Batch Interview Analysis Workflow

Product Management OS feature — This workflow is included in the Product Management Operating System. Get the Product Management OS →

Saves 7-10 hours of manual synthesis

"10 interviews. 15 minutes. Done."

What you'll create

Transform a batch of user interviews into a complete research package:

OutputWhat it contains
Interview SnapshotsStructured analysis for each interview — JTBD, opportunities, key quotes
Research SynthesisThemes, patterns, contradictions across all interviews
Data-Backed PersonasPersonas grounded in actual research, not assumptions
Opportunity TreePrioritized opportunities with evidence strength

Example outputs

Interview Snapshot (one per interview)

# Interview Snapshot: Sarah Chen

**Date:** 2026-01-15
**Role:** Senior Product Manager at Acme Corp
**Context:** Exploring competitive research workflows

## Jobs-to-Be-Done
- "When preparing for a product review, I need to understand competitor
   positioning, so I can defend our roadmap decisions"
- "When a competitor launches a feature, I need to quickly assess impact,
   so I can advise leadership"

## Opportunities Identified
| Opportunity | Evidence | Notes |
|-------------|----------|-------|
| Need faster competitive intel gathering | Strong | Mentioned 3x, high frustration |
| Need to share competitive context with sales | Medium | Secondary pain point |

## Key Quotes
> "I spend half my Fridays just trying to figure out what competitors shipped"
> — on competitive monitoring

Research Synthesis (across all interviews)

# Research Synthesis: Competitive Research Workflows

**Sources:** 8 interviews
**Date:** 2026-01-20

## Top Themes

### 1. Time sink on monitoring (6/8 interviews)
PMs spend 2-4 hours weekly just tracking competitor activity...

### 2. Sharing intel is broken (5/8 interviews)
Competitive knowledge stays siloed in product managers' heads...

## Prioritized Opportunities
| Opportunity | Evidence | Impact | Confidence |
|-------------|----------|--------|------------|
| Automated competitor monitoring | Strong (6/8) | High | High |
| Shareable competitive briefs | Strong (5/8) | Medium | Medium |

How to run this workflow

Step 1: Add your interviews

Open the workflow's input folder:

workflows/batch-interview-analysis/inputs/

Drop your interview transcripts here.

Supported formats:

  • Text files (.txt)
  • Markdown files (.md)
  • Any plain text format

Naming tip: Name files descriptively so outputs are easy to identify:

  • ✅ interview-sarah-pm-acme.md
  • ❌ notes-1.txt

Step 2: Run the workflow

In Claude Code, navigate to your Product Management OS folder and run:

claude

Then tell Claude:

Run the batch interview analysis workflow

Pro tip: Add context about what you're exploring:

Run the batch interview analysis workflow.
I'm exploring how product managers handle competitive research.

Step 3: Answer Claude's question

Claude will ask: "What were you trying to learn from these interviews?"

This focuses the analysis. Examples:

  • "We're exploring why enterprise users churn after trial"
  • "We're investigating onboarding pain points"
  • "Find the top themes across all research"

Step 4: Review outputs

Find your results in:

workflows/batch-interview-analysis/outputs/

You'll see:

  • interview-snapshot-[name].md — One per interview
  • research-synthesis.md — Cross-interview analysis
  • personas.md — Data-backed personas
  • opportunity-tree.md — Prioritized opportunities

Best results tips

  1. Include full transcripts — The more detail, the better the analysis
  2. Add research context — Tell Claude what you were exploring
  3. Name files clearly — Helps track which snapshot came from which interview
  4. Review and edit — AI analysis is a starting point, not the final word

If something doesn't look right

Tell Claude what to change:

  • "The themes are too broad — break into more specific clusters"
  • "You missed insights about pricing from the third interview"
  • "Add a section comparing enterprise vs. SMB users"
  • "The evidence strength ratings seem off"

Claude will revise. You can keep iterating until it's right.

Troubleshooting

ProblemFix
"Claude only processed some files"Check files are in the inputs/ folder, not a subfolder
"Synthesis is too surface-level"Be more specific about your research question in Step 3
"Quotes attributed to wrong person"Make sure file names include interviewee name
"Missing opportunities I expected"Ask Claude to look specifically for that topic

Framework reference

This workflow uses Teresa Torres' Continuous Discovery Habits:

  • Opportunities are framed as user needs, not solutions
  • Evidence strength helps prioritize what to act on
  • Opportunity trees connect research to product decisions

What's next?

After running this workflow, you might want to:

Previous
Overview
Next
Using Agent Teams