System

A system comprising:

• circuitry configured to:
• receive, via a communication interface coupled to a packet-switched network, question information expressed in a natural language and utilization information of a user from a terminal device;
• apply a natural language processing model to the question information to extract a category parameter, a temporal condition, a proximity condition, and a preference condition as search parameters;
• generate a structured database query as inquiry information based on the search parameters weighted by a utilization tendency derived from the utilization information, and acquire search result information from a storage device;
• compute a relevance score for each item in the search result information based on a purchase history and a utilization frequency of the user, and reorder items by the relevance score;
• construct a prompt sentence by encoding the question information, the search parameters, and the reordered items as structured text segments and concatenating the segments into an input sequence;
• input the prompt sentence to a transformer-based generative neural network model to generate response information expressed in a natural language, and transmit presentation information comprising the response information to the terminal device;
• and
• associate question information and response information with dialogue identification information, store dialogue history in the storage device, and update subsequent prompt sentences by incorporating a summarized representation of stored dialogue history.

A method comprising:

• receiving, via a communication interface coupled to a packet-switched network, question information expressed in a natural language and utilization information of a user from a terminal device;
• analyzing the question information using a natural language processing model to extract search parameters, the search parameters comprising at least a category parameter, a temporal condition, a proximity condition, and a preference condition;
• generating inquiry information for a storage device based on the search parameters and the utilization information, transmitting the inquiry information to the storage device, and acquiring search result information from the storage device;
• extracting and reordering presentation-target information from the search result information based on the utilization information;
• constructing a prompt sentence for input to a generative neural network model, the prompt sentence comprising at least the question information, the search parameters, and the reordered presentation-target information;
• inputting the prompt sentence to the generative neural network model to generate response information expressed in a natural language, and transmitting presentation information comprising the response information to the terminal device via the communication interface;
• and
• associating a plurality of pieces of the question information and the response information with each other based on dialogue identification information, storing dialogue history information in the storage device, and updating a content of a subsequent prompt sentence based on the dialogue history information.