I record interviews and team calls pretty often. The raw transcripts that come out are rough — filler words everywhere, speakers mislabeled half the time, zero structure. These three prompts turn that mess into something I can actually search, pull quotes from, and turn into action items.
[Claude]
You are my transcription cleanup editor. Given the raw transcript of a 45-minute client interview, rewrite it into a structured brief: a 3-line summary, a list of every concrete commitment or action item with the person who owns it, and the 5 strongest direct quotes with timestamps. Keep quotes verbatim. Output as a markdown document with clear headers.
[ChatGPT]
Here is the transcript of a 90-minute product demo call. Extract every feature the customer asked for, then group them into Must Have, Nice to Have, and Not Needed. For each, note the exact words they used and the timestamp. End with a ranked priority list of the top 5 features to build next. Output as a table.
[Cursor]
I have a folder of 12 interview transcripts as .txt files. Write a Python script that reads all of them, strips filler words like um and uh, merges repeated speaker turns, and exports one combined markdown file with a speaker index and a keyword search index. Run it and show me the output.
More prompt packs live in the archive at aitoolalliance.com/tag/prompt-pack.