Assisted repurposing pipeline (L2)
One source asset becomes the social, email, and enablement versions through a shared assistant workspace loaded with the voice file. Humans approve every output. The gain is coverage, not volume for its own sake.
The steps
- 01
Write the voice file
Tool: Notion
Banned words, sentence length, claim rules, examples of good and bad. Without a voice file, repurposing produces generic output at scale, which is worse than no output. Owner: brand or content lead. DoD: voice file exists with at least five bad examples.
- 02
Define the output set per source
Tool: Manual
For each source type, name the exact derivatives: three social posts, one email, one enablement one pager. Fixed sets prevent endless generation. Owner: content lead. DoD: derivative map published.
- 03
Generate from the source, never from a derivative
Tool: Claude
Every derivative is generated from the original asset. Generating from a summary of a summary is how claims drift away from evidence. Pitfall: chained rewrites that quietly change a number. DoD: generation logs show the source asset each time.
- 04
Cap output at review capacity
Tool: Manual
Never produce more than the review queue can approve in a week. Backlogged AI content is the fastest route to something wrong going out unreviewed. Owner: content lead. DoD: weekly cap set and enforced.
Tools in this playbook
Next playbooks
Unfamiliar terms are defined in the AI and Revenue Dictionary. Related frameworks live in the framework library.