@neilpatel

Google told us exactly how SEO would change back in 2,015. Almost nobody listened because the old ways still worked. It doesn't anymore.

AI just made the thing Google warned us about impossible to ignore. Let me show you what everyone missed. But before we talk about where SEO is headed, we need to go back to where it all began, the keyword era.

Every marketer I talk to says SEO used to be easier. It wasn't easier. We just got lucky.

Back then, a lot of what we call great SEO strategy wasn't strategy at all. It was just good keyword matching. And those aren't the same thing.

We weren't understanding the user better. We were matching to the exact words they typed. The system rewarded.

So it felt strategic. Looking back, we were playing the system, not solving search. Think about it.

If you're selling shoes, did you know if your customer was a brand new runner, someone with a knee injury, or a dad buying shoes for his kid? No. And you didn't need to.

The keyword was a good enough standing for what they actually wanted. But Google was already trying to close that gap. Feature snippets, people also ask zero quick answers.

Every one of those was Google reaching for the moment behind the search. The crack was, it's already there. Before AI overviews even rolled out broadly, 57% of all searches were already ending without a single click.

After 59, I felt it on My own site pages I built keywords for were losing traffic even when the rankings help. At first I thought it was a ranking problem. It wasn't.

And on individual pages, it's worse. When a, a or you shows up above the result, it acts like a giant sponge. It soaks up the traffic.

Even when you're holding the No. 1 spot, the ranking still exists. Clicks don't sear.

Interactive track thousands of these searches and the click

fic. Even when you're holding the No. 1 spot, the ranking still exists.

Clicks don't sear. 76% down to point six one %. More than a 60% drop for the exact spot everyone's fighting to rank in.

So if Google was already moving away from keywords, what were they moving toward? Let me tell you what Google told us. In 2015, Google published something called Micro Moments.

It said people don't search at random. They search from four states of mind. Google broke it down into four moments.

Number one, want to know. That's when someone's researching. They're asking how does this actually work?

2, want to go. That's when someone is looking for a place like best running store near me. 3, want to do, that's when someone is trying to solve a problem like, like how do I fix knee pain when I run?

And 4, want to buy. That's when someone is ready to make a decision like Best Running Shoes for Flat Feet. Same keyword, completely different moment, which means you need completely different content.

That was a big Lesson, Google was telling us the keyword is just the surface. The real opportunity is understanding the moment underneath. And almost everyone ignored it because keywords were still a good enough cheat.

You could fake the moment with the keyword and the algorithm would let you get away with it until AI made faking it impossible. Nobody types in keywords into ChatGPT. They describe the whole situation.

My 7 year old had headaches every afternoon for 2 weeks. She drinks water. She's not stressed about school.

It only happens after she uses her tablet. What's going on? Business owners do the exact same thing.

I run a local bakery. I've posted on Instagram for 6 months. Decent likes

e uses her tablet. What's going on? Business owners do the exact same thing.

I run a local bakery. I've posted on Instagram for 6 months. Decent likes, zero new customers walking in, no budget for ads.

What am I doing wrong? Those aren't keywords. They're whole situations.

And the data shows people search this way now. Short queries, one to three words trigger an AI answer just 23% of the time. Six words or more, 77%.

The longer the search, the closer it is to a real human moment and the more AI steps in to answer it directly. And 76% of people now say they want a conversational search experience. They're not typing keywords, they're describing the problem and they want an answer, not 10 links to dig through.

So the real question is how to find the moment? How do you reverse engineer the moment behind the search? And why won't Google rankings get you there?

We studied 500 commercial keywords and ran over 4,000 prompts to multiple AI tools. 4% shot at showing up in the AI answer. I rank 4.

6%. And the one that floored me, 90% of the pages AI actually cited ranks 21 or lower on Google. Your Google rank and your AI visibility are two separate scores.

Now. AI used to pull most of its citation from Google's top 10 results, 76% of them. Today, that's down to 38%.

Three out of every four things AI sites now comes from outside Google's top results entirely. If you're only building for ranking, you're winning the wrong game. So what's the right game?

Understanding the moment. And there's a tool I use to map it that most people underutilize. Most people open answer the public to find blog topics.

That's using a telescope as a paperweight. What it's really showing you is how people feel right before they search. Every questio

o find blog topics. That's using a telescope as a paperweight. What it's really showing you is how people feel right before they search.

Every question cluster is a different moment. Take running shoes. Running shoes for bad knees is a pain moment.

Best running shoes for under $100 is a budget moment. And running shoes even worth it is a doubt moment. They need trust before they need information.

Three people, three completely different answers. Even though the keyword looks almost identical, when someone hands AI a situation, the model hunts for content that matches the whole moment. Content built around a moment gets cited Content built around a keyword string gets passed over.

This is the game. How to get cited. Getting cited by AI comes down to two questions.

Can AI pull a clean answer off your page? And does AI even find you in the places it already trusts? Almost every single SEO guide only answers the first one.

We're going to do both. Start with your own page because you can match the moment perfectly and still get ignored. The reason is structural.

A doesn't read your page the way a person does. It chunks it and it scans. For one thing, can I lift a complete confident answer to this situation right out of this page?

The answer is buried in paragraph 8. After the three paragraphs of Warm Up, AI skips you and grabs the page that leads with it. Here are three ways you can do to fix it fast.

1. Leave with the answer. Cut the and this section will explore.

Nobody reads that and neither does AI give them what they came for in the first line. 2, write the question as a headache in the exact words a person would use. What is the best moisturizer for acne prone skin under $100?

That's a prompt someone types at 11 at night, matches it word for word and you're halfway to

is the best moisturizer for acne prone skin under $100? That's a prompt someone types at 11 at night, matches it word for word and you're halfway to being the answer. 3, build a real FAQ section, the way people actually talk, not just in keyword speak.

And this is where answer the Public pays off again. Those question clusters you pulled earlier are almost word for word the prompts people type into AI. Turn your top 10 into a clean FAQ sections on your main pages and you'll move your citation rate more than months of technical SEO ever would.

Now, once you do that, does your page get cited? Sometimes a lot. When ChatGPT move from one model to the next, citations pointing straight at brand sites jump from 8% to 56% seven times in a single update.

Then the next model pulled it back down. That's a real lesson. The number swings every version.

So you build a page a can always extract from and you stop chasing any single model. But the bigger piece is the one that SEO experts skip even as peak. Most of what AI sites isn't your page at all.

Remember that 75% from earlier? Most of what AI sites comes from somewhere other than your own website. So winning your own page is maybe a quarter of the job.

The other three quarters happen out of your own website. On pages AI already trust, three places dominate right now. Reddit, it's one of the most dominant and cited sources across major AI engines, whether it's Google, Gemini, or even chat.

GPT review sites are massive as well. Brand profiles on G2, Captair and Trust Pilot get cited far more often than brands that aren't on them. And the randoms like the Best Running Shoes for Flat Feet or other top alternative to list.

AI leans on those constantly. So the move isn't writing more blog posts, it's getting your brand mentioned in the

et or other top alternative to list. AI leans on those constantly. So the move isn't writing more blog posts, it's getting your brand mentioned in the Places a already pulls from.

Now the blog posts that you write on your website are still valuable, but that's for the 25% you need to also focus on the other 75%. Hence you want to be on other people's websites as well. And you don't have to guess where those are.

Run your category questions through Uber's AI Visibility Report. It shows you the exact sources AI citing and it shows you for not only you but also your competitors. That list is your outreach plan.

Get onto review platforms, get into roundups that are already winning. Show up in the communities where your buyers already ask questions, match the moment on your page and earn your name in the places AI trust. Do both and you stop being invisible.

Now let me show you what 20 years of SEO actually taught me about marketing. Every big Google update, Panda, Penguin, Hummingbird, Burt, helpful content a or views move the target closer to what people actually needed and further from what markers could fake and moments can't be faked. Everyone is calling this shift from SEO to GEO.

It's bigger than that. For 20 years, we would optimize for the algorithm that we can measure. Now we optimize for what the person needed.

All along, Google's been inching towards That the whole time AI just closed the gap across 22 companies we track. 4% in a single year, doubled. And those were the early movers, the ones who started back when most people were still debating whether any of this matter.

The window is still open, but the compounding has already started. What does this mean for you? If you're building real content that answers real

ter. The window is still open, but the compounding has already started. What does this mean for you?

If you're building real content that answers real questions for people, you're not behind, you're ahead. You just have to reform at it for the new surface. Here's exactly what to do this week.

Take your 5 highest traffic pages open, answer the public for each one's core topic. Find the top 10 what, when, where, why questions. Pick three that match a want to know or a want to buy moment.

Write a clean FAQ section at the bottom of each page answering those three exact questions. Someone would say to AI, that's it for the week. 20 years goes by fast.

And the one thing that never changed through every update and every platform shift is this. The person on the other side of that search has a real problem that they need to solve. Build content that generally matches that and you'll outlast every update that's coming.

If you want our team to build this out for you, head to npdigital. Com. It's what we do every day for brands across every industry around the globe.

One question before You go look at your content right now. Are you writing for a keyword or for a moment? And if you want a deeper playbook, exactly how to get your brand onto the sources AI trust step by step, I broke that down here.