Claude Skill · User Research Suite
Executing User Research
An AI skill that helps you run interviews, usability tests, and contextual sessions —
capturing structured, synthesis-ready data at every step.
What this skill does
Once you have an approved research plan, this skill helps you move into fieldwork.
It coaches facilitation, structures your note-taking in real time, and ensures
every session ends with clean, tagged data that's ready to hand off to synthesis —
without extra cleanup.
It doesn't replace your judgment as a researcher. It gives you frameworks, prompts,
and templates so you can focus on the participant instead of remembering procedure.
How to trigger it
Just describe what you're trying to do. Here are example prompts that activate this skill:
Pre-session
"My first participant is in 30 minutes. Help me prepare."
Facilitation
"How do I run a think-aloud test without leading the participant?"
Live session
"The participant just went silent — what do I say?"
Note-taking
"Tag these session notes and extract the key moments."
Post-session
"I just finished 6 interviews. Help me organize everything before synthesis."
Surveys
"Set up a survey to run alongside my interviews — same research questions."
Research session workflow
Prepare
Run through the pre-session checklist — logistics, consent, recording tools, prototype access, file structure. Claude will flag anything missing.
Facilitate
Run the session using the 5-Act structure. Claude suggests neutral probes, helps you avoid leading questions, and reminds you when to follow curiosity over script.
Capture
Take notes using a shared tagging taxonomy — [OBS], [PAIN], [DELIGHT], [QUOTE], [NEED] — so all sessions produce consistently structured data.
Archive (within 24h)
Clean the transcript, apply tags, write the top 3 moments, flag cross-session quotes, and archive recordings with consistent naming. Memory decays fast.
Hand off to synthesis
All outputs — transcripts, tagged notes, observer logs, survey CSVs — are organized and ready. No extra cleanup before analysis begins.
Facilitation methods
The skill applies four proven frameworks depending on your session type. Tell Claude what you're running and it applies the right one.
Default
5-Act Interview Structure
A structured arc — welcome, context, prototype intro, core tasks, debrief — that creates psychological safety while keeping sessions on track.
Use for: most interviews and usability tests
Usability
Think-Aloud Protocol
Participants narrate their thoughts in real time. You observe and probe neutrally, capturing hesitation, backtracking, and surprise.
Use for: prototype and product testing
Field
Contextual Inquiry
Observe users in their real environment — you're the apprentice, they're the master. Behavior in context reveals more than recall in a meeting room.
Use for: workplace and home-environment studies
Depth
Laddering
Ask "why" 3–5 times to move from surface behavior to underlying motivation. Stops when answers become abstract values like freedom or control.
Use for: motivation and attitude research
Note-taking tag system
Every session uses the same tags. This makes notes synthesis-ready before you've left the call —
clusters form naturally instead of requiring a separate sorting pass.
| Tag |
What to mark |
| [OBS] |
Direct observation
What happened — verbatim or near-verbatim. No interpretation.
|
| [QUOTE] |
Participant's exact words
Use sparingly — only for striking, memorable statements worth pulling into a report.
|
| [PAIN] |
Friction or frustration
Moments of confusion, failure, hesitation, or expressed frustration.
|
| [DELIGHT] |
Surprise or success
Moments of unexpected pleasure, ease, or positive reaction.
|
| [NEED] |
Expressed or latent need
What they want the product or world to do — stated or implied.
|
| [Q] |
Follow-up question
Something you wanted to probe but couldn't in the moment. Revisit in debrief or next session.
|
| [INSIGHT] |
Your interpretation
Clearly marked as analyst voice — never mixed with raw observation.
|
File and folder structure
Create this structure before your first session. Claude will help you name and organize files consistently across the study.
research-sessions/
├── session-[participant-id]-[date]/
│ ├── raw-transcript.md
│ ├── session-notes-tagged.md
│ └── observer-log.md
├── survey-responses/
│ └── survey-responses-[study-name]-[date].csv
└── recording-archive/
└── recording-[participant-id]-[date].mp4
Naming matters at scale. With 8+ sessions across 3 researchers, inconsistent file names create chaos at synthesis. Claude enforces this structure automatically when connected to Google Drive.
Connected tools
Connect these tools to Claude to unlock deeper automation — scheduling, recording, transcription, and archiving in one flow.
What you'll have at the end
By the time all sessions are complete, this skill ensures you have everything synthesis needs — nothing missing, nothing to reformat.
- Raw transcripts — one per session, auto-generated or manual, language preserved
- Tagged session notes — inline tags, 3 top moments, open follow-up questions
- Observer logs — timestamped, separate from facilitator notes
- Survey response CSV — with open-text fields tagged to match interview data
- Recording archive — named by participant ID and date, organized by study