The model confidence score reveals the level of assurance the underlying algorithm has in extracting the right value. It expresses how confident the machine learning model or NLP service is that the relevant intent has been accurately assigned. A fallback intent is provided if the confidence level drops below a predetermined threshold.

‍