AI Search Isn't Killing SEO — It's Breaking Something Bigger

Every conversation about building a website eventually comes back to SEO. It’s the default assumption: good content plus good optimization equals traffic. That formula held for a long time.

But the conversation is quietly shifting. GEO — how your content gets picked up and used by AI systems — is becoming just as important. Maybe more so, for certain types of content.

The problem is most people are still optimizing for 2020.

01 — The Click Is Disappearing

I spent the past year doing SEO the right way.

Keywords. Content. Internal linking. Optimization. All the things that are supposed to compound over time. And for a while, the logic held. Write good content, structure it well, wait for traffic.

Then something stopped making sense.

Traffic wasn’t growing the way it should have. The searches were happening — I could see that. But the clicks weren’t following. People were finding answers without ever arriving.

If you’ve used Google recently, you’ve seen why.

At the top of many search results, there’s now an AI-generated summary — Google calls them AI Overviews. A paragraph or two that answers the question directly, pulled from various sources, presented before any links.

Users don’t need to click anymore. The answer is already there.

The old path was simple: Search → Click → Website. That was the deal. You optimized for search engines, they sent you traffic, you built an audience.

The new path looks more like: Search → AI Answer → Done.

For anyone who’s spent time building a content site, this is not a small shift. It’s the assumption the whole model was built on, quietly breaking.

02 — What This Feels Like From the Inside

I want to be honest about what I actually observed, rather than just citing industry statistics.

My traffic data started showing something strange. Impressions were stable — people were seeing my content in search results. But click-through rates were dropping. The content was being found, but not visited.

At first I blamed myself. Maybe the titles weren’t compelling enough. Maybe the meta descriptions needed work. So I optimized those. The numbers moved a little, then settled back down.

Eventually I had to consider a different explanation: the content was doing its job — it was answering questions — but AI was doing the delivery. My site was the source. The AI was the last mile. And the user never had to come to me at all.

That’s a strange position to be in. Useful, but invisible.

03 — SEO vs GEO

This is where a new concept becomes relevant: GEO, or Generative Engine Optimization.

It sounds like jargon, but the distinction is real.

SEO was always about one thing: getting humans to click your link. Rankings, traffic, conversions. The search engine was the gatekeeper, and you were optimizing for its algorithm so that humans would choose you.

GEO is about something different: getting AI to understand, extract, and use your content. Not clicks. Citations. Not traffic. Inclusion.

The difference in practice: SEO rewards content that ranks. GEO rewards content that answers. SEO cares about keywords and backlinks. GEO cares about clarity and structure. SEO measures success in clicks. GEO measures success in whether your content shows up in AI-generated responses — whether or not anyone visits your site.

One way to put it: SEO gets you found. GEO gets you quoted.

04 — What I’m Actually Changing

I’ve started making small adjustments to how I write, and I want to be specific rather than vague about what that looks like.

First, I’m writing more directly. Less throat-clearing, fewer qualifications before the main point. If the article is answering a question, I try to answer it in the first two paragraphs, then elaborate. AI systems favor content that gives the answer early.

Second, I’m structuring content for extraction. Clear headings. Concrete statements that can stand alone. If a sentence only makes sense in context, it’s less likely to be pulled into an AI summary. If it can stand on its own as a useful piece of information, it’s more likely to be used.

Third, I’m thinking about what I can offer that AI can’t easily replicate. Original data. Personal experience. Specific examples from my own work. These are things AI can quote but not generate on its own — which makes them more valuable, not less.

05 — This Isn’t the End of SEO

I want to be careful not to overclaim.

SEO isn’t dead. Search traffic still exists. Clicks still happen. For many queries — especially transactional ones, local searches, anything where the user needs to actually go somewhere — traditional SEO still matters.

But something has shifted in the middle layer. Informational content — the “how does this work,” “what is this,” “explain this to me” queries — is increasingly being answered without a click. That’s where the old model is breaking.

The question isn’t whether to abandon SEO. It’s whether to add a new layer of thinking on top of it.

The Real Question

For most of the history of the web, the goal was to be found by people.

Now there’s a second goal: to be understood by machines that talk to people on your behalf.

That’s a genuinely new thing. It changes what good content looks like. It changes what “distribution” means. It changes who the audience is — because sometimes the first reader of your content isn’t a human at all.

I don’t have a complete answer for what this means yet. I’m still figuring it out in real time, on this site, with these articles.

But I think the people who adapt earliest — who start writing for both humans and AI, who understand that being cited is becoming as valuable as being clicked — are going to be in a better position when the dust settles.

The rules are changing. That’s uncomfortable. It’s also an opening.