# Sentiment analysis

Analyzes reviews, surveys and other feedback to find user segments, how satisfied each one is and what to fix.

- Kind: a skill for AI agents
- Version: 1.0, updated 2026-11-18
- Shared by: Workotron
- Written by: [Pawel Huryn](https://github.com/phuryn/pm-skills/tree/9af7bfd0bc47fdad43c2eaa7a43cd13918e7aaef/pm-market-research/skills/sentiment-analysis)
- License: MIT
- Business area: Operations
- Tags: #operations #claude
- Rank: not ranked yet, with 0 copies, 0 thumbs up and 0 thumbs down
- Permanent address: https://skills.workotron.com/sentiment-analysis

## Who is this for?

- Owners with a pile of customer reviews or survey answers
- Product managers who need to find patterns in feedback

## What it does

- Reads the CSV files, PDFs, survey answers, reviews or social listening reports you give it
- Finds at least three distinct user segments
- Pulls out themes, praise and complaints for each segment, with quotes when available
- Scores sentiment for each segment from minus 1 to plus 1
- Recommends two or three high-impact improvements for each segment

## When to use it

- When you have feedback at scale to make sense of
- After a survey closes, to find themes in the answers
- When you want to know which customers are unhappy and why

## About this skill

It reads the files you provide and cites its sources. It notes minority views and flags segments with small samples. It may give a rough Net Promoter Score where it applies.

The output is a written report with one profile for each segment.

## Supported platforms

- Workotron
- Claude Code
- ChatGPT
- Codex

## Install

Get https://skills.workotron.com/sentiment-analysis.zip and extract it into your skills folder. If you cannot unzip, get each file:

- https://skills.workotron.com/sentiment-analysis/files/LICENSE.txt
- https://skills.workotron.com/sentiment-analysis/files/SKILL.md

## Discussion (0)
