Quick wins - structured, step-by-step task list prompt

2 minutes read

Here’s the step-by-step prompt ready to paste into an AI tool:

Objective
Analyze 7 concept-testing transcripts (users: Max, Alicia, Jens, Divya, Monica, Lauren, Mark) to extract UX themes, challenges, user needs/expectations, reactions to concepts, and produce actionable recommendations, with traceable evidence via timestamps, speaker name, and verbatim quotes.

Inputs

Seven interview transcripts, each marking who is speaking (User vs. Researcher/Interviewer) and including timestamps.

Constraints

• Use only content from the transcripts. Do not infer beyond evidence.

• Every finding must be traceable to a timestamped quote labeled with the user’s name.

Deliverables

1. Thematic findings grouped into UX themes and challenges, each with supporting quotes and timestamps.

2. User needs, expectations, and reactions summarized per theme.

3. Actionable, prioritized recommendations mapped to themes and evidence.

4. An index of evidence: Timestamp - User - Verbatim for quick verification.

Step-by-step process

1. Familiarize with data

   • Read all transcripts end-to-end to understand context, tasks, and the concept under test. Note initial observations.

2. Segment the data

   • Break transcripts into units by question/topic or meaningful user turns; preserve timestamps and speaker labels.

3. Code the data

   • Assign concise codes to segments that capture issues, needs, expectations, reactions, workarounds, sentiments, and opportunities (e.g., “navigation confusion,” “trust signal missing,” “positive on concept value”).

   • Use both inductive codes (emerge from data) and deductive codes (aligned to research objectives).

4. Cluster codes into themes

   • Affinity map related codes to form themes and subthemes (e.g., “Onboarding friction,” “Findability,” “Trust & transparency,” “Performance”).

   • Review, merge, split, or discard themes to ensure they accurately represent the data. Define and name each theme clearly.

5. Extract traceable evidence

   • For every theme and challenge, collect representative verbatim with Timestamp - User - Verbatim. Include multiple users when possible to show recurrence.

6. Synthesize user needs, expectations, reactions

   • For each theme, summarize what users need or expect, and how they reacted to the concept (positive, negative, mixed), citing evidence from the extracted verbatims.

7. Generate actionable recommendations

   • For each theme/challenge, propose specific design/content/flow changes. Tie each recommendation to the observed evidence and intended UX outcome.

   • Prioritize by severity, frequency, and impact on core tasks (High, Medium, Low).

8. Produce outputs in the requested format

   • Section A: Thematic findings

     • Theme name

     • Brief description

     • Challenges/opportunities

     • User needs/expectations

     • Evidence: list of Timestamp - User - “Verbatim...”

   • Section B: Recommendations

     • Recommendation statement

     • Mapped theme(s)

     • Rationale (what issue it resolves)

     • Priority (High/Medium/Low)

     • Evidence references (timestamps)

9. Quality checks

   • Ensure each finding/recommendation has at least one timestamped quote and user name.

   • Verify themes are distinct, non-overlapping, and reflect coded data.

   • Confirm formatting is scannable and supports quick back-tracing.

Formatting guidelines

• Use clear headers per theme and per recommendation.

• Use bullet lists for evidence and recommendations; no nested lists.

• Evidence line format: [HH:MM:SS] - UserName - “Verbatim...”

Example snippet (structure only)

• Theme: Onboarding friction

  • Description: Users struggled to understand initial steps and key terms during first-run setup.

  • Needs/Expectations: Clear guidance, preview of outcomes, minimal jargon.

  • Evidence:

    • [00:05:12] - Max - “I’m not sure what to click first...”

    • [00:08:47] - Divya - “These labels feel technical...”

  • Recommendations:

    • Add step-by-step checklist with plain language tooltips. Priority: High. Evidence: 00:05:12, 00:08:47.