Interview Transcript Cleanup

Clean up raw interview transcripts while preserving meaning and speaker voice.

When to Use

After conducting user interviews when you have raw transcripts that need cleanup before analysis.

The Process

1. Read the full transcript to understand context and flow

2. Remove noise without losing meaning:

   • Filler words (um, uh, like, you know)
   • False starts (incomplete/abandoned sentences)
   • Redundancies (repeated words that don't add meaning)

3. Preserve what matters:

   • Original meaning and intent
   • Speaker's unique voice and style
   • Speaker labels (Interviewer:, Interviewee:)
   • Unclear/inaudible markers [inaudible]

4. Light editing only:

   • Fix punctuation and capitalization
   • Don't over-formalize casual language
   • Don't cut unnecessarily

Quality Check

• Does the cleaned transcript read naturally?
• Is the original meaning preserved?
• Can you still hear the speaker's voice?
• Are all speakers clearly labeled?

When Not to Use

Do not use this skill when the request is unrelated, low-stakes, or better handled by a simpler direct response.