My Intern Turned In 1,000 Keywords in 20 Minutes. Then I Asked Him One Question.

Those Three Seconds

When the intern handed me the keyword list, I was at the water dispenser filling my cup. He called out, "Zhou, it's done." I walked over. The screen held more than a thousand keywords. The formatting was neat. The categories were clear. He had even color-coded the search-volume ranges.

I asked which of them had commercial value.

He stared at the screen. His hand came off the keyboard and rested on the edge of the desk. Three seconds. The computer fan was spinning. At the next desk, someone was typing on a mechanical keyboard, click-click-click. Then he said he would have it filter them again.

I set my cup on his desk. A little water splashed out. I told him not to filter. I told him to name one keyword he thought was worth money.

He looked for a while and pointed at the third row. I asked why. He said the search volume was big. I asked what came after that. He said a big search volume means people are searching. I asked what came after that. He stopped talking.

That night I went home and opened my computer to write a client proposal. The cursor blinked in the input box. I typed half a line, deleted it. Typed half a line again, deleted it. My hand rested on the mouse, and I opened the chat box. I asked how an SEO strategy framework should be written in the GEO era. Three seconds later a B-plus structure appeared. Clear logic. Neat sections. Subheadings. I copied it, pasted it, changed a few words, and sent it to the client. The client replied, "Received."

I closed the computer and felt that something was off. I could not say what.

The next week I wrote another proposal. This time I did not hesitate. I opened the chat box directly. The third week, the same. The fourth week, the client called. He asked where a certain number came from: the line claiming industry search volume had grown 400 percent in three years.

I opened the document and found the line. I opened the AI chat history. It said "according to industry reports" and did not name one. I searched for ten minutes and found nothing. The client said on the phone, "Zhou, you've been doing this for twenty years. You didn't used to make this kind of mistake."

Ash fell from my cigarette onto my pants. I did not brush it off.

That night I unplugged the wireless mouse and threw it into a drawer. In the drawer was a handwriting tablet. Its battery had leaked, and white and green corrosion had crusted across the surface. I found a sheet of A4 paper, took a pen, and sat there. The first line I wrote was "client's core needs." I looked at it for five minutes and crossed it out. The second line was "gaps in competitors' strategy." I wrote three words, and the pen tip pierced a hole in the paper. I turned the paper over and kept writing on the back.

Forty minutes. One page. Seven points. I crossed out three and kept four. I photographed it and sent the image to myself. The next morning I looked again. One of the four was useless. The other three could be used.

I opened the chat box, typed in the four points, and asked which one was most likely to draw a challenge from the client, and why. It replied with a paragraph and said the second point lacked data support. I told it to go find data, and to make sure the source could be traced. It found three. One would not load. One was from 2019. I spent twenty minutes searching manually, found a 2023 industry report, and checked it myself. Page 14, the numbers matched. I pasted the source link into the proposal.

This time the client did not call. He replied by email: "OK."

I started spending Wednesday evenings on the analytics dashboard. No tools, no generated reports. Just me, line by line, for two hours. The first hour was painful. My eyes hurt. I wanted to open AI and have it summarize everything. I put my phone in the living room to charge and kept looking. In the second hour I saw that search terms for one category were rising while our landing page was still two years old. I wrote it in a notebook. The next day I told the team to change it. Two weeks later, traffic for that category was up thirty percent.

Later the intern asked what I write on that sheet of paper before every proposal. I told him to go look. He flipped through the stack on my desk. The top sheet had "write first, ask later" in the upper left corner. The handwriting was messy. Next to it was a tear where the pen had cut through. He asked what it meant. I told him to write a garbage version first, then open the chat box. He said that sounded slow.

I said: last time you handed in a thousand keywords in twenty minutes, and when I asked you to screen them for commercial value, you couldn't. Spend twenty minutes writing ten keywords you think are worth money. Then let AI add more. See which is faster.

He tried it. Twelve keywords, five crossed out, seven kept. He opened the chat box and asked for more. AI gave him forty. He checked them one by one and kept nine. He told me four of his seven were not in what AI produced. I said those four are yours.

The day before yesterday he handed in a competitor analysis. I turned to page three and asked where a piece of data came from. He said he had checked the original report, page such and such, line such and such, and the link was in the appendix. I flipped to the appendix. Three links. Two worked. One was dead, but he had taken a screenshot. I closed the document and said fine.

This morning the stack of handwritten drafts was still on my desk. The top sheet was yesterday's core judgment for a client proposal: four points, one crossed out. Beside it sat the unplugged wireless mouse, under a layer of dust. The intern walked past, glanced at it, and did not ask. He was holding a pen and a sheet of paper.