Product Management
AI-prepared customer interviews
Run /customer-interview to get a research-backed interview guide with open-ended questions, follow-up prompts, and a framework for analyzing what you hear.
/customer-interview
Prepare customer interview guides with research-backed questions, follow-up prompts, and analysis frameworks. Then summarize insights from completed interviews.
Capabilities
Talk to customers better
Research-backed questions
The agent generates open-ended questions that avoid leading bias. Questions are tailored to your product area and research goals.
Interview guides
Complete interview guides with intro script, core questions, follow-up prompts, and closing. Ready to use or customize.
Insight extraction
After the interview, paste notes and the agent extracts key insights, quotes, and patterns organized by theme.
How It Works
How /customer-interview works
1
Type /customer-interview
Run the command and describe the research topic, customer segment, and what you're trying to learn.
2
Agent creates the guide
An interview guide is generated with warm-up questions, core topics, follow-up prompts, and a framework for analysis.
3
Conduct and analyze
Use the guide in your interview. Afterward, paste your notes and the agent extracts insights and patterns.
Try It
Example prompts
/customer-interview prep for onboarding experience research /customer-interview questions about collaboration workflows for enterprise PMs /customer-interview analyze these notes from today's call Full Skill Source
Use this skill in your project
Copy the full text below or download it as a markdown file. Place it in your project's .claude/commands/ directory to use it as a slash command.
--- name: interview-guide-builder description: Prepare customer interview guides with research-backed questions, warm-up structure, and synthesis templates. Use when planning user research interviews, customer discovery calls, or feedback sessions. --- # /customer-interview Generate a structured customer interview guide tailored to your research goals, with open-ended questions, follow-up prompts, and a synthesis template. ## What This Command Does 1. Clarifies the research objectives and target participant profile 2. Structures the interview into phases (warm-up, core exploration, wrap-up) 3. Generates open-ended, non-leading questions for each phase 4. Includes follow-up probes for deeper exploration 5. Creates a post-interview synthesis template for capturing insights 6. Saves the guide as a reusable document ## Usage ``` /customer-interview [research topic or goal] ``` **Examples:** - `/customer-interview Why do users abandon the onboarding flow after step 3?` - `/customer-interview Discover how engineering managers track team productivity` - `/customer-interview Validate demand for automated code review features` ## File Location **Location**: `nimbalyst-local/research/interviews/[topic-slug]-interview-guide.md` ## Execution Steps 1. **Define research objectives** Ask the user (or extract from the provided topic): - **Primary question**: What is the single most important thing to learn? - **Secondary questions**: What else would be valuable to understand? - **Participant profile**: Who should be interviewed? (role, experience level, usage patterns) - **Interview type**: Discovery (exploring a problem space), validation (testing a hypothesis), or feedback (evaluating a solution) If the user provides a brief topic, infer reasonable defaults and confirm. 2. **Select the question framework** Choose the appropriate questioning approach based on interview type: | Type | Framework | Focus | |------|-----------|-------| | Discovery | Jobs-to-be-Done | Understand what people are trying to accomplish and what gets in the way | | Validation | The Mom Test | Get honest signal without leading the witness toward your idea | | Feedback | Usability Probing | Understand reactions, confusion points, and mental models | **The Mom Test principles (apply to ALL interview types):** - Ask about their life, not your idea - Ask about specifics in the past, not hypotheticals about the future - Talk less, listen more - Never pitch during a research interview 3. **Generate the interview guide** Structure the guide into these phases: ### Phase 1: Warm-Up (3-5 minutes) - Build rapport and set expectations - Explain the purpose without biasing responses - Get consent for recording if applicable - 2-3 easy, non-threatening opening questions about their role and context ### Phase 2: Context Setting (5-7 minutes) - Understand their current workflow and environment - Map out the relevant tools, people, and processes they work with - 3-4 questions about their typical day/week related to the topic ### Phase 3: Core Exploration (15-20 minutes) - Deep dive into the primary research question - 5-7 main questions, each with 2-3 follow-up probes - Questions should move from broad to specific - Include "tell me about a time when..." prompts for concrete examples - Include "walk me through..." prompts for process understanding ### Phase 4: Pain Points and Workarounds (5-10 minutes) - Explore frustrations, unmet needs, and creative workarounds - 3-4 questions about what is hard, broken, or missing - These often reveal the strongest insights ### Phase 5: Wrap-Up (3-5 minutes) - Open floor for anything not covered - Ask what they wish you had asked about - Thank them and explain next steps - Request referrals to other potential participants 4. **Write follow-up probes for each question** Every core question should have probes like: - "Can you tell me more about that?" - "What happened next?" - "How did that make you feel?" - "Why do you think that is?" - "What would have made that easier?" Probes are used when the participant's initial answer is too surface-level. 5. **Generate the synthesis template** Include a template at the bottom of the guide for post-interview notes: ```markdown ## Post-Interview Synthesis **Participant**: [Name/ID] **Date**: [Date] **Interviewer**: [Name] **Duration**: [minutes] ### Key Quotes - "[Verbatim quote]" -- Context: [what they were talking about] ### Observed Behaviors & Workarounds - [What they actually do vs. what they say they do] ### Pain Points (ranked by intensity) 1. [Strongest pain point] 2. [Second pain point] ### Surprises - [Anything unexpected or that challenged assumptions] ### Jobs to Be Done - When [situation], I want to [motivation], so I can [outcome] ### Implications for Product - [What this means for our roadmap/feature decisions] ### Follow-Up Actions - [ ] [Action item] ``` 6. **Save the guide** Write to `nimbalyst-local/research/interviews/[topic-slug]-interview-guide.md` ## Question Quality Rules Generate questions that follow these principles: - **Open-ended**: Never yes/no questions. Start with "how", "what", "tell me about", "walk me through" - **Non-leading**: Never suggest the answer. BAD: "Don't you find X frustrating?" GOOD: "How do you feel about X?" - **Specific**: Ask about concrete past experiences, not hypothetical futures. BAD: "Would you use a tool that...?" GOOD: "Last time you needed to do X, what did you do?" - **Layered**: Start broad, then drill down. "Tell me about your workflow" before "How do you handle error cases?" - **Neutral**: Use their language, not your product's terminology ## Error Handling - **Topic too vague**: Ask the user to specify what decision this research will inform - **Too many objectives**: Recommend splitting into multiple interview rounds; one guide should have 1-2 primary objectives max - **Solution-focused topic**: Reframe toward the problem space. "Validate our new dashboard" becomes "Understand how users currently monitor their metrics" ## Best Practices - Keep interviews under 45 minutes; attention drops sharply after that - Interview 5-8 people per segment; patterns emerge by interview 5 - Schedule synthesis immediately after each interview while memory is fresh - Look for behavioral evidence over stated preferences - One interviewer asks questions; a second person takes notes if possible - Never combine a sales call with a research interview
Explore More