AI Search Engines vs Google: SEO Impact Explained

A client rang me in a bit of a panic in February. Traffic down 22% year on year, rankings unchanged, nothing broken on the site. Position two for their main term, same as always. Fewer people clicking.

I’ve had some version of that call maybe fifteen times in the last year and a half. It’s the defining problem of search right now, and it doesn’t show up in your rank tracker at all.

What’s actually changed

Google didn’t get replaced. Let’s start there, because a lot of the commentary on this is wildly overstated. Google still handles the overwhelming majority of searches globally. ChatGPT, Perplexity, Claude, Copilot and the rest have taken a real slice, and it’s growing, but it’s a slice, not the pie.

What changed is what a search result looks like when it arrives.

Ten years ago, a search gave you ten blue links and you picked one. Now a decent chunk of queries return an AI-generated answer sitting above everything else, stitched together from several sources, with small citation links most people never touch. Google calls theirs AI Overviews. ChatGPT does something similar when it searches. Perplexity built its whole product around it.

So the user gets their answer without leaving the results page. Your ranking is fine. Your impressions might even be up. But the click didn’t happen, because there was nothing left to click for.

That’s the whole story, really. Everything else is a detail.

Why “zero-click” is the term everyone’s using

A zero-click search is one where the person gets what they needed and never visits a website. These have existed for years (weather, currency conversion, what year a film came out) but AI answers have massively expanded the category.

Anything that can be answered in a paragraph is now at risk. Definitions. How-to steps. Comparisons. Quick factual lookups. If your traffic was built on informational blog posts that answered simple questions, you’ve probably already felt this, and it’s probably not going to recover.

What’s held up better is anything where the person needs to do something afterward. Buy, book, compare specific prices, contact someone, download a tool, read a real opinion from someone with skin in the game. AI can summarise information. It can’t complete a transaction for you, and it can’t tell you what it was actually like to work with a particular supplier.

Do AI search engines rank things differently?

Sort of, but less differently than people assume.

Google’s AI Overviews pull heavily from pages that already rank on page one. So conventional ranking still matters enormously. The overview isn’t ignoring the index, it’s summarising it.

The AI assistants that search the live web (ChatGPT, Perplexity, Claude) work a bit differently. They tend to run several queries behind the scenes, pull from a handful of sources, and favour pages that state things clearly and directly. They also lean noticeably on aggregator content: Reddit threads, review sites, “best X for Y” listicles, forum discussions. Which is why so many brands suddenly care what’s being said about them on Reddit.

The bigger difference is volume of sources. A traditional search gives you ten options and you pick. An AI answer picks three or four for you and names them. Being fifth is no longer a consolation prize, it’s invisible.

What this means for how SEO services actually work now

The fundamentals held up better than anyone expected. Technical health, crawlability, site speed, internal linking, genuine authority, real backlinks. All of it still matters, because AI systems are reading the same web everyone else is and they trust the same signals.

What’s shifted is emphasis. A few things I’ve changed in how I approach this:

Answer the question in the first two sentences. Not after 400 words of throat-clearing about how the digital landscape is evolving. AI systems extract passages, and a passage that states the answer cleanly gets pulled far more often than one that builds to it. This is also just better writing, so no great loss.

Structure properly. Real headings that match real questions. Short paragraphs. Lists where lists make sense. Tables for comparisons. It makes content easier to parse for a machine and easier to skim for a human, and those two goals have basically converged.

Have something to say. This is the uncomfortable one. Content that summarises what’s already on the internet is now competing with a machine that summarises the internet for a living, instantly, for free. You lose that fight. Original data, first-hand experience, actual opinions, numbers from your own work, case studies with specifics in them. That’s what survives, and honestly it’s what should have been getting written all along.

Watch your brand mentions, not just your links. AI models build a sense of who’s credible in a space partly from how often and how positively you’re discussed across the web. Being mentioned in a comparison article you don’t control can matter as much as ranking for a term you do.

Fix your measurement. If you’re still judging SEO services purely on sessions, you’ll misread everything happening right now. Track branded search volume, direct traffic, conversion rate, referrals from AI tools in your analytics, and whether you appear in AI answers for your key terms. Traffic can fall while the business does better. It happens more than you’d think.

Should you stop investing in SEO?

No, and the people saying otherwise usually have something else to sell you.

Organic search still drives more qualified traffic than any other channel for most businesses. AI assistants are reading search results, so ranking is a prerequisite for being cited rather than an alternative to it. And if you abandon the channel for two years while competitors keep building, you don’t come back easily. Authority takes a long time to accumulate and no time at all to lose.

What’s fair to say is that the old playbook, publish volume, rank for long tail, harvest clicks, has a much shorter shelf life than it did. Agencies still selling twenty blog posts a month with no differentiation are selling something that stopped working. Anyone providing serious SEO services in 2026 should be talking to you about content quality, entity authority, conversion, and measurement, not just keyword counts.

The honest summary

Google is still the main event. AI search is a genuine and growing second front. Both of them increasingly answer questions on the page instead of sending people to yours.

So you’ll probably get less traffic for the same rankings, and the traffic you keep will be worth more per visit, because the people still clicking are the ones who wanted more than a summary.

Build for those people. The machines tend to follow.

Author