Name: AI Sentiment Analysis Lab
Author: MySimulator

Alerting policy

Negative-sentiment alert threshold: 0.65

This case study models 2,000 social media posts mentioning a brand, each scored for negative-sentiment intensity. Alerting the PR team at a low threshold catches early signs of a real backlash forming, but also floods them with everyday grumbling that never escalates.

Day outcome

Posts flagged for PR review—

True backlash posts caught—

PR review hours—

Posts by negative-sentiment score (red = flagged, ring = real backlash signal)