What this case covers
The author, Tun Xiaozu, uses Workbuddy and self-iterating prompt V5.0 to operate vertical small accounts on workplace-related WeChat Official Accounts. The article reviews the complete process from fragmented information to prompt iteration, formatting, and mass sending.
Questions it answers
- What changed in the prompts from V3.0 to V5.0?
- Why was it only after writing 10 articles that the lack of human feel was discovered?
- How can a style model be built using articles from benchmark accounts?
- How can formatting and mass sending be automated?
- What pitfalls should be noted when downloading benchmark articles?
Our observations
- Topics are taken directly from benchmark viral hits, which is more reliable than coming up with topics oneself.
- Fundamentals determine the upper limit of AI output; tools are not the key.
- The public-domainization of WeChat Official Accounts gives small accounts a baseline read count.
From the author
A flash of inspiration: I can have AI, before each creation, first search for excellent cases related to the title based on the title, and then do original writing.