When the search changed its owner: three reminders brought to me by a GEO sharing

Tonight, I attended the “GEO Generative Engine Optimization: Making Your Brand Visible in AI Search” lecture organized by Brand Marketing Exchange, which was shared by two of my friends, Chao Ge and Ruoyun. One is a practical veteran who has been running a marketing community for more than ten years, and the other is the director of Today Weekly Academy, a sub-brand of Today Weekly. The two of them sang together and talked about this new topic that is rapidly taking shape in a high-level and down-to-earth manner. I had just finished dinner and originally only meant to show up to support my friends, but after sitting through the whole session, what I took away was not just methods, but also a kind of anxiety — it turns out that many people always assume they are working hard to accumulate, yet may be betting on the wrong place.

The speaker, Brother Chao, said something that made me restless after listening to it: He has been running a marketing community for 12 years and has written more than one million words on Facebook. As a result, if you search on Google, you can’t find even half a word unless you type his name. Twelve years and a million words later, the digital presence is almost zero.

This is not one person’s story. This is the common blind spot of our generation of content workers - we put our efforts in places that will bring instant applause, but do not put them in the soil that will accumulate assets for us. Now, search is changing owners: from Google’s crawlers to AI. The problem is that most people haven’t even won the last one (SEO), and the next one (GEO) has already started.

After listening to this sharing, I condensed it into three reminders.

1. Why GEO is worth starting now

The first reason is that the track is being reshuffled. AI is rewriting the rules of search, and old SEO bonuses are loosening up. This is pressure for old players, but it is a gift for newcomers: Brother Chao puts it bluntly, he has nothing in SEO, and now that he has entered AI search, he is at least on the same starting line as everyone else. In other words, just because it didn’t work in the past doesn’t mean it won’t work this time - this is the time window to enter the market.

The second reason is that AI recommendations have built-in credibility. More than 20 people voted at the scene, and almost unanimously believed that the brands recommended by AI were trustworthy. This intuition is actually very reasonable: AI should cherish its own feathers, and it will not randomly recommend a bad product, otherwise users will leave after a bad experience. So once you are selected by AI, what you get is not just exposure, but a transfer of trust.

The third reason is the most realistic - slots are extremely scarce. For one question, the recommendation list given by AI is only about five to eight. This is not long-tail profit sharing, this is winner-take-all. If you’re not on that list, you don’t exist.

Stacking these three points together, the conclusion is clear: GEO is not a multiple-choice question about whether to do it or not, but a time question about whether to do it early or later.

2. What can be done: Write the content to the “robot that solves problems for people”

After taking over, Director Ruoyun of Today Weekly made what I think is the most critical mental change: in the past, we wrote content for people to read, and it had to be beautiful, have a story, and have emotions; but GEO is writing for robots, and this robot is solving problems for a living person.

This sentence sounds simple, but it requires a complete overhaul of the writing method. I organized what she said into four actionable actions:

1. **Write articles using the logic of a TV series. ** The crawler era captures keywords, while the AI ​​era captures answers segment by segment. Therefore, each paragraph must be able to be picked out independently and form a complete unit; but when all are combined, it will be a good-looking TV series for humans. The unit dramas are independent and the series are integrated into one - I like this metaphor very much.

2. **Change the subscript into a question. ** In the past, we loved subscripting adjectives (he finally succeeded after many twists and turns), but AI doesn’t know how to ask questions about an adjective. Change it to “How can we make GEO better?” This kind of question-style subscript, and the answer is given directly below, can only be grasped by AI and can be understood by humans.

3. **Draw the target first, then shoot. ** This is the most useful sentence to me in the whole scene. Most people shoot arrows first and then draw targets - they go back and test the old articles they have written to see if they can be searched on AI. The actual test is very difficult. The correct sequence is the reverse: first break down who your audience is, what problems they will encounter, and how they will ask AI. Use these questions as targets, and then write them into articles. The case of the Dragon Boat Festival gift box she cited is very convincing: clear questions are used as the big tags, and each brand is used as the small tags. As soon as the article is put on the shelves, it is immediately caught by multiple AI questions and disappears after being removed from the shelves. Operable and verifiable, this is the method.

4. **Use five whys to identify the most urgent pain points. ** Asking all the way down: Does anyone ask about AI on this topic? Are there real pain points behind it? Is this pain point just needed? Is it painful enough that he would take the initiative to ask the AI? The last question is the most important - does this have anything to do with my product? The fifth why is actually conversion. The end point of writing GEO articles is never exposure, it is for people to remember you, come back to you, and finally close the deal.

3. Don’t forget authority: if the structure is right, someone will endorse it for you.

If you only remember the above paragraph, you will think that you will win if you adjust the structure of the article correctly. But Ruoyun reminded a second half that is easily overlooked: E-E-A-T, which is authority.

She said it very honestly: Your own official website is of high quality and has sufficient digital traces. These are necessary conditions, but they are no longer sufficient. What really widens the gap is third-party authority—academic papers have big bonus points but the threshold is high, Wikipedia can be used but has rules, LinkedIn is helpful for personal brands, and financial media like Today Weekly, which has been in business for 20 to 30 years and has extremely high domain trust, is a lever that a few ordinary brands can really leverage. This also explains why the domain of old media has suddenly become so valuable in the AI ​​era - its reputation was built over twenty years and with the expertise of a whole group of journalists, and it cannot be learned or bought.

But here’s the detail where I think it’s important to draw the line: authority must be real and work both ways. The person being quoted must have a real person and a digital trace, and that trace must include you; more importantly, it cannot be just the other person mentioning you, your official website must also link back and expose each other. Let the robot confirm that the relationship is not a one-sided advertisement, but a real connection with two-way evidence. Many people spend money to buy a promotional article and think it’s over. In fact, that’s only half the job.

Write at the end: I plan to do this

The greatest value of this sharing to me is that it turns anxiety into a list. If you, like me and Brother Chao, have bet on content that cannot grow assets in the past, then I will make up for it like this:

• Build your home first: Move the core content back to your own official website, so that each article can grow on the soil that will accumulate domain weight, instead of just gaining applause in the community.
• Rewriting method: For new articles, always draw a target first - first think about how the audience will ask AI, and then write it using a series structure with question-type subscripts and paragraphs as answers.
• Pick three urgently needed questions to start with: Use five reasons to screen out the pain points that will really be actively searched for and linked to my product. Write an article for each pain point, rather than writing what comes to mind.
• Build two-way authority: Take stock of third parties (media, experts, platforms) that I can cooperate with and that are truly reciprocal, so that the references are two-way connections and mutual support.

Search has changed its owner, but what it wants remains the same—truth, usefulness, and trustworthiness. GEO just asked this thing again in a way that the machine can understand. The earlier you start, the sooner you will be seen.

---