SEO Is Not Dead, but AI Search Has Changed How Visibility Is Earned

SEO is not dead. What has changed is where the competition happens and what counts as “ranking.” Google’s blue links still matter for plenty of searches, but a growing share of questions now get answered inside ChatGPT, Perplexity, Gemini and Google’s AI Overviews before a user ever clicks through to a website. If your business isn’t structured in a way these tools can verify and trust, you can have solid rankings and still be invisible at the exact moment a customer is deciding who to call. That’s not SEO dying. It’s the visibility system adding a new layer on top of the old one.

Table of Contents

• What specifically has changed since AI search tools launched?
• Do Google rankings still matter if AI is answering directly?
• What should a business stop doing and start doing?
• The bottom line

What specifically has changed since AI search tools launched?

The mechanics of how people find a tradie, a cafe or a clinic have shifted in three concrete ways.

First, the starting point has changed for a meaningful chunk of searches. Someone looking for a plumber in Richmond might still Google it, but increasingly they’ll ask an AI assistant directly, phrased as a full question rather than a string of keywords. That assistant then has to decide who to mention, often without the user ever seeing a list of ten options to scroll through.

Second, the criteria for being chosen have shifted. Classic SEO rewarded relevant keywords, backlinks and a technically sound site. AI engines still care about some of that, but they weight it against different signals: whether your business name, address and phone number are consistent everywhere you’re listed, whether your site has structured data (schema markup) that explicitly tells machines what kind of business you are and where you operate, and whether your content answers questions plainly enough to be lifted and cited rather than just read.

Third, the output itself has changed. A Google ranking gives you a clickable listing. An AI citation gives you a mention inside a generated answer, sometimes with a link, sometimes without. This means visibility can happen even when a click doesn’t, which complicates how businesses should measure whether their efforts are working.

None of this makes traditional SEO irrelevant. It makes it necessary but no longer sufficient.

Do Google rankings still matter if AI is answering directly?

Yes, for several practical reasons, though their role has changed.

AI Overviews and chatbots don’t invent answers from nothing. They draw heavily on sources that already rank well organically, that have authoritative backlinks, and that demonstrate topical consistency over time. A site with poor technical SEO, thin content or inconsistent business information is less likely to be treated as a reliable source regardless of how well-written its answers are. In that sense, solid SEO fundamentals are increasingly the entry ticket to being considered for AI citation, not a separate competing strategy.

There’s also the simple fact that not every search triggers an AI answer. Many local, transactional searches (“plumber near me,” “book a table tonight”) still return standard map packs and organic listings. Ranking well there remains directly tied to enquiries.

The comparison worth making is not “SEO versus AI search” but what each layer is optimising for:

• Traditional SEO optimises for ranking position in search engine results pages, built on technical health, backlinks, keyword relevance and user experience signals.
• Generative Engine Optimisation (GEO) optimises for citation inside AI-generated answers, built on structured data, entity verification, NAP consistency and content written to directly answer questions.

They share a foundation (a technically healthy, well-structured, genuinely useful website) but diverge in what they’re ultimately trying to win. Businesses that treat them as rivals for budget are missing that the second layer depends on the first being done properly.

What should a business stop doing and start doing?

Stop treating keyword density and link volume as the whole game. These still have a place, but pouring resources into adding more content without structure or chasing backlinks while ignoring how a bot “reads” your business is increasingly inefficient. AI engines are far less interested in how many keywords appear on a page and far more interested in whether the page clearly, factually answers a question a real person might ask.

Stop assuming consistency across the web doesn’t matter. A surprisingly common failure point is simple inconsistency: the business’s name, address or phone number differs slightly between its website, Google Business Profile and directory listings. Humans glance past this. AI engines, which are trying to verify that a business is real and trustworthy before citing it, treat it as a reliability flag. According to RankMeFirst.ai’s work auditing Melbourne businesses, inconsistent NAP data, missing or invalid schema markup, and content that isn’t structured to directly answer questions are the three most common reasons AI engines skip over otherwise legitimate, good-quality businesses.

Start auditing your structured data. JSON-LD schema is the technical language that tells an AI engine what kind of business you are, where you’re located, what you offer and how to verify you. Many small business websites either have none of this or have it implemented incorrectly. This is usually one of the faster fixes available, because it’s a technical correction rather than a content overhaul.

Start writing content that answers questions directly, near the top of the page, rather than burying the useful information under long preambles designed for keyword padding. If a customer might ask “how much does X cost in Melbourne” or “what’s the best type of Y for a renovation,” your page should answer that plainly, early, in a form that’s easy to extract and cite.

Start checking your entity signals. This means things like having a complete, verified Google Business Profile, a consistent presence on LinkedIn, and any other third-party data points that let an AI engine cross-reference that your business is genuinely what it claims to be. Thin or absent entity signals make it harder for AI tools to trust a citation, even when the website content itself is good.

A sensible sequence for most small businesses is: fix NAP consistency first (it’s usually cheap and fast), then implement or correct schema markup, then restructure key pages to be answer-first, then build out entity signals over time. Structural fixes tend to show results faster than content rewrites, since AI engines need to re-crawl and re-index content changes before they reflect in citations.

The bottom line

SEO isn’t dead, but the definition of “visible” has expanded. Ranking on page one of Google used to be the finish line. Now it’s one of two finish lines, and the second one (being the business an AI tool actually names when someone asks it a question) runs on a different, more structural set of rules. Businesses that keep investing in sound SEO while adding the technical and structural work GEO requires are positioning themselves for both. Businesses that ignore the second layer risk doing everything right by 2015’s standards and still not getting mentioned when it counts.

If you’re not sure where your own site stands on the structural signals AI engines check, a GEO audit is a reasonable first step: it shows specifically what’s missing before you spend money guessing.