Back to changelog
NewRules & Automation

A New "Set Sentiment" Action Has Been Added To The Rule Engine To Prevent Misclassification Of Slang

Social media is full of slang such as "sick" or "GOAT," which AI-based sentiment analysis can mistakenly classify as neutral or negative.

Previously, misclassified messages had to be corrected manually โ€” but with the new Action: "Set Sentiment" added to the Rule Engine, messages containing slang are automatically classified with the correct sentiment. This also improves reporting accuracy.

**How to set it up **๐Ÿ‘‰ Engage โ†’ Rule Engine โ†’ Add New Rule โ†’ Create manually โ†’ Actions โ†’ Set Sentiment โ†’ Choose from Positive / Neutral / Negative / Mixed โ†’ Click Save when all other settings are complete.

This way, any conversation matching the filter will have its sentiment classified as the one you've set, preventing misclassification.

Example: Filter โ†’ Message Content contains "GOAT" โ†’ Action โ†’ Set Sentiment โ†’ Select Positive ๐Ÿ‘‰ Messages containing "GOAT" will be classified as Positive