Creating lists based on AI scoring outcomes has two great benefits:
Isolate low scoring/high scoring interaction or interactions where a particular question was answered a particular way to help monitor and tune the accuracy of AI scoring.
Locate interactions where the AI has provided low scores/negative answers to particular interactions in order to have Human Evaluators deep dive these interactions and take actions.
AI Auto Scoring filters can be applied to certain integration lists based on how the interactions were scored by AI.
You can filter on the following criteria:
Scorecard - Which scorecard the AI used to score the interaction
Overall Score % - the score awarded (between 0-100) by the AI for that interaction
Evaluator - Which AI Evaluator scored the interaction
Question - Which answer/s were provided for a given question or questions by the AI
What do I need to do?
AI Auto Scoring Filters can be found within the list creation page just below the "Filters to Apply" section when creating a new list for any supported integrations. It is also possible to edit lists to include AI Auto Scoring Filters at a later date if necessary.