I get asked some version of this question most weeks now, usually by a client who has spotted a competitor showing up in a Google AI Overview and wants to know why they haven’t. It’s a fair question, and I’ll be honest with you the same way I am with every client: there isn’t a separate playbook for AI search that nobody’s told you about. There’s what already works, and there’s being more deliberate about how you use it.
I’ve spent a fair bit of time this year testing this properly across client sites rather than just reading about it, and what follows is genuinely what I’ve found holds up, alongside the bits of common advice floating around that don’t.
How Does AI Search Actually Work?
When you see a paragraph of text sitting above the normal search results on Google, that’s an AI Overview. It’s generated using something called retrieval-augmented generation, which is a fancy way of saying Google’s AI pulls from pages already sitting in its search index and summarises them into a direct answer. It isn’t running a separate crawl of the web or judging your site by a different set of rules. Google has said this outright in its own developer documentation: the best practices for SEO remain relevant because these features are rooted in the same core ranking and quality systems as the rest of search.
ChatGPT and Perplexity work a little differently, since they’re standalone tools people go to directly rather than typing into Google. When browsing is switched on, they’ll pull live content from the web and cite sources as they go, though a good deal of what they know still comes from their training data rather than a live search.
Do You Need a Different Strategy for AI Search?
There’s a genuine split in opinion on this, and I’ve watched it play out across SEO forums for a while now. One camp says just do SEO properly and you’ll appear in AI Overviews as a natural consequence, since the retrieval step depends on whether Google already trusts your page for the topic. The other camp is convinced there’s a whole new discipline to learn, complete with its own acronym.
My honest take sits closer to the first camp, and it matches what Google itself has published. There’s no separate technical requirement to unlock, and no schema type or formatting trick guarantees inclusion. What I have noticed, though, is a sharper and more useful point that came up in an SEO discussion I was reading recently: AI Overview inclusion appears to be constrained by something like contradiction risk.
Before an answer gets generated, the retrieved passages seem to get checked against each other for consistency. If your website says one thing about your delivery times, your Google Business Profile says something slightly different, and a review mentions something else again, that inconsistency makes your content harder to trust, even if any single page reads perfectly well on its own. I think that’s a genuinely underrated point, and one worth taking seriously.
Writing Content AI Tools Can Actually Use
Answer the Question First, Then Expand
If a page takes three paragraphs of throat clearing to reach the actual answer, there’s a real chance it gets skipped in favour of something that leads with the substance. Start each section with the core point, then support it with the detail and context that shows you actually know what you’re talking about. Depth still matters here, this isn’t a case for being needlessly brief, it’s a case for not making the reader work to find the answer.
Write Like You’re Explaining It to a Person
Short sentences, plain language, and no keyword stuffing. AI tools are trained on human language, so content that reads the way you’d actually explain something to a customer standing in front of you tends to be far easier for them to interpret and quote back accurately.
Structure With Clear Headings and FAQs
A logical run of H2s and H3s gives both search engines and AI tools a map of what your page covers. Question-led FAQs, phrased the way a real customer would ask them rather than how you’d write a heading, are one of the more reliable ways I’ve seen content get picked up, since they mirror exactly how people now talk to AI tools in the first place.
You might also find this helpful
Why Consistency Across the Web Matters
This is the bit I don’t see covered nearly enough, and it’s become one of the first things I check when I’m auditing a client’s technical SEO setup for AI visibility specifically. AI tools don’t just read your website in isolation. They seem to weigh up whether what your site says lines up with what your Google Business Profile says, what review sites say, and what’s said about you on forums like Reddit.
If those sources tell slightly different versions of the same story, your business becomes a slightly riskier source to quote from, even though nothing on your own site is actually wrong.
Keeping your opening hours, service areas, pricing and contact details consistent everywhere you appear online is a genuinely simple, low-effort way to remove that friction, and it’s something most businesses have never thought to check.
Schema Markup: What It Does and Doesn’t Do
I want to be straight with you on this one, because a lot of guides oversell it. Schema markup, the code that tells search engines what type of content a page contains, does help machines parse and categorise your pages more reliably. What it doesn’t do is guarantee you a spot in an AI Overview.
Google’s own Search Central documentation is explicit that there’s no special schema requirement for AI features, and no formatting trick unlocks inclusion on its own.
So add FAQ and Article schema because it’s good practice and helps machines understand your site properly, not because it’s a lever you can pull to force your way into an AI Overview. If you’re on WordPress, Rank Math makes adding this fairly straightforward without needing a developer involved.
Building External Authority and Backlinks
AI tools, particularly ChatGPT, seem to weigh brand mentions across the web more heavily than most people expect, not just what’s on your own domain. Being referenced in comparison articles, genuine reviews, industry roundups and forum threads all seems to build the kind of distributed trust that gets a business mentioned when someone asks an AI tool for a recommendation.
This is really E-E-A-T in practice, Experience, Expertise, Authoritativeness, and Trustworthiness. Named authors with real credentials, links to credible sources, and being talked about favourably elsewhere on the web all feed into this. None of it is instant, and none of it is a shortcut, it’s built the same way credibility has always been built.
The Technical Foundations That Still Matter
None of the content work above matters much if the basics underneath are shaky. A slow site, particularly on mobile, is still a negative signal. Check that your key pages aren’t accidentally blocked from crawling and that your sitemap is current, since AI tools can only surface content that’s actually been indexed in the first place. A sensible internal linking structure, where a blog post links to a relevant service page and that page links back to supporting content, also helps paint a clearer picture of where your expertise sits across a topic.
You might also find this helpful
How to Audit Where You Stand in AI Search Right Now
Before changing anything, it’s worth getting a proper sense of where you currently stand, rather than guessing.
Start by typing your key search terms directly into ChatGPT, Gemini and Perplexity and noting honestly whether you appear, and if so, what’s being said about you. Check whether GPTBot and other AI crawlers are able to access your site by reviewing your robots.txt file, since some hosting providers block these by default without telling you.
In Google Analytics 4, go to Explore, create a blank report, set the dimension to Page Referrer, and filter for sources including chat.openai.com, perplexity.ai and gemini.google.com. It’s an imperfect measure, results vary between sessions even for the same prompt, but it gives you something concrete to track over time rather than a vague sense of whether things feel better.
If you’d rather have someone run this audit properly rather than piecing it together yourself, this is exactly the kind of thing I help clients with as part of our SEO services.
So, Does Traditional SEO Still Work for AI Search?
Yes, and I’ll say that plainly because there’s a lot of noise out there suggesting otherwise, usually from people selling a shortcut. The businesses I see doing well in AI search are, without exception, the ones who were already doing SEO properly. Well organised, genuinely helpful content on a technically sound site doesn’t need to start again from scratch for AI. It needs a bit more deliberate attention to structure, consistency across the web, and honest expertise, which is exactly the same advice I’d give for good SEO content generally.
I won’t promise you a guaranteed spot in an AI Overview, because nobody honestly can, Google’s own systems change constantly and no agency controls that. What I can promise is that the work involved is the same work that’s always paid off in search, just applied a little more thoughtfully. If you’d like a second opinion on how your content is holding up right now, get in touch or book a free call and we’ll talk it through properly.
