Small teams with visible leaders or subject-matter experts
Keep expert voices consistent without flattening them.
Standardize how AI helpers represent key voices, review public posts, and prepare engagement suggestions.
The gap
What this should prevent
Team AI helpers blur personal voice, brand voice, and generic support tone.
Reviewers need to know which source justified a suggestion before trusting it.
Every new workflow recreates voice rules because the context is not portable.
Workflow
How the spec becomes useful
Start with one voice
Build a single source-backed spec before introducing shared team rules or multiple reviewers.
Add review surfaces
Route draft suggestions, engagement opportunities, and export artifacts through explicit human checkpoints.
Reuse the context
Feed the same voice system into skills, content inboxes, reply workflows, and future internal agents.
Artifacts
What you get out of it
Voice registry
Separate specs for people and brand surfaces, with boundaries and source context attached.
Review queue
A workflow for draft ideas and reply opportunities that keeps approval explicit.
Agent context pack
Reusable files and prompts so each tool does not invent its own version of the same person.
Starting signal
Sources this use case cares about
Public social profilesTeam Slack laterDocs and changelogs laterGitHub and support later
Prompts
Examples this should unlock
Draft this reply as the founder, not the brand account.
Show why this engagement opportunity is worth reviewing.
Export the current voice rules for a support agent prompt.