System

A system comprising:

• circuitry configured to:
• receive, via a communication interface coupled to a packet-switched network, historical interaction data, perform data preprocessing to generate numerical feature information, train a prediction machine learning model on the numerical feature information and outcome information, and extract activity patterns and subject segments with high success probability;
• generate structured prompt sentences incorporating virtual subject conditions and dialog structures from the extracted patterns, input the prompt sentences to a generative AI model to generate AI agent scenario information, and provide interactive sessions in which a user interacts with the AI agent;
• receive utterance information and behavioral information during live interaction sessions, apply an emotion recognition model to generate an emotional state representation, generate real-time advisory content incorporating the emotional state representation, and transmit the advisory content as an overlay display to a terminal device;
• collect feedback data from interactive sessions and feed back the feedback data as training data to update the prediction machine learning model and the generative AI model;
• and
• apply a performance evaluation model to user performance metrics to generate skill assessment scores, and generate personalized training recommendation content by supplying the scores to the generative AI model.

A method comprising:

• receiving, via a communication interface coupled to a packet-switched network, historical interaction data from an information storage device, and performing data preprocessing comprising completion of missing values, removal of abnormal values, and encoding of categorical values to generate numerical feature information;
• constructing and training a prediction machine learning model on the basis of the numerical feature information and outcome information to obtain a trained prediction model, and extracting activity patterns and subject segments having a relatively high success probability using output results of the trained prediction model;
• generating structured prompt sentences comprising virtual subject conditions and dialog structures based on the extracted activity patterns and subject segments, and providing the structured prompt sentences as input to a generative AI model to generate AI agent scenario information;
• providing an interactive session in which a user interacts with the AI agent via the communication interface, receiving prompt inputs from the user, and collecting feedback data comprising interaction outcomes and user performance metrics;
• and
• feeding back the collected feedback data as training data to update the trained prediction model and the generative AI model, and storing updated model parameters in the information storage device.