When Users Search, They No Longer See Your Website — They See AI's Answer. What Do You Do?

A few days ago, a friend of mine came to me with a complaint.

He said his company's website had been doing SEO for seven or eight years. Keyword rankings had always been solid, traffic was steady. But over the last six months, traffic had been mysteriously dropping. He pored over server logs, crawl records, backlink quality — couldn't find a single thing wrong.

I said to him: have you considered that maybe you didn't do anything wrong — maybe the way users search has simply changed?

He froze for a second.

I pulled out my phone and searched a term: "best robot vacuum." And guess what? The first result wasn't a website at all — it was an AI-generated answer. The list of websites only appeared below that.

What users see is the AI's answer, not your first-ranked page.

That's what I've been thinking about lately. And it has a name: GEO.

What Is GEO?

GEO stands for Generative Engine Optimization.

Sounds intimidating. But stripped down to one sentence —

You need to make sure AI mentions you when it answers users' questions.

In the old days of SEO, you competed with your peers on one thing: who ranks first in the search results. The user has to click through to your site for you to win.

With GEO, you compete on something else entirely: whether the answer AI generates includes your name, your product, your perspective. If ChatGPT, Perplexity, or Google's AI Overviews recommend products to users and never mention you — users won't even know you exist.

You don't even get a chance to show up.

Think about that for a moment. It's a scary thought.

Is This Just SEO with a Different Name?

A lot of people's first reaction: isn't that just SEO rebranded?

No. It isn't.

Let me give you an analogy.

SEO is like opening a store in a mall and fighting to get your sign in the most prominent spot by the entrance. When customers walk in, they see you first.

GEO is like — the customer never walks into the mall at all. They stand outside the door, ask a robot "I want to buy a robot vacuum, which one's good," and the robot gives them the answer right there. The customer takes the answer and walks away.

Position doesn't matter anymore. Whether the robot is willing to mention you — that's what matters.

But these two things aren't opposed to each other. And that's the second point I want to make clear —

AI's answers don't come from nowhere. They're built from information scraped across the internet. All that content you've painstakingly created over the years, the backlinks you've accumulated, the word-of-mouth discussions about you on Reddit and YouTube — that's exactly what AI feeds on.

So if your company has built a solid SEO foundation, you already have a head start on GEO. You've already won more than half the battle. The remaining half is about making it easier for AI to "understand" you and "talk about" you.

So, How Do You Actually Do It?

After looking into it, I found that the factors that genuinely move the needle on whether AI mentions you come down to a few key points. I'll highlight the most important ones.

First: How Often Your Brand Gets Mentioned Matters More Than Links

This is pretty counterintuitive.

In the old SEO playbook, we obsessed over backlinks — who's linking to us, how authoritative those links are. But AI doesn't care much about links. What AI cares about is how much buzz there is around your brand across the internet.

Even if the page mentioning you doesn't link to you at all. Even if it's just a Reddit user casually dropping a line like "I've used brand X, it's decent."

That kind of unlinked, offhand mention might be more valuable to AI than a high-authority backlink.

Why? Because AI cares about "how many people are actually talking about this brand," not "how impressive its backlink profile looks."

Second: Your Content Needs Substance

Someone ran a study analyzing ten thousand real queries.

The result? Pages with data, citations, and concrete numbers were 30% to 40% more likely to be cited in AI answers than pages that were "all opinions, all sentiment."

Let that number sink in for a moment.

AI is a picky eater. It likes things it can grab and use directly — a point that can be made in a single sentence, a specific number, a quotable passage. If everything you write is filler — "we believe," "we should," "the future looks bright" — AI wouldn't know what to pull even if it wanted to mention you.

Third: Your Pages Need to Be Machine-Readable

This one leans technical.

A lot of websites today load their content dynamically in the browser using JavaScript. Humans see the page just fine, but AI crawlers sometimes can't execute JavaScript — which means your content is effectively invisible to AI.

No matter how well you write, if AI can't see it, it might as well not exist.

That's why technical details like server-side rendering have become important again in the GEO era.

Fourth: You Need to Look "Fresh" to AI

AI tools generally prefer giving users the most up-to-date information. If the content on your site was written five years ago and hasn't been touched since, AI sees it as an "expired" source.

Regularly updating your content and publishing new material isn't just about keeping search engine crawlers happy — it's about signaling to AI that "this brand is alive and still has things to say."

Fifth: Wikipedia and UGC Platforms Are Goldmines

Wikipedia is one of the most important sources of AI training data. If your brand has an accurate Wikipedia entry, the chances of being mentioned by AI go up significantly.

Then there are platforms like Reddit and YouTube, where users generate their own content. I've noticed something interesting — when AI answers questions like "which product is best," it loves referencing genuine user discussions on these platforms. What ordinary users say about you carries more weight with AI than what you say about yourself.

How Do You Measure This?

You've taken a whole series of actions — but are they actually working? You need to be able to measure it.

That's what Semrush does. Their Enterprise AIO can track your brand's share of voice across major AI search engines, how often you're mentioned, and whether the sentiment around those mentions is positive or negative. Smaller teams can use their AI Visibility Toolkit — lighter on features, but enough to see which AI platforms mention you, how many times you've been cited, and which pieces of your content are getting picked up.

If you're doing GEO, you need to put these metrics alongside your traditional keyword rankings and organic traffic in your weekly reports.

Otherwise you'll do a ton of work and have no idea whether any of it made a difference.

One More Thing

For years, those of us in marketing have been fixated on one thing: "clicks." Clicks, rankings, impressions — we know this playbook inside and out.

GEO means marketers now need to keep an eye on something else: "being cited."

Being cited means AI is willing to single you out in its answer — to say "this brand is pretty good" or "this company's perspective is such-and-such."

This isn't a replacement for SEO. It's an addition. Your past SEO work wasn't wasted — it's still laying the groundwork. But if you don't start tracking "being cited," in six months or a year, you'll find that your rankings are still there, your traffic is still there, but when users are ready to buy — AI didn't mention you.

My friend eventually figured out why his traffic was dropping.

It wasn't that his rankings had fallen. It was that users simply weren't scrolling down far enough to see him anymore. AI had handed them the answer right on the first screen.

That day, I told him —

The question you need to ask yourself from now on isn't "where do I rank?" — it's "is AI willing to mention me?"

Those two questions sound similar. But the answers could be completely different.