System

A system comprising:

• circuitry configured to
• receive, via a packet-switched network, attribute data and preference data associated with an entity, store the attribute data and the preference data in a storage device, and compute a classification output by inputting interaction data to a trained classification model,
• analyze the attribute data, the preference data, and the classification output, and generate a prompt data sequence comprising an instruction portion and a constraint portion based on an analysis result, the constraint portion specifying exclusion of previously processed candidates based on tracking data stored in the storage device,
• input the generated prompt data sequence to a generative neural network model comprising a transformer-based architecture to acquire candidate output data, and parse the candidate output data into structured candidate records,
• transmit, via the packet-switched network, a processing request to an external data source based on identification information in the structured candidate records, receive a processing result, and register the processing result as tracking data in the storage device, and
• generate notification data based on the processing result and the structured candidate records, and transmit the notification data via the packet-switched network to a remote device for presentation.

A method performed by a system comprising circuitry, the method comprising:

• receiving, via a packet-switched network, attribute data and preference data associated with an entity, and storing the attribute data and the preference data in a storage device;
• computing a classification output by inputting sensor data or interaction data to a trained classification model, and storing classification output data in association with the attribute data and the preference data;
• analyzing the attribute data, the preference data, and the classification output data, and generating a prompt data sequence based on an analysis result;
• inputting the generated prompt data sequence to a generative neural network model to acquire candidate output data;
• transmitting, via the packet-switched network, a processing request to an external data source based on identification information corresponding to the candidate output data, and receiving a processing result;
• and
• generating notification data based on the processing result and the candidate output data, and transmitting the notification data to a remote device for presentation.