether it answers a question well enough to reference. Pages that are clearly written, logically structured, and genuinely answer the questions your audience is asking tend to perform better. Vague, thin, or overly promotional content is unlikely to be surfaced as a reliable source by any platform.

Schema markup is worth paying attention to, despite what Google's guidance implies about its limited role in AI search. Our own client experience and a growing body of industry observation suggests structured data may be doing more work on other platforms than is officially acknowledged. It helps AI systems understand exactly what a page is about, who it's for, and what information it contains.

Brand consistency across the web also appears to play a role. AI platforms seem to favour businesses and sources that appear coherently across multiple touchpoints: their own website, third party listings, review platforms, social profiles, and external mentions. A business that exists clearly and consistently in multiple places online is simply easier for an AI system to understand, verify, and recommend with confidence.

And underpinning all of it, the quality and depth of your content remains the common thread. AI platforms are designed to give people reliable, helpful answers. Content that demonstrates genuine expertise, draws on real experience, and goes beyond surfacelevel information is consistently what gets surfaced across every platform, not just Google.

The bottom line

Google's new guidance is useful, but it only tells part of the story. The AI search landscape is broader, more varied, and evolving faster than any single platform's documentation can capture right now.

The smartest position for any business is to build content that demonstrates real expertise and genuine experience, the kind that serves a reader well regardless of how they found it. That approach travels well across platforms, even as the rules continue to develop.

If you're unsure how your site is currently set up for AI search visibility whether that's schema markup, content structure, robots.txt configuration, or how you're appearing across platforms like ChatGPT and Perplexity that's something we look at as part of our free 30 minute SEO review.