Automated agent performance ranking on key performance indicator impact factors

An apparatus configured for providing automated agent performance ranking, comprising:

• one or more memories comprising processor-executable instructions;
• and one or more processors configured to execute the processor-executable instructions and cause the apparatus to:
• obtain, for each agent of a plurality of agents, a performance score for each feature associated with a plurality of customer-agent interactions, wherein each feature is associated with a key performance indicator (KPI) metric and is under control of the agent;
• determine task labels for the plurality of customer-agent interactions, wherein a task label for a customer-agent interaction of the plurality of customer-agent interactions is determined by:
• detect, with a large language model, a first intent from an interaction transcript and an engineered prompt of the customer-agent interaction;
• generate, with the large language model, a first narrative for the first intent and a first confidence score corresponding to a probability that the first intent is an intention in fact expressed in the interaction transcript;
• determine that the first confidence score satisfies a threshold;
• verify that the first narrative conforms with one or more fluency and completeness rules based on a determination that the first confidence score satisfies the threshold;
• update the first narrative based on one or more modified hyperparameters of the large language model when the first narrative does not conform with the one or more fluency and completeness rules;
• output the first narrative as the task label for the customer-agent interaction,
• group the plurality of customer-agent interactions into one or more task groups based on the task label for each of the plurality of customer-agent interactions;
• determine, for each agent of the plurality of agents, a task-feature performance value, wherein the task-feature performance value is a combination of the performance score for each feature associated with the one or more task groups;
• obtain, for a first agent of the plurality of agents, a time series of the task-feature performance value for a corresponding one of the one or more task groups,
• determine that a predicted performance score for each feature in a new customer-agent interaction for the first agent indicates a decrease in performance based on a time series of the task-feature performance value for the corresponding one of the one or more task groups and the decrease in performance meets or exceeds a divergence threshold;
• and
• output an indication when the decrease in performance meets or exceeds the divergence threshold.