Patterns: Skip the hands
Also Known As
Direct to AI, Voice Channel
Context
Remote mob working with AI agents. The talker dictates exact words for the typist to type into AI prompts. This is slow and wastes human effort on transcription.
Purpose
Remove the typist as human transcriber. Let the talker speak directly to the AI at natural speed.
How To
- Typist sets up voice transcription tool (Wispr Flow, etc.) and uses speakers so the tool can pick up remote voices
- Talker speaks their prompt naturally (Talker -> mic -> video conf -> speakers -> mic -> transcribe)
- Typist clicks to capture and insert transcription into AI prompt
- Mob reviews before sending (good practice)
Consequences
Much faster prompting. Talker can speak at natural pace instead of dictating slowly. Typist freed from transcription duty but still controls when to capture.
Preserves “everybody hears the intention, everybody in the know.”
Can extend to capture multi-person discussion and feed to AI.
Requires audio setup where transcription picks up remote voices from speakers.
Collaborators
Narrate the Agent - Related pattern for making AI work visible to mob
Mind Dump - Voice-to-AI pattern for individual work