How Do You Analyze Customer Interview Transcripts?
Upload each recording or transcript to one place, keep the participant's words exactly as spoken, and let Deck pull out the insights. Each insight links to the quote that supports it, and insights group into themes across all your interviews. You then read patterns across calls instead of rereading every transcript before a planning meeting.
Why Interview Insights Get Lost After the Call
Interview notes end up in docs, recordings sit in a drive, and the best quotes live in one researcher's head. A month later, nobody can say how many customers raised the same problem. So decisions lean on the last call someone remembers, not on everything customers said.
How Deck Turns Each Interview into Evidence
Deck transcribes each recording, extracts insights, and links every insight to the verbatim quote behind it. Insights join your existing themes or start new ones, next to support tickets, surveys, and other feedback. When you open an interview, Deck highlights the transcript passages connected to each insight.
How to Analyze Interview Transcripts in Deck
Run it yourself in Deck, or hand it to your AI agent. Same steps, same result.
- 01
Pick the Right Upload Path
Recordings and transcripts go through Manage → User Interviews → Add Interview. Many transcripts in one table go in as a CSV of the type User Interview. Sales calls in Gong and interview notes in Notion come in through those integrations.
- 02
Check What Deck Already Has
Look through Manage → User Interviews for the participant before you upload, so nothing is processed twice. Interview processing uses credits, so for a batch of more than 10, check Settings → Billing first.
- 03
Clean Each Transcript
Put one speaker turn on each line as “Speaker Name: spoken text” and keep the labels consistent. Remove timestamps, join notices, and recording disclaimers. Never summarize or change what the participant said.
- 04
Upload the Recording or Transcript
Choose Interview Video for MP4, MOV, MP3, WAV, and other recordings up to 500 MB each. Choose Interview Transcript to upload a .vtt, .txt, or .docx file or paste the text. A transcript needs at least 100 words and 10 lines.
- 05
Fill In the Interview Details
Add the Interview Name and Interview Date, which are required, and the User Email if you have it. You can pick a segment here too. Then submit.
- 06
Review the Insights
A transcript takes about 2 minutes and a recording about 5. Open each processed interview to read its insights, the highlighted quotes, and the themes they joined.
Import Interviews with Your AI Agent
Prefer Claude, ChatGPT, Cursor, or another MCP client? Paste this prompt. Your agent runs the same workflow through Deck MCP and checks with you before it changes anything.
You need
- Deck MCP connected to your agent
Your agent starts with list_sources.
Agent Prompt
Paste This into Your Agent
Use Deck MCP. Fetch and follow https://docs.getdeck.io/docs/agents/workflows/import-interviews.md
Interview Transcript Prep Checklist
## For each interview- Interview Name: [Participant or session name]- Interview Date: [YYYY-MM-DD]- User Email: [optional]- Segment: [optional]## Transcript format- Interviewer: [spoken text]- Participant: [spoken text]## Before upload[ ] Not already in Manage → User Interviews[ ] One speaker turn per line, as Speaker Name: spoken text[ ] Speaker labels consistent (Interviewer and Participant when names are unknown)[ ] Timestamps, join notices, and recording disclaimers removed[ ] Participant's words unchanged, nothing summarized or added[ ] At least 100 words and 10 lines, at most 50,000 words[ ] Recordings: a supported format, 500 MB or less## After processingInterview | Status | Insights | Top themes-----------------|-------------------|--------------|-----------------[Interview Name] | Complete / Failed | [Deck count] | [Theme], [Theme]