user-interview

User Interview Processing Workflow

Systematically process user interviews to extract actionable insights using Claude Code.

Quick Start

/user-interview

Just finished user interviews? Drop your transcripts and I'll extract insights.

Tell me:
1. How many interviews do you have? (paste transcripts or point me to files)
2. What was the research goal? (what were you trying to learn?)
3. Any specific hypotheses you were testing?

I'll cross-reference with your existing research, extract pain points,
feature requests, JTBD insights, and generate a shareable report.

Output: thoughts/shared/pm/interviews/[date]-interview-insights.md

Scope Boundary: This Skill vs /user-research-synthesis

Use /user-interview when:

• Processing 1-3 interviews from a single research session
• Extracting quotes, themes, and action items from individual conversations
• Doing a quick post-interview debrief (same day)
• Output: Per-interview insight cards with VALIDATED/CHALLENGED/NEW theme labels

Use /user-research-synthesis when:

• Synthesizing 4+ interviews into a unified research report
• Looking for cross-interview patterns, personas, or journey insights
• Applying advanced frameworks (Mom Test validation, Five Whys root cause, affinity mapping)
• Output: Research report with theme hierarchy, evidence strength, and strategic recommendations

Handoff: After processing 3+ individual interviews with /user-interview, suggest: "You now have [N] processed interviews. Want to run /user-research-synthesis to find cross-interview patterns and build a research report?"

Context Routing Logic (Internal - for Claude)

Automatic Context Checks: When this skill is invoked, immediately check:

| Source | What to Check | Priority | | -------------------------------------------- | ----------------------------------------------- | -------- | | thoughts/shared/pm/research/ | Previous interview syntheses, existing themes | High | | thoughts/shared/pm/prds/ | Active PRDs that this interview relates to | High | | thoughts/shared/pm/context/stakeholder-template.md | If interviewee is a known stakeholder/customer | Medium | | thoughts/shared/pm/context/business-info-template.md | Company context, product positioning | Medium | | thoughts/shared/pm/launches/ | Recent launches the interviewee might reference | Low |

Context Priority:

1. Existing research and validated themes FIRST
2. Active PRDs and problem statements SECOND
3. Business context and positioning THIRD
4. Recent launches and stakeholder info FOURTH

Cross-Skill Links:

• Before interviews → /interview-guide for question preparation
• After processing → /user-research-synthesis for deeper analysis
• If insights inform a feature → /prd-draft to reference findings
• If insights reveal strategic shifts → /write-prod-strategy

Overview

Tools needed: Claude Code (primary), research repository (optional) When to use: After conducting 3+ user interviews

Workflow

Step 0: Connect to Existing Research

Before processing new interviews, check existing context:

1. Search thoughts/shared/pm/research/ for previous interview syntheses and findings
2. Identify validated themes from prior research
3. Note which PRDs or strategic initiatives these interviews relate to
4. Check thoughts/shared/pm/context/business-info-template.md for product positioning context

For each theme that emerges in the new interviews:

• If it matches a previous finding: "This validates theme X from [previous research file]. Now supported by [N total] interviews."
• If it contradicts a previous finding: "This challenges theme X from [previous research file]. Previous research said [Y], but this interview suggests [Z]. Needs further investigation."
• If it's entirely new: "This is a NEW theme not seen in prior research. First surfaced in this batch. Recommend probing in future interviews."

Show the PM a brief summary:

## Existing Research Context
- Found [N] previous interview syntheses in thoughts/shared/pm/research/
- Validated themes from prior research: [list]
- Active PRDs from thoughts/shared/pm/prds/ this research relates to: [list]
- New themes to watch for: [will identify during analysis]

Step 1: Transcribe

Option A: transcription tool (Recommended)

• Upload recordings
• Auto-transcribe
• Download as text files

Option B: Your recording tool

• Zoom, Meet, Teams all auto-transcribe
• Download transcripts
• Clean up formatting

Option C: Claude

• Upload audio files directly
• Ask for transcription
• Save outputs

Step 2: Load Transcripts into Claude Code

1. Place all transcript files in a folder
2. Start Claude Code in that directory
3. Tell Claude: "I have [X] interview transcripts to analyze. Please read all the files in this folder."

Step 3: Extract Insights

Run these queries in order:

Query 1: Pain Points
"Analyze all transcripts. What are the top 5 pain points customers mentioned?
For each:
- How many customers mentioned it
- Specific quotes
- Severity (how painful)
- Current workarounds"

Query 2: Feature Requests
"What features did customers request?
For each:
- Frequency
- User quotes
- Underlying need
- Priority (must-have vs nice-to-have)"

Query 3: Jobs to Be Done
"For each major use case, what job is the customer trying to accomplish?
Include:
- The task/goal
- Why it matters to them
- What triggers the need
- What success looks like"

Query 4: User Segments
"Do different user types express different needs?
Identify:
- 2-3 distinct user archetypes
- Unique needs per archetype
- Common needs across all"

Query 5: Surprises
"What was surprising or unexpected in these interviews?
What assumptions were challenged?"

Step 4: Create Output Doc

Ask Claude:

"Create an executive summary covering:
- Top 3 insights
- Top 5 pain points (ranked)
- Top 3 feature requests
- Recommended next steps
Format as a report I can share with my team.
Save to thoughts/shared/pm/interviews/[date]-interview-insights.md"

Deep Dive

For more thorough analysis, add these steps:

Step 5: Sentiment Analysis

"Analyze sentiment across all transcripts:

For each user:
- Overall sentiment (positive/neutral/negative)
- What they love
- What they're frustrated by
- Quote that captures their feeling

Then: Overall sentiment trends across all users"

Step 6: Opportunity Scoring

"Based on these interviews, score each opportunity:
[paste pain points and feature requests from Step 3]

For each opportunity, rate 1-5:
- Pain level (how much it hurts)
- Frequency (how often encountered)
- Reach (how many users affected)
- Strategic value

Recommend top 3 to pursue."

Step 7: Create Quote Bank

"Extract 15-20 powerful quotes from these interviews that:
- Illustrate key pain points
- Show user mental models
- Express emotional impact
- Could be used in PRDs or presentations

For each quote:
- The quote
- User context (role, company size, use case)
- What it illustrates"

Save these quotes for later use in PRDs, presentations, and marketing.

Step 8: Competitive Mentions

"What did users say about competitors?
- Which competitors were mentioned
- What users like about them
- What users don't like
- Why they chose us (or didn't)

Identify: What can we learn?"

Step 9: Workflow Analysis

"For each key workflow discussed:
- Step-by-step how users complete it today
- Where they get stuck
- What tools they use
- What they wish was different

Visualize: Current state → Desired future state"

Using research repository (If You Have It)

research repository is purpose-built for research synthesis:

Setup:

1. Create new project
2. Import transcrip