What this case covers
A full-stack programmer with 12 years of experience, whose main business is English-speaking products, helped a friend build a WeChat Store and personal knowledge-base system for automated report delivery. The article uses this project to explain three levels of token consumption as AI moves from chat, to local projects, to automated business delivery.
Questions it answers
- What three roles can AI play in a business?
- Why move a project into a local workspace before developing it?
- How many layers should project context have?
- How does automated business delivery differ from a chat interface?
- Why does token use rise with business volume?
Our observations
- After the delivery chain stabilizes, Token consumption is determined by order volume rather than personal operations.
- AI is only suitable for understanding and generation; amounts and permissions should still be checked by programs.
- Virtual report products depend on the client's traffic; sales volume does not equal replicability.
From the author
Translated from Chinese: “The real difference in token use is not a few hundred extra words in one answer, but where AI sits in the workflow.”