Your customers have stopped searching. They just ask the AI directly.

A while ago, a friend of mine who runs a specialty coffee business came to me complaining.

He said, Liu Run, I spent three years getting my company's SEO (Search Engine Optimization) to number one in our category. But recently he noticed something that sent a chill down his spine.

His potential customers had started skipping the search results.

They open an AI chat box and just ask: "I want to buy a Fair Trade specialty coffee, any recommendations?" Then the AI gives back a paragraph listing several coffee brands, with their prices and reputations all written out.

No links clicked. No scrolling to page two.

The AI served the answer directly to the customer.

He panicked. If customers never come to my website again, he said, haven't I wasted three years of SEO?

I told him, hold on, don't panic. Let me tell you a story.

From "being found" to "being quoted"

What is SEO?

SEO stands for Search Engine Optimization. You guess a hundred keywords, rank on the first page, and wait for customers to search their way in.

That logic held true for twenty years.

But have you ever wondered — when a customer just asks the AI directly, where does the information in that answer come from?

The AI "reads" it from pages all over the web, then "summarizes" it.

In other words, the AI is doing the reading comprehension for the customer. It sifts through countless web pages, finds an answer it considers reliable, and then copies your name, your product, your pricing into its response.

This process has a new name: GEO (Generative Engine Optimization).

What is GEO?

Put simply, it's making your content easy for large language models to "understand, trust, and quote."

Your old SEO competitor was the ten pages ranking above you. Now your GEO competitor is whether the AI is willing to write you into its answer at all.

These are two completely different games.

So, do small companies still have a chance?

This friend runs a small team of just eight people. What worried him most was: are those big brands with their massive budgets going to crush us again?

I told him, hear me out — you might not only stop panicking, you might even feel like laughing.

Why?

Because this GEO game doesn't reward whoever has the most — it rewards whoever says it clearly.

Big company websites often have tens of thousands of pages, but much of it is templates, boilerplate, marketing copy. The AI reads it and can't extract a single crisp answer.

And you, a small team — if you can answer one specific question clearly, with data, with sources, the AI will actually prefer to quote you.

You were never going to win on volume. You win on that one "ready-to-copy standard answer."

This is fascinating.

In the past, big companies threw money at pages and backlinks to push their way up the rankings. Now, the AI doesn't want piling — it wants "I can read this paragraph and use it to answer the user right away." Whoever can provide that kind of content gets quoted.

Take selling coffee. There are two approaches:

The first: write ten shallow blog posts with titles like "Ten ways to drink coffee," "The story of where coffee beans come from," "How to brew a good cup of coffee." Two thousand words each, all vague generalities.

The second: write one deep piece called "The complete Fair Trade coffee sourcing process: from the estate to your cup." It has specific estate names, sourcing price ranges, certification standards, expert quotes.

Which one do you think the AI will quote?

The second one. Because in the second one, every sentence is solid, citable material that can be lifted directly and used as an answer. The first one — the AI reads it and still doesn't know which sentence to copy.

Four steps to make your content "quotable"

So how do you actually do it? I'll give you four steps you can start today.

Step one: dig through your existing content and find the "copyable parts."

Open your website backend and look at your most-viewed posts. Read them from start to finish and ask yourself one question: is there a single sentence in this article that a customer would genuinely ask, and that gets answered clearly?

Think about it — many blogs pile on preamble and only circle back to the answer ages later. The AI reads it and has no idea which sentence is the conclusion.

Find those paragraphs that already answer crisply, and note them down. These are your "quotable assets."

Step two: install a "translator" on your web pages.

This translator is called Schema — structured data markup.

What does that mean?

When the AI reads your web page, it's essentially guessing. "Is this string of numbers a price? Is this address a store location? Is this paragraph a user review?" It's guessing.

And Schema is you proactively telling the AI: this is the price, this is the address, these are the business hours. No need to guess — I'm telling you directly.

How do you do it specifically?

• For product pages, use Product Schema — mark up the product name, price, stock, user reviews.
• For service pages, use Service Schema — write out what service you provide and which region you serve.
• For pages with a physical store, use LocalBusiness Schema — list the address, business hours, contact info.

This step isn't hard. Most website builders have plugins for it, and you can set it up in a few hours. But the effect is immediate. The AI goes from "guessing" to "confirming," and the probability of quoting you goes way up.

Step three: change your content-writing habit to "answer first, explain later."

This might be the hardest step, and the most important.

Most people, when they write, like to start with preamble, tell a story, add an intro, and only reveal the conclusion at the very end. That's how we were taught to write essays growing up.

But the AI doesn't play by those rules.

The AI is looking for a "ready-to-copy standard answer." If you bury the answer in the fifth paragraph, the AI might move on before it ever gets there.

So new content should be written like this: right at the top, the first sentence is the definition. The first sentence is the conclusion. The first sentence is that clean, directly quotable answer. Then you expand and explain.

If you look back at the opening of this very article, that's exactly how it was written. Conclusion first, then the why.

Step four: don't try to write about everything. Be the "ultimate authority" in one narrow niche.

A small team's resources are limited. You can't out-write big companies on every topic.

But you can pick one very narrow, very specific area and master it completely.

What does that mean?

If you sell Fair Trade coffee, then don't write ten vague coffee blogs. Focus all your energy on writing the most complete, deepest, most data-backed content in this niche. Make it so that every time the AI encounters a "Fair Trade coffee" question, you're the first one it thinks of.

Being number one in a narrow field is far more useful than being number one hundred in a broad one.

How do you know if GEO is actually working?

At this point, you'll ask a question: how do I know if it's having any effect?

A lot of people ask this.

In the old days of SEO, you'd watch the traffic in your backend. Traffic goes up, that means your ranking went up.

But GEO is different. The customer never clicked into your website. What they saw was the paragraph the AI gave directly. How do you even track that?

This calls for a change in mindset.

You need to look at two things:

First, how many times your brand name shows up in the AI's generated answers.

Second, whether any new customers found you because they saw the AI's answer.

For the first one, you can periodically go to each AI platform and ask it "any recommendations for such-and-such category" and see whether it mentions you. Manually or with tools — build up a sense of your "citation frequency."

For the second one, that's where your customer management system comes in.

At this point, I have to mention Salesforce. Not as an ad — it's because its CRM (customer relationship management) really can help on this front.

How exactly?

Your customer management system can record "which channel did this customer come from." Did they come from search? From an ad? Or from seeing the AI's recommendation?

Then you look at all that data together. For the customers who came in after being quoted by the AI through your GEO optimization, what's their quality? Are they closer to closing? Is their average order value higher?

If the answer is yes, then your GEO has ROI (return on investment).

That turns something very intangible — "did the AI mention me" — into something very concrete: "how much money did those mentions make me."

A few final words

That friend of mine who sells coffee listened to all this, then went quiet for a moment.

He said, so you mean I don't need to beat those big brands' websites. I just need to do one thing: make the AI believe that, in my field, I'm the most trustworthy source of answers.

I said, that's exactly right.

He asked again, then what if one day the AI stops citing any web page at all and just generates answers entirely on its own?

I said, that question, I can't answer right now either.

But I do know one thing: in this window of time while the AI still needs to cite external information, whoever can turn themselves into that "standard answer that gets quoted" will have picked up their ticket to the next round.

As for how long that ticket stays valid, I don't know either.

But at the very least, don't keep playing the old game, fighting for a ranking that nobody is looking at anymore.

Maybe, next time you ask the AI "how do I get my small business noticed," the answer it quotes will be the one you wrote.