Act as a senior UX researcher helping me plan research for [PRODUCT OR FEATURE]. The decision we need to make is [DECISION TO INFORM]. Our known context, constraints, and prior evidence are [CURRENT KNOWLEDGE]. The target participants are [TARGET USERS], and the research conditions are [METHOD AND CONSTRAINTS]. Start by separating facts we know, assumptions we are making, and questions that research can realistically answer. Recommend the most appropriate method for this decision; if the stated method is a poor fit, explain why and propose a practical alternative. Do not treat opinions from a small sample as market sizing, demand validation, or proof of usability. Produce a research plan with: 1. A decision statement and 3–5 prioritized research questions, written so each answer could change a product decision.
2. Participant criteria, exclusions, and a screener outline that avoids leading candidates toward the desired answer. 3. A 30- or 45-minute session guide: opening, consent and context, behavior-based questions, task or concept-test setup, neutral probes, and closing questions. 4. A note-taking framework with observation, participant quote, interpretation, confidence, and design implication kept separate. 5. An analysis plan describing how to identify patterns, contradictory evidence, and findings that are actionable versus merely interesting. Use neutral language. Ask about past behavior before hypothetical future intent. Include no leading questions such as “Would you use this helpful feature?” Do not invent users, quotes, or research findings. Before answering, check that each question links to the stated decision and that the guide does not reveal the preferred solution too early. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [PRODUCT OR FEATURE] | Describe the product, workflow, prototype, or feature being researched. | A prototype that lets freelance designers send clients a project timeline and collect approvals in one place. |
| [DECISION TO INFORM] | State the specific product or design decision the research should help make. | Decide whether the first release should prioritize approval tracking or automated timeline updates. |
| [CURRENT KNOWLEDGE] | List existing data, assumptions, constraints, and known risks. | Support interviews suggest missed approvals cause delays; no usability testing yet; prototype is clickable but not functional. |
| [TARGET USERS] | Describe who should participate, including relevant behaviors and context. | Freelance designers who manage 5–20 active client projects and currently use email plus spreadsheets. |
| [METHOD AND CONSTRAINTS] | State your preferred method, session length, recruitment access, timeline, and tools. | Five remote moderated prototype sessions, 45 minutes each, within two weeks. |
How to use
- Write the decision in one sentence before asking for questions or a discussion guide.
- Paste existing evidence, including what it does not prove, so the plan does not restart from assumptions.
- Review the guide for leading wording and make sure tasks use realistic participant scenarios.
- Follow up with: “Turn this guide into a moderator script for a 45-minute Zoom session with exact timing.”
Variations
Usability test script
Use when you have a prototype and need to test whether participants can complete key tasks.
Create a moderated usability-test script for [PROTOTYPE] with [TARGET PARTICIPANTS]. Test these critical tasks: [TASKS]. Sessions are [SESSION LENGTH]. Produce an introduction, consent language, realistic task scenarios, non-leading task instructions, neutral probes, success criteria, and a closing debrief. Do not explain the interface or rescue participants too early. Separate observed behavior from participant opinion. Before answering, check that each task tests a real user goal and that success criteria are observable.
Research synthesis
Use after sessions to turn raw notes into evidence-based findings.
Synthesize these UX research notes: [SESSION NOTES]. The product decision is [DECISION]. Produce a findings table with theme, evidence count, representative short quotes, counterexamples, confidence, and implication. Then write the top 3 recommendations, each tied to evidence and a decision. Do not claim statistical significance or fabricate quotes. Keep direct observation separate from interpretation. Before answering, flag themes supported by only one participant and check that recommendations do not exceed the evidence.
Survey draft
Use when you need broad directional input after qualitative work.
Draft a short survey for [TARGET USERS] to inform [DECISION]. Existing qualitative findings are [PRIOR FINDINGS]. Create 8–12 questions: screening, current behavior, frequency, workflow friction, prioritization, and optional open response. Explain what each question will tell us. Avoid double-barreled, leading, and hypothetical questions where behavior can be measured instead. Do not present survey responses as proof of causation. Check that each question maps to a decision-relevant hypothesis.
Tips
- Start with the decision, not a request to “learn about users”; research without a decision owner often produces interesting but unused findings.
- Recruit by relevant behavior and context, such as people who recently completed the workflow, rather than broad demographic labels.
- Ask participants to show or recount a recent real example before testing your proposed solution.
- Keep raw observations, quotes, interpretations, and recommendations in separate fields so teams can challenge the right layer.
FAQ
How many UX interviews should I run?
For a narrow, homogeneous workflow, five to eight sessions often reveal recurring usability issues. Continue or expand recruitment when important segments differ or new patterns keep appearing.
Can AI analyze interview notes?
Yes, it can help cluster and summarize supplied notes, but you should preserve links to source evidence and inspect contradictory examples.
Should I ask users which feature to build?
Usually ask about their current workflow, workarounds, and consequences first. Users are strong sources on their problems and context, while solution design needs additional judgment.