Synthetic VOC Comment Insights

Problem

Early product discovery often starts with sparse notes: an industry, a persona, and a few scenarios, while real user feedback is still missing. A plain prompt to “write user comments” tends to produce advisory language or product-manager phrasing instead of the raw frustration, hesitation, and friction that make VOC data useful. This skill frames the task as synthetic customer voice generation, producing first-person comments that can be reviewed and refined rather than jumping to recommendations.

How it works

Given industry, product, target roles, and usage scenarios, it follows a fixed reasoning path. First, Role Modeling creates 3–5 user archetypes and examines skill level, stress sources, accountability, usage frequency, and current workarounds. Next, JTBD decomposes each task chain into surface actions, real goals, and emotional objectives. Then Failure Mapping predicts where users are most likely to click wrong, waste time, give up, or suffer serious consequences under pressure. After confirming the direction, it outputs at least 60 first-person comments grouped by role × scenario, preserving colloquial tone, emotion, emoji, and details typical of e-commerce or local-life reviews.

Boundaries

It is useful for early discovery, comment-style simulation, and failure-scenario exploration. It is not a replacement for user interviews, surveys, or compliant data collection. If the input is incomplete, the skill may fill in reasonable assumptions, so key premises need human confirmation. For B2B or SaaS products, the tone should lean toward internal work-chat complaints and colleague-oriented gripes rather than consumer e-commerce praise.