A confidence score reflects the model's estimated probability that its prediction is correct. Scores near 1.0 indicate high certainty; scores near 0.5 suggest the model is uncertain. Platforms use confidence thresholds to route low-confidence predictions to human review, ensuring accurate classification while maintaining automation benefits. Setting thresholds too high reduces automation; too low increases error rates.
← Glossary
Confidence Score
A number between 0 and 1 indicating how certain the AI model is about its classification decision.
Related terms
- AI Classifier — A machine learning model that automatically categorizes inputs—like return reasons or customer feedback—into predefined groups.
- Threshold — The minimum confidence score required for the AI to auto-assign a classification versus flagging for human review.
- False Positive — When the AI incorrectly assigns a return reason that doesn't match the actual reason.
- False Negative — When the AI fails to identify the correct return reason and assigns an incorrect one.