Stats
Actions
Tags
From rz-ux-researcher
Synthesize themes from multiple user interview transcripts. Use when the ticket provides N interview transcripts (typically 3+) and asks what common threads, pain points, or jobs-to-be-done emerge.
How this skill is triggered — by the user, by Claude, or both
Slash command
/rz-ux-researcher:interview-synthesisThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Invoke when the ticket:
Invoke when the ticket:
For persona creation, use persona (often follows synthesis). For mapping the flow users described, use journey-map.
# Interview synthesis — {topic}
**Date:** {YYYY-MM-DD}
**Target app(s):** {app_id}
**Triggering ticket:** [CAR-{n}]({URL})
**N interviewed:** {number}
**Segments:** {e.g., "4 early SIA users + 3 website chatbot users"}
**Confidence:** {low/medium/high}
## Methodology
- **Collection method:** Interviews, Zoom / in-person / async
- **Protocol:** Semi-structured / structured / open
- **Duration:** {min–max} per interview
- **Recording:** Transcribed by {tool}
- **When:** Date range of interviews
- **Known biases:** Self-selected respondents? Beta users only? Power users? Name the sampling bias explicitly.
## Themes (ranked by strength)
### Theme 1 — {short phrase}
**Confidence:** {low/medium/high}
**N supporting:** {count} of {total}
**Supporting quotes:**
> "{anonymized direct quote}" — {segment}, {date}
> "{second quote, different person}" — {segment}, {date}
> "{third quote}" — {segment}, {date}
**Synthesis:**
One paragraph explaining what the theme means. Not a restatement of quotes — the interpretation.
**Implications:**
- For Designer: …
- For PM-lite / Riché: …
---
### Theme 2 — {short phrase}
…
### Theme 3 — {short phrase}
…
(Aim for 3–6 themes. More than 6 is usually not synthesis; it's a transcript dump.)
## Discarded themes
Themes that appeared but didn't hold up on examination. One line each on why.
- **{Theme}** — only 1 person; not a pattern.
- **{Theme}** — contradicted by behavioral data.
- **{Theme}** — specific to prior product version we've since changed.
## Surprises
What did you NOT expect going in? Surprises are often the highest-signal finding.
## Gaps
What we still don't know after this synthesis. What the next round of research would need to cover.
## Anonymization note
All direct quotes have been stripped of: names, company names, emails, phone numbers, and other PII. If any quotes reveal segment in a way that identifies a specific user, they've been further generalized.
## Sources
- Interview IDs or Notion page links
- Date range: {start} – {end}
- Interviewer(s): {if applicable}
## Confidence: {low/medium/high}
Rationale: {e.g., "medium — themes consistent across segments, but small N (n=7)"}
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub rczamor/rz-agent-team --plugin rz-ux-researcher