# Uxr Interview Debrief

> Turns one research session into same-day debrief notes with participant id only, verbatim quotes with timestamps by guide arc, observations kept apart from interpretations, surprises and questions for the next session. Use for "run uxr-interview-debrief", "debrief this interview", "structure my session notes", "here is today's transcript", "I just got off a user call", "notes are a wall of text", "same-day debrief", part of the UX Research with Claude Pack by 90n.

- Skill: `polar-bear-org/uxr-interview-debrief` (Agent Skill)
- Install (CLI): `npx get-skill add polar-bear-org/uxr-interview-debrief`
- Raw SKILL.md: https://api.skillmd.com/api/skills/polar-bear-org/uxr-interview-debrief/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: polar-bear-org (https://skillmd.com/u/polar-bear-org)
- Updated: 2026-11-14
- Page: https://skillmd.com/skills/polar-bear-org/uxr-interview-debrief

---


# Interview Debrief Notes

## When To Use
You just finished three calls in a row and the notes are a wall of text. Run this the same day for each session, before memory turns what was said into what you expected to hear. It answers: what did this one participant say and do, what surprised us, and what do we ask next time?

## When Not To Use
If all the sessions are done and you need patterns across them, use Thematic Analysis; a debrief covers one session and never names themes. If the session was a usability test and you need issues by task, use Usability Test Findings.

## Inputs
- The transcript, typed notes or both from one session, with the date and method
- The interview guide (or its arcs) and the guide version used
Anonymise first: replace names with P1, P2, remove contact details, employers and anything that identifies a person.
If you have none of this, I start from your raw notes for one session and mark the output as a first draft.

## Approach
The GOV.UK Service Manual pages Analyse a research session and Taking notes and recording user research sessions: debrief straight after each session, one observation per note, and keep what was seen and heard apart from what you think it means. The judgment: raw words stay raw, because "it's fine, I've stopped noticing it" and "it's fine" are different findings. The failure it prevents: a paraphrase in quotation marks that reaches the readout as a user quote nobody said.

## Workflow
1. Ask three questions: which session this is (participant id, date, method), which guide version was used, and who observed.
2. Header and evidence base: participant id only, date, method, guide version, observers, and "evidence base: one session". If a name, employer or contact detail arrives, I leave it out and ask you to remove it from the chat.
3. Key quotes verbatim with timestamps, grouped by the guide's arcs. Their exact words; a gap in the transcript is marked "[inaudible, hh:mm]", never filled.
4. Observations, one per line: what they did or said, as seen or heard. Interpretations go in their own list, each marked "interpretation" and linked to the observation it reads.
5. Surprises: anything that broke an expectation, with the assumption or research question it touches.
6. Next session and gaps: questions to add or change in the guide, things to verify, and gaps in the record (recording failed, notes only, observer joined late).

## Output Format
```markdown
# Interview Debrief Notes
**Participant:** [P1] | **Date:** [date] | **Method:** [interview / contextual visit / other] | **Guide version:** [v] | **Observers:** [roles]
**Evidence base:** one session; [transcript and notes / notes only]
## Key quotes by arc
| Arc | Timestamp | Quote (verbatim) |
|---|---|---|
| [arc from guide] | [hh:mm] | ["exact words"] |
## Observations
- [O1] [what P1 did or said, as seen or heard]
## Interpretations
- [I1] interpretation of [O1]: [what it might mean]
## Surprises
| Surprise | Assumption or research question it touches |
|---|---|
| [what broke an expectation] | [assumption or RQ] |
## Follow-up
- For the next session: [question to add or change]
- To verify: [item] | Gaps in the record: [what is missing]
## Decision
[Moderator] decides by [date, before the next session] which guide changes to make; debriefs go to [analysis owner].
```

## Done When
- The header holds a participant id and no name, employer or contact detail
- Every quote is verbatim with a timestamp, and every gap is marked rather than filled
- Observations and interpretations sit in separate lists, and no theme is named

## Quality Bar
- One session per debrief, written the same day
- No paraphrase inside quotation marks, and no merged or sharpened quotes
- No comment on the participant as a person; notes describe what happened
- Quotes stay verbatim and traceable to a session; Claude never fills a gap in the notes

## Next
Run uxr-thematic-analysis (Thematic Analysis) to code all the debriefs together.

## About the makers

This pack is made by 90n ([90n.ai](https://90n.ai)): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).

