What this case covers
Ji Zhong hosted an AI practice roundtable at Shengcaiyoushu with Yin Xinghe, Lin Huanyu, and Lu Shuye, three community members applying AI in content businesses. The article documents their workstation, content factory, and Agent employee practices, along with remarks from the live discussion and replication materials.
Questions it answers
- How can a Xiaohongshu workstation bring topic selection and publishing into one place?
- How did Bilibili product-video production go from one video in two weeks to one in an hour?
- How were more than 30 affiliate accounts assigned to different Agents?
- Where should someone without a project begin learning AI?
- How can participants get help with errors while reproducing the workflows?
Our observations
- What the three cases share is not a tool, but identifying the most painful repetitive step first.
- AI takes on execution; people still define goals, examples, and acceptance criteria.
- Content can be reused across formats, but deciding what will resonate still needs human judgment.
From the author
Translated from Chinese: “Think in an AI-native way: find the task that takes the most time and repeats most often, clarify its inputs, outputs, and judgment criteria, then hand it to AI.”