System

A system comprising:

• circuitry configured to:
• acquire, via a communication interface coupled to a packet-switched network, user-related data records including behavioral event data and transaction data associated with a user, and store the user-related data records in a storage area;
• apply a data analysis program to the user-related data records to identify attribute information and segment information for the user;
• construct a prompt sentence based on the attribute information, the segment information, and rendering style parameters for an interactive agent entity, and transmit the prompt sentence to a generative language model to obtain a response text including proposal content;
• transmit the response text as one or more data packets to a terminal device via the communication interface for presentation via the interactive agent entity;
• receive, from the terminal device, customization data specifying at least one of appearance settings and voice settings for the interactive agent entity, store the customization data in the storage area, and incorporate the customization data into a writing style and an expression style of subsequently constructed prompt sentences;
• record, in the storage area, operation history data and result data received from the terminal device after presentation of the proposal content, compute a quality index and a performance index from the recorded data, and update generation rules and template information for the prompt sentence based on the quality index and the performance index;
• and
• packetize updated proposal content into one or more transport-layer data packets and transmit the one or more transport-layer data packets to a registered network address of the terminal device.

A method performed by circuitry, the method comprising:

• retrieving, from a storage area, interaction records and behavioral log data associated with a user identifier, and applying statistical processing and feature extraction processing to generate user attribute information;
• constructing a first prompt sentence including the user attribute information and transmitting the first prompt sentence via a communication interface coupled to a packet-switched network to a generative language model;
• receiving first response information from the generative language model, structuring the first response information as content data for an interactive agent entity, and transmitting display control data including visual rendering parameters and audio rendering parameters to a terminal device via the communication interface;
• storing session history data, computing quality indicators and performance indicators, generating a second prompt sentence including the quality indicators and the performance indicators, and updating at least one of a strategy parameter and the content data based on second response information from the generative language model;
• and
• acquiring feedback data from the terminal device, generating a third prompt sentence for the generative language model to analyze the feedback data, and modifying at least one of the visual rendering parameters, the audio rendering parameters, and a dialogue policy of the interactive agent entity based on third response information from the generative language model.