Back to results
US 2026/0288909 A1Published application
System
AssigneeSoftBank Group Corp.
InventorTatsuya SUZUKI
Life of this patent
Estimating the term…
Abstract
A system includes a processor that is configured to; use a virtual private network connection to securely receive information from a user, generate, by using a generative artificial intelligence model, a prompt sentence for analyzing the received information and extract important words from the received information, and automatically classify the information into an appropriate service or department based on the extracted words.
Claims19
A system comprising:
- a communication interface connected to a packet-switched network;
- and
- circuitry configured to:
- establish, via the communication interface, an encrypted communication channel with a terminal device over the packet-switched network;
- receive, through the encrypted communication channel, input data comprising a natural-language query from the terminal device;
- perform natural-language processing on the input data to extract a set of feature terms from the natural-language query;
- construct, based on the input data and the set of feature terms, a prompt data structure configured to instruct a generative neural network model to produce an analysis of the input data;
- transmit the prompt data structure to the generative neural network model via the communication interface;
- receive, from the generative neural network model, output data comprising a classification result and response content generated in response to the prompt data structure;
- determine, based on the classification result and the set of feature terms, a routing destination from among a plurality of processing modules stored in a data storage device;
- and
- transmit, via the encrypted communication channel, the response content to the terminal device.
9 more elements
A system comprising:
- a communication interface connected to a packet-switched network;
- a data storage device;
- and
- circuitry configured to:
- establish an encrypted communication channel with a terminal device via the communication interface;
- receive input data comprising a natural-language query through the encrypted communication channel;
- extract feature terms from the input data using natural-language processing;
- search an information storage structure based on the feature terms to acquire related information;
- construct a prompt data structure based on the input data, the feature terms, and the related information;
- transmit the prompt data structure to a generative neural network model and receive output data comprising a classification result and response content;
- associate source identification information with the response content;
- determine a routing destination based on the classification result;
- and
- transmit the response content and the source identification information to the terminal device via the encrypted communication channel.
11 more elements
A method comprising:
- establishing, by circuitry via a communication interface connected to a packet-switched network, an encrypted communication channel with a terminal device;
- receiving, through the encrypted communication channel, input data comprising a natural-language query;
- performing, by the circuitry, natural-language processing on the input data to extract a set of feature terms;
- constructing, by the circuitry, a prompt data structure based on the input data and the set of feature terms, the prompt data structure configured to instruct a generative neural network model;
- transmitting the prompt data structure to the generative neural network model and receiving output data comprising a classification result and response content;
- determining, based on the classification result, a routing destination from among a plurality of processing modules;
- and
- transmitting the response content to the terminal device via the encrypted communication channel.
5 more elements
Dependencies are inferred from the claim text.