Conversation provision system and device

Abstract

We provide a conversation delivery system with improved usability. [Solution] The conversation provision system comprises conversation units 84, 86, 88 that instruct a task processing device 50, which processes tasks using artificial intelligence in response to task processing instructions, to generate conversation content for at least one of a plurality of virtual agents for the user, and acquire the conversation content generated by the task processing device 50 and output it to the user; and notification units 78, 82, 92 that, for multiple individual notifications to the user from multiple virtual agents, summarize and combine the individual notification sources and individual notification content of the multiple individual notifications, and notify the user of the main agent among the plurality of virtual agents as the notification subject.

Description

Technical Field

[0001] The present invention relates to a conversation providing system and apparatus that provides a conversation with a plurality of virtual agents to a user using artificial intelligence.

Background Art

[0002] For a user, a conversation with a plurality of virtual agents is provided using artificial intelligence. As such a plurality of virtual agents, a main agent that conducts a conversation with the user and a specialized expert agent specialized in various services are used (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When using a plurality of virtual agents, if notifications are respectively made to the user from the plurality of virtual agents, the user experience will deteriorate.

[0005] An object of the present invention is to provide a conversation providing system and apparatus with improved user experience.

Means for Solving the Problems

[0006] One aspect of the present invention is a conversation provision system comprising: a conversation unit that instructs a task processing device, which processes tasks using artificial intelligence in response to task processing instructions, to generate conversation content for at least one of a plurality of virtual agents to a user, and acquires the conversation content generated by the task processing device and outputs it to the user; and a notification unit that, with the main agent among the plurality of virtual agents acting as the notification subject, notifies the user of a summary notification in which the individual notification sources and individual notification content of a plurality of individual notifications to the user, with the individual notification source and individual notification content of the plurality of individual notifications of the plurality of virtual agents acting as the notification subject.

[0007] Another aspect of the present invention is a conversation providing device comprising: a conversation output unit that instructs a task processing device, which processes tasks using artificial intelligence in response to task processing instructions, to generate conversation content for at least one of a plurality of virtual agents to a user, and acquires the conversation content generated by the task processing device and outputs it to the user; and a notification output unit that, for a plurality of individual notifications to a user with the plurality of virtual agents as individual notification sources, outputs an aggregated notification in which the individual notification sources and individual notification content of the plurality of individual notifications are summarized and combined, with the main agent among the plurality of virtual agents as the notification subject. [Effects of the Invention]

[0008] This invention makes it possible to improve usability. [Brief explanation of the drawing]

[0009] [Figure 1] A block diagram showing the hardware configuration of a conversation provisioning system according to one embodiment of the present invention. [Figure 2] A block diagram showing the functional configuration of a conversation provision system according to one embodiment of the present invention. [Figure 3A] A first sequence diagram illustrating a conversation delivery method according to one embodiment of the present invention. [Figure 3B] A second sequence diagram illustrating a conversational delivery method according to one embodiment of the present invention. [Figure 3C] A third sequence diagram illustrating a conversational delivery method according to one embodiment of the present invention. [Figure 3D] A fourth sequence diagram illustrating a conversation delivery method according to one embodiment of the present invention. [Figure 3E] A fifth sequence diagram illustrating a conversation provision method according to one embodiment of the present invention. [Figure 4] A table and diagram illustrating how to set individual priorities and contributing factors for individual notifications in one embodiment of the present invention. [Figure 5] A table and diagram illustrating the contents of individual notices and aggregated notices for each embodiment of the present invention. [Figure 6] A schematic diagram showing the agent selection screen of one embodiment of the present invention. [Modes for carrying out the invention]

[0010] An embodiment of the present invention will be described with reference to Figures 1 to 6. In the following, a virtual agent that engages in conversation with a user using artificial intelligence will be referred to as an AI agent. Among the multiple AI agents prepared in a conversational application, the AI ​​agent that primarily engages in conversation with the user will be referred to as the primary agent. While some AI agents provide information to the user according to a specific purpose or assist the user in performing various procedures, the primary agent that primarily engages in conversation with the user is expected not only to provide information and perform procedures according to a specific purpose, but also to build a close relationship with the user, gain the user's trust, and become a supportive presence for the user.

[0011] In this embodiment, for multiple individual notifications sent to the user from multiple AI agents, the individual notification sources and content of the multiple individual notifications are summarized and compiled into an aggregated notification, which is sent to the user by the main agent as the notification entity. The main agent acts as if it were a personal secretary for the user, improving usability and enabling the building of a better relationship between the user and the main agent.

[0012] As shown in Figures 1 and 2, the conversation provision system of this embodiment includes a conversation provision server 10, a generation server 50 as a task processing device, a plurality of client terminals 60, and a user terminal 70. In the figures, the plurality of client terminals 60 are represented by a single client terminal 60.

[0013] Referring to Figure 1, the hardware configuration of the conversation provisioning system of this embodiment will be described. Since the hardware configurations of the conversation provision server 10 and the generation server 50 are the same, they will be described collectively as servers below. Similarly, since the hardware configurations of the client terminal 60 and the user terminal 70 are the same, they will be described collectively as terminals below.

[0014] As shown in Figure 1, servers 10 and 50 are physically configured as computers including a processor 211, memory 212, storage 213, communication device 214, input device 215, output device 216, and buses connecting them. Each of these devices operates on power supplied from a battery (not shown). In the following description, the term "device" can be read as a circuit, device, unit, etc. The hardware configuration of servers 10 and 50 may include one or more of the devices shown in the figure, or it may be configured without some of the devices. Alternatively, multiple devices with different enclosures may be connected via communication to constitute servers 10 and 50.

[0015] Each function in servers 10 and 50 is realized by causing a processor 211 to perform operations, control communication by a communication device 214, or control at least one of reading and writing data in a memory 212 and a storage 213, by loading a predetermined software (program) onto hardware such as the processor 211 and the memory 212.

[0016] The processor 211 controls the entire computer by operating, for example, an operating system. The processor 211 may be constituted by a central processing unit (CPU: Central Processing Unit) including an interface with peripheral devices, a control device, an arithmetic unit, registers, and the like. Also, for example, a baseband signal processing unit, a call processing unit, and the like may be realized by the processor 211.

[0017] The processor 211 reads a program (program code), a software module, data, etc. from at least one of the storage 213 and the communication device 214 into the memory 212, and executes various processes according to these. As the program, a program for causing a computer to execute at least a part of the operations described later is used. The functional blocks of the servers 10 and 50 may be stored in the memory 212 and realized by a control program operating in the processor 211. Various processes may be executed by one processor 211, or may be executed simultaneously or sequentially by two or more processors 211. The processor 211 may be implemented by one or more chips. Note that the program may be transmitted to the servers 10 and 50 via a telecommunication line.

[0018] Memory 212 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. Memory 212 may also be referred to as a register, cache, main memory, etc. Memory 212 can store executable programs (program code), software modules, etc., for carrying out the method according to this embodiment.

[0019] The storage 213 is a computer-readable recording medium and may consist of at least one of the following: an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disc, a digital multipurpose disc, a Blu-ray® disc), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. The storage 213 may also be referred to as an auxiliary storage device.

[0020] The communication device 214 is hardware (transceiver / receiver device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc.

[0021] Each device, such as the processor 211 and memory 212, is connected by a bus for communicating information. The bus may be configured using a single bus, or different buses may be used for each device.

[0022] Servers 10 and 50 may be configured with hardware such as microprocessors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable logic devices (PLDs), and field-programmable gate arrays (FPGAs), and some or all of each functional block may be implemented by such hardware. For example, processor 211 may be implemented using at least one of these hardware components.

[0023] Terminals 60 and 70 are computers such as smartphones, mobile phones, tablets, or wearable devices. Physically, terminals 60 and 70 are configured as computer devices including a processor 261, memory 262, storage 263, communication device 264, input device 265, output device 266, and a bus connecting them. The processor 261, memory 262, and storage 263 are hardware similar to the processor 211, memory 212, and storage 213 of the information processing device.

[0024] The communication device 264 may include, for example, a high-frequency switch, duplexer, filter, frequency synthesizer, etc., to implement at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the transmitting and receiving antennas, amplifier section, transmitting and receiving section, transmission path interface, etc., may be implemented by the communication device 264. The transmitting and receiving section may be physically or logically separated into a transmitting section and a receiving section.

[0025] The input device 265 is an input device that accepts input from an external source (e.g., a key, microphone, switch, button, sensor, reader, scanner, etc.). The output device 266 is an output device that outputs to an external source (e.g., a display, speaker, LED lamp, etc.). The input device 265 and the output device 266 may be configured as an integrated unit (e.g., a touchscreen).

[0026] Referring to Figures 2 and 3, the functional configuration and conversation provision method of this embodiment will be described. Hereinafter, each functional configuration of the conversation provision system will be described in accordance with each step of the conversation provision method. The conversation provision program of this embodiment works in cooperation with the hardware configuration of the conversation provision system of this embodiment described above to realize each functional configuration of the conversation provision system of this embodiment and to execute each step of the conversation provision method.

[0027] Furthermore, in one embodiment described below, the user is provided with a conversation with the AI ​​agent through voice input and voice output, but the user may also be provided with a conversation with the AI ​​agent through text input and text output on the screen.

[0028] As shown in Figure 2, in the conversation provision system, the conversation provision server 10 stores user database 14 (hereinafter also referred to as "user DB14") which is associated with user identification information that identifies the user, main agent identification information that identifies the main agent that primarily engages in conversations with the user, affinity information that indicates the user's affinity to each AI agent, user preference information that indicates the user's preferences, and conversation history information that indicates the conversation history between the user and each AI agent.

[0029] Furthermore, the agent database 16 (hereinafter also referred to as "Agent DB16") stores agent feature information for numerous AI agents, associated with each agent's identification information, indicating the characteristics of each AI agent. Agent feature information includes voice information, character feature information, field-specific information, and importance information. Voice information is used to generate the voice of the AI ​​agent. Character feature information indicates the character characteristics of the AI ​​agent. Field-specific information identifies the field of expertise of the AI ​​agent. Here, the AI ​​agents used include conversational agents whose primary purpose is to converse with the user, and specialized agents that provide information to the user or assist the user in various procedures in specific specialized fields, and are operated by various clients such as businesses and government offices. The field of expertise of an AI agent is general conversation if the AI ​​agent is a conversational agent, and the specific specialized field handled by that specialized agent if the AI ​​agent is a specialized agent. The importance information indicates the importance of the AI ​​agent to the user and is determined according to the field of expertise of the AI ​​agent and the client operating the AI ​​agent.

[0030] As shown in Figures 2 and 3A, the conversation provision system and conversation provision method allow the user to select the primary agent with whom they will primarily engage in conversation, and also allow the user to select the current agent with whom they are currently engaging in conversation, as described below.

[0031] Conversation application startup step S12 The user terminal 70 has a conversation application installed for engaging in conversation with the AI ​​agent. On the user terminal 70, the application launch unit 72 launches the conversation application in response to the user's operation to launch the conversation application.

[0032] Primary agent selection step S14 On the user terminal 70, the main agent selection unit 74 presents the user with several conversation agents, primarily designed for conversation with the user, from among the multiple AI agents prepared for the conversation application, and allows the user to select the main agent with whom they will primarily converse. The user can select the main agent through on-screen selection operations, voice input, etc. The multiple conversation agents prepared for the conversation application correspond to existing characters and original characters created for the conversation application, and the user can select their preferred character as the main agent with whom they will primarily converse.

[0033] Primary agent designation request step S16 On the user terminal 70, the main agent selection unit 74 sends a main agent designation request to the conversation provision server 10, specifying the main agent selected by the user, along with user identification information to identify the user and main agent identification information to identify the main agent.

[0034] Primary agent designation step S18 In the conversation server 10, the agent designation unit 12 stores the main agent designation information in the user DB 14 in association with the user identification information, in response to a main agent designation request received from the user terminal 70 along with the user identification information and main agent identification information.

[0035] Current agent selection step S22 On the user terminal 70, the current agent selection unit 76 allows the user to select the current agent to converse with at the moment from among the multiple AI agents prepared in the conversation application, and sets current agent identification information to identify the current agent. The user can select the current agent by selecting it on the screen, by voice input, etc. In the conversation application, in addition to the multiple conversation agents mentioned above, there are also multiple specialized agents that are operated by various clients such as businesses and government offices, which provide information to the user in specific areas of expertise, assist the user in various procedures, etc. The main agent may be selected as the current agent, and when the conversation application is started, the main agent is usually automatically selected as the current agent.

[0036] Current agent image display step S24 On the user terminal 70, the image generation unit 78 generates a current agent image indicating the current agent, and the image display unit 82 displays the current agent image generated by the image generation unit 78 to the user.

[0037] As shown in Figures 2 and 3B, the conversation provision system and conversation provision method, as described below, involve the current agent responding to utterances from the user to the current agent.

[0038] Speech acquisition step S28 In the user terminal 70, the voice acquisition unit 84 acquires the voice spoken by the user to the current agent and obtains voice information indicating the spoken voice.

[0039] Request for response step S32 On the user terminal 70, the conversation execution unit 88 sends a response request to the conversation provision server 10, along with the utterance voice information, user identification information, and current agent identification information acquired by the voice acquisition unit 84, requesting a response from the current agent to the user. The response generation request here instructs the conversation generation AI to generate a response from the AI ​​agent to the user.

[0040] Speech analysis step S34 In the conversation provision server 10, the speech analysis unit 18 analyzes the utterance audio information received from the user terminal 70 along with user identification information and current agent identification information, generates utterance audio text data by converting the utterance audio into text, and outputs it to the conversation generation instruction unit 24 along with the user identification information and current agent identification information.

[0041] Answer generation instruction step S36 In the conversation provision server 10, when the conversation generation instruction unit 24 receives utterance text data along with user identification information and current agent identification information from the speech analysis unit 18, it extracts conversation history information associated with the user identification information and current agent identification information from the user DB 14, and extracts agent feature information associated with the current agent identification information from the agent DB 16. The conversation history information shows the conversation history between the user and each AI agent, and if the conversation history information is large, only a certain amount of the most recent conversation history information is extracted. The agent feature information shows the characteristics of the AI ​​agent and includes character feature information that shows the character characteristics of the AI ​​agent, and field identification information that identifies the field in which the AI ​​agent handles. The field in which the AI ​​agent handles is set to normal conversation if the AI ​​agent is a conversation agent, and to a specific specialized field handled by the specialized agent if the AI ​​agent is a specialist agent.

[0042] Next, the conversation generation instruction unit 24 generates a response generation prompt based on the utterance speech text data, conversation history information, character feature information, and field identification information, and sends it to the generation server 50 along with user identification information, current agent identification information, and field identification information. The response generation prompt instructs the conversation generation AI corresponding to the field handled by the current agent to generate response text data that shows the current agent's response to the most recent utterance from the user indicated by the utterance speech text data. The instruction is to ensure that the current agent's response is natural in light of the conversation history and matches the character features and field handled by the current agent.

[0043] Answer generation step S38 In the generation server 50, the conversation generation unit 52 utilizes LLM (Large Language Models) and is equipped with multiple conversation generation AIs corresponding to the fields handled by each AI agent. These multiple conversation generation AIs include a general conversation generation AI and conversation generation AIs corresponding to the specialized fields handled by each specialized agent. The specialized conversation generation AIs are fine-tuned based on information about their respective fields.

[0044] The conversation generation unit 52 receives a response generation prompt from the conversation provision server 10 along with user identification information, current agent identification information, and field identification information, inputs it to a conversation generation AI corresponding to the field of handling identified by the field identification information, receives response text data indicating the current agent's response from the conversation generation AI, and transmits the response text data to the conversation provision server 10 along with the user identification information and current agent identification information.

[0045] Conversation history update step S42 In the conversation provision server 10, the conversation generation instruction unit 24 updates the conversation history information by adding the utterance text data input from the speech analysis unit 18 along with the user identification information and the current agent identification information, as well as the response text data received from the generation server 50 along with the user identification information and the current agent identification information, to the conversation history information associated with the user identification information and the current agent identification information in the user DB 14. The conversation generation instruction unit 24 also outputs the response text data along with the user identification information and the current agent identification information to the speech generation unit 22.

[0046] Answer voice generation step S44 In the conversation provision server 10, when the voice generation unit 22 receives response text data along with user identification information and current agent identification information from the conversation generation instruction unit 24, it extracts voice information of agent feature information associated with the current agent identification information from the agent DB 16, generates response voice data by converting the response text data into voice using the current agent's voice, and transmits it to the user terminal 70 identified by the user identification information. In this embodiment, the voice generation unit 22 and the conversation generation instruction unit 24 form the conversation output unit.

[0047] Answer voice output step S46 In the user terminal 70, the conversation execution unit 88 acquires the response voice data received from the conversation provision server 10 and outputs it to the voice output unit 86. The voice output unit 86 then outputs the current agent's response to the user based on the response voice data input from the conversation execution unit 88. In this way, in this embodiment, the conversation unit is formed by the voice output unit 86 and the conversation execution unit 88.

[0048] As shown in Figures 2 and 3C, the conversation provision system and conversation provision method set or update familiarity information indicating the user's familiarity with each AI agent and user preference information indicating the user's preferences in the user DB14, as described below.

[0049] Conversation analysis instruction step S52 In the conversation provision server 10, the conversation analysis instruction unit 26 extracts conversation history information associated with user identification information and agent identification information from the user DB 14, and generates sensitive information extraction prompts and preference information extraction prompts based on the extracted conversation history information, and sends them to the generation server 50 along with the user identification information and agent identification information. The sensitive information extraction prompt instructs the generation AI for conversation analysis to extract sensitive information indicating the degree of sensitivity of the conversation between the user and the current agent from the conversation history information. The preference information extraction prompt instructs the generation AI for conversation analysis to extract user preference information indicating the user's preferences from the conversation history. The conversation analysis instruction unit 26 then outputs the extracted conversation history information along with the user identification information and agent identification information to the intimacy setting unit 28.

[0050] Conversation analysis step S54 In the generation server 50, the conversation analysis unit 54 is equipped with a generation AI for conversation analysis that utilizes LLM. The conversation analysis unit 54 inputs the sensitive information extraction prompt and the preference information extraction prompt, received from the conversation provision server 10 along with user identification information and agent information, to the generation AI for conversation analysis, receives the sensitive information and user preference information outputs from the generation AI for conversation analysis, and transmits the sensitive information and user preference information to the conversation provision server 10 along with user identification information and agent identification information.

[0051] Intimacy level setting step S56 In the conversation provision server 10, the intimacy setting unit 28 sets the intimacy level of the user identified by the user identification information to the AI ​​agent identified by the agent identification information, based on the amount of conversation indicated by the conversation history information received from the conversation analysis instruction unit 26 along with the user identification information and agent identification information, and the sensitivity level of the conversation indicated by the sensitive information received from the generation server 50 along with the user identification information and agent identification information, and stores the intimacy information indicating the intimacy level in the user DB 14.

[0052] Regarding the setting of intimacy, instead of analyzing conversation history using a generative AI for conversation analysis, it may be set based on the amount of conversation between the user and the AI ​​agent, or it may be set based on the number of times, frequency, and total time the AI ​​agent has operated as the primary agent or current agent, or these may be combined with the analysis of conversation history using a generative AI for conversation analysis.

[0053] User preference setting step S58 In the conversation provision server 10, the user preference setting unit 32 sets user preference information based on the user preference information received from the generation server 50 along with the user identification information, and stores the user preference information in the user DB 14 in association with the user identification information.

[0054] Regarding the setting of user preferences, instead of analyzing conversation history using a generative AI for conversation analysis, user preferences may be set based on user input, user purchase history on the user's device, content viewing or browsing history, website search and browsing history, etc., or these may be combined with the analysis of conversation history using a generative AI for conversation analysis.

[0055] Update step S62 If conversation history information is updated in User DB14, the familiarity information and user preference information will be updated in User DB14 as appropriate, in the same manner as in the steps described above.

[0056] As shown in Figures 2 and 3D, the conversation provision system and conversation provision method, as described below, summarizes and combines multiple individual notifications to the user, each originating from multiple AI agents, based on priority according to the user's interest information. This aggregated notification is then sent to the user by the main agent acting as the notification entity.

[0057] Individual notification provision step S68 On the client terminal 60, the individual notification provision unit 62 generates individual notification information indicating individual notifications from a specialized agent to a user in response to input from a client such as a company operating a specialized agent or a government office, and transmits it to the conversation provision server 10. The individual notification information includes individual notification identification information that identifies the individual notification, individual notification destination information indicating the user identification information of the individual notification recipient, individual notification source information indicating the agent identification information of the individual notification source agent, individual notification content information indicating the content of the individual notification, and individual notification related information indicating information related to the individual notification. Examples of individual notification content will be described in detail later.

[0058] Individual notification generation step S72 In the conversation server 10, the individual notification generation unit 34 generates individual notification information indicating individual notifications from conversation agents to users and outputs it to the individual notification acquisition unit 36. The individual notification information indicating individual notifications from conversation agents to users is the same as the individual notification information indicating individual notifications from specialized agents to users, and examples of the content of individual notifications will also be described in detail later.

[0059] Individual notification acquisition step S74 In the conversation server 10, the individual notification acquisition unit 36 ​​acquires individual notification information received from the client terminal 60 and individual notification information input by the individual notification generation unit 34, and outputs it to the priority setting unit 38.

[0060] Individual priority setting step S76 In the conversation provision server 10, the priority setting unit 38 sets individual priority information, which is the priority of individual notifications, based on the individual notification information input from the individual notification acquisition unit 36, according to the interest information about the user's interests, adds the individual priority information indicating the individual priority to the individual notification information, and then outputs the individual notification information to the aggregated notification generation unit 44.

[0061] Specifically, the priority setting unit 38 extracts user preference information and affinity information from the user DB 14, which are associated with the user identification information of the individual notification recipient user and the agent identification information of the individual notification source agent, respectively, as indicated by the individual notification recipient information and individual notification source information. It also extracts importance information from the agent DB 16, which are associated with the agent identification information of the individual notification source agent, as indicated by the individual notification source information. Here, the importance information indicates the importance of the AI ​​agent and is determined by the field of the AI ​​agent and the client operating the AI ​​agent. For each individual notification identified by the individual notification identification information, the priority setting unit 38 sets individual priority information indicating the individual priority based on affinity information indicating the affinity of the individual notification recipient user to the individual notification source agent, importance information indicating the importance of the individual notification source agent to the user, notification content information indicating the content of the individual notification, and user preference information indicating the preferences of the individual notification recipient user. Furthermore, if an individual priority is set high, the priority setting unit 38 sets factor information indicating the factors that led to the high individual priority. Examples of how to set individual priorities and contributing factors for individual notifications will be described in detail later. Similar to the setting of familiarity levels or user preferences described above, a conversational analysis generation AI may be used to set individual priorities and contributing factors for individual notifications.

[0062] Aggregated notification generation request step S78 In the conversation server 10, the aggregate notification provision unit 42 provides aggregate notifications at appropriate time intervals or frequencies.

[0063] The aggregated notification provision unit 42 outputs an aggregated notification generation request to the aggregated notification generation unit 44 a predetermined time before the notification time of the aggregated notification, requesting the generation of aggregated notification information indicating the aggregated notification to be notified at that notification time. The aggregated notification information includes aggregated notification identification information that identifies the aggregated notification, aggregated notification destination information that indicates the recipient of the aggregated notification, aggregated notification source information that indicates the source of the aggregated notification, individual notification group information that indicates the group of individual notifications aggregated in the aggregated notification, and aggregated notification content information that indicates the content of the aggregated notification.

[0064] Aggregation notification destination and aggregation notification source setting step S82 In the conversation provision server 10, when an aggregated notification generation request is input by the aggregated notification provision unit 42, the aggregated notification generation unit 44 extracts individual notification information from each individual notification information that has the same user identification information of the individual notification destination user indicated by the individual notification destination information, from each individual notification information that has been input from the priority setting unit 38 and temporarily stored, and sets the user identification information of the common individual notification destination user as aggregated notification destination information that indicates the notification destination user of the aggregated notification. In addition, the aggregated notification generation unit 44 extracts the main agent identification information associated with the user identification information indicated by the aggregated notification destination information from the user DB 14, and sets it as aggregated notification source information that indicates the notification source agent of the aggregated notification.

[0065] Individual notification group setting step S84 In the conversation provision server 10, the aggregated notification generation unit 44 sets each individual notification information that indicates an individual notification sharing the user identification information of the individual notification recipient users as described above, as individual notification group information that indicates a group of individual notifications to be aggregated into an aggregated notification. Note that individual notification information that has been extracted once will be excluded from extraction in subsequent instances.

[0066] Summary notification content generation instruction step S86 In the conversation provision server 10, the aggregated notification generation unit 44 generates an aggregated notification content generation prompt based on the individual notification group information and sends it to the generation server 50 along with the aggregated notification identification information. The aggregated notification content generation prompt instructs the conversation generation AI, which normally generates conversations, to generate aggregated notification content information that indicates the aggregated notification content. The aggregated notification content that is instructed to be generated is a summary of the individual notifications indicated by each individual notification information included in the individual notification group information, based on the individual priority indicated by the individual priority information, the individual notification source agent indicated by the individual notification source information and the individual notification content indicated by the individual notification content information. Furthermore, if factor information is set, the factor information indicated by the factor information is summarized in detail. An example of the aggregated notification content of an aggregated notification will be described in detail later.

[0067] Aggregated notification content generation step S88 In the generation server 50, the conversation generation unit 52 inputs the aggregated notification content generation prompt received from the conversation provision server 10 along with the aggregated notification identification information to the conversation generation AI for generating normal conversations, receives the aggregated notification content information output from the conversation generation AI, and transmits the aggregated notification content information along with the aggregated notification identification information to the conversation provision server 10.

[0068] Summary notification content setting step S92 In the conversation provision server 10, the aggregated notification generation unit 44 sets the aggregated notification content information received from the generation server 50 along with the aggregated notification identification information as the aggregated notification content information of the aggregated notification associated with the aggregated notification identification information, and outputs the aggregated notification information to the aggregated notification provision unit 42.

[0069] Aggregated notification provision step S94 In the conversation provision server 10, when the aggregated notification provision unit 42 receives aggregated notification information from the aggregated notification generation unit 44, it transmits the aggregated notification information to the conversation provision server 10 at the notification time of the aggregated notification. Thus, in this embodiment, the aggregated notification provision unit 42 and the aggregated notification generation unit 44 form a notification output unit.

[0070] Aggregation notification step S98 In the user terminal 70, when the aggregate notification unit 92 receives aggregate notification information from the conversation provision server 10, it generates an aggregate notification image based on the aggregate notification content information using the image generation unit 78, and displays the aggregate notification image to the user using the image display unit 82. The aggregate notification image is displayed in the form of a push notification, a widget display, or the like. Thus, in this embodiment, the notification unit is formed by the aggregate notification unit 92, the image generation unit 78, and the image display unit 82.

[0071] As described above, for multiple individual notifications originating from multiple AI agents, each individual notification is assigned an individual priority based on the user's interests, and the factors that led to the higher priority are also identified. The individual notification sources and content of the multiple individual notifications are then summarized in detail based on the individual priority and the factors that led to the higher priority, and this is sent as an aggregated notification from the main agent. This avoids the user receiving individual notifications from multiple AI agents, prevents the aggregated notification from becoming cluttered with the content of individual notifications of low interest to the user, and allows the user to appropriately grasp the overview of the individual notifications of high interest from the aggregated notification. In this way, the main agent acts as if it were a personal secretary for the user, improving usability and enabling the building of a better relationship between the user and the main agent.

[0072] As shown in Figures 2 and 3E, the conversation provision system and conversation provision method select the current agent with whom the user is currently having a conversation from the individual notification source agents of each individual notification included in the aggregated notification, as described below, and initiate a conversation with that current agent, starting from the individual notification.

[0073] Response operation response step S102 On the user terminal 70, if the user responds to the aggregated notification image while the conversation application is not running, or if the user launches the conversation application while the aggregated notification image is displayed, the application launch unit 72 launches the conversation application. Furthermore, if the conversation application is launched by the application launch unit 72, or if the user responds to the aggregated notification image while the conversation application is running, the current agent selection unit 76 selects the main agent as the current agent to converse with the user at this time, the image generation unit 78 generates a main agent image representing the main agent, and the image display unit 82 displays the main agent image generated by the image generation unit 78 to the user. Responses to the aggregated notification image are performed by the user via voice input, touch operation, click operation, etc.

[0074] Aggregated notification voice command step S104 On the user terminal 70, the aggregate notification unit 92 inputs the aggregate notification content information to the conversation execution unit 88, and the conversation execution unit 88 sends an aggregate notification speech conversion instruction, which converts the aggregate notification content indicated by the aggregate notification content information into speech, to the conversation provision server 10 along with the aggregate notification content information, the main agent identification information, and the user identification information.

[0075] Aggregated notification voice conversion step S106 In the conversation server 10, the voice generation unit 22, in response to the aggregated notification speech instruction received from the user terminal 70 along with aggregated notification content information, main agent identification information, and user identification information, extracts voice information of agent feature information associated with the main agent identification information from the agent DB 16, generates aggregated notification content speech information in the voice of the main agent, and transmits it to the user terminal 70 identified by the user identification information.

[0076] Aggregated notification step S108 In the user terminal 70, the conversation execution unit 88 acquires the aggregated notification content speech information received from the conversation provision server 10 and outputs it to the voice output unit 86. The voice output unit 86 then outputs the aggregated notification content to the user as voice based on the aggregated notification content speech information input from the conversation execution unit 88.

[0077] Agent selection screen display step S112 On the user terminal 70, when the user requests the display of the agent selection screen, the aggregated notification unit 92 generates the agent selection screen using the image generation unit 78 based on the aggregated notification information, and displays the selection screen generated by the image generation unit 78 to the user using the image display unit 82. The user's request to display the agent selection screen is made by voice input or the like. The agent selection screen is for selecting the current agent from the individual notification source agents indicated by the individual notification source information of each individual notification information included in the individual notification group information of the aggregated notification information. In this embodiment, the agent selection screen displays a list of selection icons as options for selecting the current agent, combining the individual notification source agents indicated by the individual notification source information of each individual notification information included in the individual notification group information and the individual notification content indicated by the individual notification content information. An example of the agent selection screen will be described in detail later.

[0078] Selection operation corresponding step S114 On the user terminal 70, if the current agent is selected by the user's current agent selection operation, the current agent selection unit 76, which acts as a modification unit, sets current agent identification information to identify the current agent. The user's operation to select the current agent is performed by voice input, or, if the agent selection screen is displayed, by touching or clicking the selection icons on the agent selection screen. The image generation unit 78 generates a current agent image representing the current agent, and the image display unit 82 displays the current agent image generated by the image generation unit 78 to the user.

[0079] Individual notification voice instruction step S116 On the user terminal 70, the aggregated notification unit 92 extracts individual notification information from the group of individual notification information included in the aggregated notification information, including the agent identification information of the individual notification source agent corresponding to the current agent identification information set by the current agent selection unit 76. The aggregated notification unit 92 then sends an individual notification voice conversion instruction to the conversation provision server 10, along with the individual notification content information, the current agent identification information, and the user identification information, to convert the content of the extracted individual notification information into voice.

[0080] Individual notification voice conversion step S118 In the conversation provision server 10, the voice generation unit 22, in response to the individual notification voice conversion instruction received from the user terminal 70 along with the individual notification content information, current agent identification information, and user identification information, extracts voice information of agent feature information associated with the current agent identification information from the agent DB 16, generates individual notification content voice conversion information in which the individual notification content indicated by the individual notification content information is spoken using the voice of the current agent, and transmits it to the user terminal 70 identified by the user identification information.

[0081] Individual notification step S122 In the user terminal 70, the conversation execution unit 88 acquires the individual notification content speech information received from the conversation provision server 10 and outputs it to the voice output unit 86. The voice output unit 86 then outputs the individual notification content to the user as voice based on the individual notification content speech information input from the conversation execution unit 88.

[0082] Subsequent conversations between the user and the current agent are conducted in the same manner as the current agent's responses to user utterances as described above. However, the response generation prompts generated by the conversation generation instruction unit 24 of the conversation provision server 10 are based not only on conversation history information, character feature information, and field identification information, but also on individual notification-related information indicating information related to the individual notification that initiated the conversation.

[0083] As described above, in response to the user's selection of the current agent based on the aggregated notification sent from the main agent, the current agent with whom the user is currently conversing can be changed from the main agent to one of the individual agents that issued the individual notifications aggregated in the aggregated notification. This allows for a smooth change of conversation partner for the user. In this way, the main agent acts as a personal assistant for the user, introducing them to the individual notification-generating agents, improving usability and enabling the building of a better relationship between the user and the main agent.

[0084] Referring to Figure 4, the method for setting the individual priority and contributing factors for individual notifications in this embodiment will be explained. As shown in Figure 4, for example, the individual priorities and contributing factors for individual notifications are set as follows. (1) Setting of individual priority and contributing factors based on the level of familiarity between the individual notification recipient user and the individual notification source agent. If a user receiving individual notifications has a high level of familiarity with the individual notification source agent, it is assumed that the user will also have a high level of interest in the individual notification. Therefore, the individual priority should be set high, and the individual notification source agent itself should be set as the contributing factor.

[0085] (2) Setting of individual priorities and contributing factors based on the importance of the individual notification source agent for the individual notification recipient user. If the individual notification source agent is highly important to the individual notification recipient user, it is assumed that the recipient user is also highly interested in the individual notification. Therefore, the individual priority is set high, and the individual notification source agent itself is set as a contributing factor. For example, if the field handled by the AI ​​agent is a public field such as taxation, and the client operating the AI ​​agent is a public institution such as a tax office, the individual notification source agent will be highly important to the individual notification recipient user, the individual priority will be set high, and the AI ​​agent itself, representing a public institution, will be set as a contributing factor.

[0086] (3) Setting individual priorities and contributing factors based on the urgency of the individual notification content for the individual notification recipient user. If the content of an individual notification is of high urgency to the recipient user, it is assumed that the recipient user will also be of high interest in the notification, and therefore, a high individual priority will be set, with the deadline set as a contributing factor. For example, if the content of an individual notification concerns a user's schedule and the start time of the schedule is approaching, the urgency of the content of the individual notification to the recipient user will be high, a high individual priority will be set, and the start time of the schedule will be set as a contributing factor. Similarly, if the content of an individual notification concerns securities trading and requires immediate attention, the urgency of the content of the individual notification to the recipient user will be high, a high individual priority will be set, and the need for immediate attention will be set as a contributing factor.

[0087] (4) Setting of individual priorities and contributing factors based on the importance of the individual notification content to the individual notification recipient user. If the content of an individual notification is of high importance to the recipient user, it is assumed that the recipient user is also highly interested in the notification. Therefore, a high individual priority is set, and the important content of the individual notification is set as a contributing factor. For example, if the individual notification concerns the payment of utility bills and requires action within a certain period, the content of the individual notification will be of high importance to the recipient user, a high individual priority will be set, and the payment of utility bills will be set as a contributing factor.

[0088] (5) Setting individual priorities and factors based on the level of interest of the individual notification recipients in the content of individual notifications. If a user who is a recipient of an individual notification shows a high level of interest in the content of that notification, it is assumed that the user is also highly interested in the notification itself. Therefore, the individual notification will be given a high priority, and the content of the notification will be set as a contributing factor. For example, if the content of the individual notification concerns a topic that the recipient has recently shown strong interest in, the user will show a high level of interest in the notification, the individual notification will be given a high priority, and the content of that topic will be set as a contributing factor.

[0089] Referring to Figure 5, the individual notification contents and contributing factors, as well as the aggregated notification contents, will be explained. As shown in Figure 5, for example, the contents and contributing factors of individual notifications, as well as the contents of aggregated notifications, are as follows.

[0090] (1) Content of individual notifications and the factors contributing to them (1-1) Tax Office Agent E Regarding Tax Office Agent E, whose area of ​​expertise is tax-related and whose client is a tax office: • Individual notification content: "The deadline for filing this year's tax return is March 17th. Since you have not yet completed your filing, please file as soon as possible." This individual notification is considered highly important by the individual notification source agent to the individual notification recipient user, and individual priority is set according to that importance. The contributing factor is the individual notification source agent itself. • Factor: "Tax Office Agent E"

[0091] (1-2) Schedule Agent F Regarding Schedule Agent F, an AI agent whose area of ​​expertise is user schedule management. Individual notification content: "A meeting is scheduled for 3:00 PM today. It is currently 2:30 PM, so please prepare for the meeting." • For this individual notification, the urgency of the content for the individual notification recipient is high, and individual priorities are set according to that urgency. The contributing factors will be the deadlines. • Factor: "Meeting scheduled for 3 PM today"

[0092] (1-3) Securities Agent G Regarding Securities Agent G, whose AI agent's area of ​​expertise is securities-related and whose clients are securities companies: • Individual notification: "The share price of your owned stocks has risen significantly since the time of purchase. If you wish to sell your stocks early, please contact us as soon as possible." • For this individual notification, the urgency of the content for the individual notification recipient is high, and individual priorities are set according to that urgency. The contributing factors will be the deadlines. • Factor: "Early sale of shares"

[0093] (1-4) Waterworks Agent H Regarding Waterworks Bureau Agent H, whose area of ​​expertise is related to water supply and whose client is a waterworks bureau: Individual notification content: "This month's water bill is X yen. The payment deadline is the 25th, so please make your payment." • For this individual notification, the content of the notification is of high importance to the recipient user, and individual priorities are set according to that importance. The contributing factors will be the important items contained in the individual notification. • Factor: "Payment of water bills"

[0094] (1-5) Sports Agent I Regarding Sports Agent I, whose area of ​​expertise is sports-related normal conversation: • Individual notification content: "In the US Y Series, player Z hit a walk-off home run, and team V came from behind to win 2-3." Regarding this individual notification, if the recipient user has a high level of interest in the content of the notification, an individual priority will be set according to that level of interest, and the contributing factors will be the content of the points of interest in the individual notification. • Factor: "US Y Series"

[0095] (2) Contents of the aggregated notification For example, including the individual notifications (1-1) through (1-5) above, if (1-1) has a particularly high priority, (1-3) has a high priority, (1-2) has a relatively high priority, (1-4) has a relatively low priority, and (1-5) has a low priority: Summary of the notification: "Tax office agent E has notified you of the tax filing deadline of March 17th, G has confirmed your intention to sell your stock early due to stock price fluctuations, F has informed you about preparations for the meeting at 3 PM, and there are other notifications from H and I."

[0096] Referring to Figure 6, the agent selection screen of one embodiment of the present invention will be described. As shown in Figure 6, for example, on the agent selection screen, each individual notification that was aggregated into a single notification is displayed in a list of icons, each containing the individual notification content attached to an icon representing the agent that sent the individual notification.

[0097] In the embodiment described above, for multiple individual notifications to a user originating from multiple AI agents, the individual notification sources and content of the multiple individual notifications are summarized and compiled into an aggregated notification, which is then sent to the user by the main agent as the notification entity. This avoids the user receiving individual notifications from multiple AI agents, and also avoids the aggregated notification becoming cluttered with the content of individual notifications of low user interest, thereby improving usability. In particular, the aggregated notification is summarized and compiled based on the user's familiarity with each AI agent, the importance of each AI agent to the user, the urgency and importance of each individual notification originating from each AI agent to the user, and the user's level of interest, thus avoiding clutter from individual notifications from AI agents with low user familiarity or low importance to the user, or from the content of individual notifications of low urgency, importance, or user interest.

[0098] Furthermore, aggregated notifications provide a detailed summary of the individual notification content of high-priority individual notifications, including the factors that contributed to their high priority. This makes it possible to appropriately grasp the overview of individual notifications of high interest to the user from the aggregated notifications.

[0099] In addition, based on aggregated notifications sent from the main agent, the system allows users to select the current agent with whom they are currently conversing, changing from the main agent to one of the individual agents that sent the individual notifications aggregated in the aggregated notification. This facilitates smooth changes in the conversation partner for the user, further improving usability.

[0100] In the embodiment described above, speech audio information representing the user's utterance is converted into speech audio text data, a response generation prompt including the speech audio text data is generated and input to a conversation generation AI, the response text data is output from the conversation generation AI, and the response text data is converted into response audio information representing the AI ​​agent's response based on the AI ​​agent's voice information. Alternatively, a response generation prompt including speech audio information representing the user's utterance and the AI ​​agent's voice information may be generated and input to a conversation generation AI, and response audio information representing the AI ​​agent's response may be output from the conversation generation AI. In this case, the speech audio information and response audio information may be stored as conversation history information.

[0101] The key disclosures of this application can be summarized as follows: The first disclosure is a conversation provision system comprising: a conversation unit that instructs a task processing device, which processes tasks using artificial intelligence in response to task processing instructions, to generate conversation content for at least one of a plurality of virtual agents to a user, and acquires the conversation content generated by the task processing device and outputs it to the user; and a notification unit that, with the primary agent among the plurality of virtual agents acting as the notification subject, notifies the user of a consolidated notification in which the individual notification sources and individual notification content of a plurality of individual notifications to the user, with the individual notification source and individual notification content of the plurality of individual notifications being individually notified.

[0102] In this disclosure, regarding multiple individual notifications sent to a user from multiple virtual agents, a consolidated notification is sent to the user, with the main agent acting as the notification entity, summarizing and combining the individual notification sources and content of each individual notification. This avoids the user receiving individual notifications from multiple virtual agents, thus improving usability.

[0103] The second disclosure item is the conversation provision system described in the first disclosure item, in which the aggregated notification summarizes and compiles the individual notification sources and content of the multiple individual notifications based on priority according to the user's interest information.

[0104] In this disclosure, aggregated notices are compiled by summarizing and combining the individual notice sources and content of multiple individual notices based on priority according to the user's interests. This avoids the aggregated notice becoming cluttered with the content of individual notices of low user interest, thereby further improving usability.

[0105] The third disclosure is the conversational delivery system of the second disclosure, in which the interest information includes the user's level of familiarity with each of the virtual agents.

[0106] In this disclosure, aggregated notifications are summarized and compiled based on a priority level corresponding to the user's familiarity with each virtual agent, thus avoiding the cumbersome nature of individual notifications from virtual agents with low familiarity levels for the user.

[0107] The fourth disclosure is the conversational delivery system of the second disclosure, in which the interest information includes the importance of each virtual agent to the user.

[0108] In this disclosure, aggregated notifications are summarized and compiled based on priority according to the importance of each virtual agent to the user, thus avoiding the cumbersome nature of individual notifications from virtual agents of lower importance to the user.

[0109] The fifth disclosure item is the conversational delivery system of the second disclosure item, which includes the urgency to the user of each individual notification originating from each virtual agent, as the interest information.

[0110] In this disclosure, the aggregated notices are summarized and compiled based on the priority of each individual notice sent from each virtual agent, according to the urgency to the user. This avoids the issue of the notices becoming cumbersome due to the content of individual notices that are of lower urgency to the user.

[0111] The sixth disclosure item is the conversation provision system of the second disclosure item, which includes the importance to the user of each individual notification originating from each virtual agent, as the interest information.

[0112] In this disclosure, the aggregated notices are summarized and compiled based on the priority of each individual notice sent from each virtual agent, according to their importance to the user. This avoids the issue of the notices becoming cumbersome due to the content of individual notices of low importance to the user.

[0113] The seventh disclosure item is the conversation provision system described in the second disclosure item, which includes the user's level of interest in each individual notification sent from each virtual agent.

[0114] In this disclosure, aggregated notifications are summarized and compiled based on the user's level of interest in each individual notification sent from each virtual agent, thus avoiding the issue of notifications becoming cumbersome due to the content of individual notifications of low user interest.

[0115] The eighth disclosure item is the conversation provision system described in the second disclosure item, which, in the aggregated notice, provides a detailed summary of the individual notices of the high-priority individual notices, including the factors that contributed to their high priority.

[0116] In this disclosure, the aggregated notice provides a detailed summary of the content of high-priority individual notices, including the factors that contributed to their high priority. Therefore, it is possible to appropriately grasp the overview of the content of individual notices of high interest to the user from the aggregated notice.

[0117] The ninth disclosure is the conversation provision system of the first disclosure, further comprising a modification unit that, in response to a request from the user based on the aggregated notice notified by the notification unit, changes the virtual agent that acts as the conversation partner for the user from the main agent to one of the virtual agents that are the source of the individual notices included in the aggregated notice.

[0118] This disclosure explains that, in response to user requests based on aggregated notifications sent from the main agent, the virtual agent that the user is conversing with is changed from the main agent to one of the individual virtual agents that are the source of the individual notifications included in the aggregated notification. This allows for smooth changes in the user's conversation partner and further improves usability.

[0119] The tenth disclosure is a conversation providing device comprising: a conversation output unit that instructs a task processing device, which processes tasks using artificial intelligence in response to task processing instructions, to generate conversation content for at least one of a plurality of virtual agents to a user, and acquires and outputs the conversation content generated by the task processing device; and a notification output unit that, for a plurality of individual notifications to a user, with the plurality of virtual agents being the individual notification sources, outputs an aggregated notification in which the individual notification sources and individual notification content of the plurality of individual notifications are summarized and combined, with the main agent among the plurality of virtual agents being the notification subject. This disclosure has the same effect as the first disclosure. [Explanation of symbols]

[0120] 10...Conversation server 12...Agent designation unit 14...User database 16…Agent DB 18…Voice Analysis Unit 22…Voice Generation Unit 24…Conversation Generation Instructions 26...Conversation analysis instruction unit 28...Intimacy setting unit 32...User preference setting unit 34...Individual notification generation unit 36...Individual notification acquisition unit 38...Priority setting unit 42... Aggregated Notification Provision Unit 44... Aggregated Notification Generation Unit 50... Generation Server 52... Conversation Generation Unit 54...Conversation Analysis Unit 60...Client Terminal 62...Individual Notification Provision Unit 70...User Terminal 72...App launch section 74...Main agent selection section 76...Current agent selection section 78...Image generation unit 82...Image display unit 84...Audio acquisition unit 86...Audio output unit 88...Conversation execution unit 92...Aggregation and notification unit

Claims

1. A task processing device that processes tasks using artificial intelligence in response to task processing instructions, a conversation unit that instructs the task processing device to generate conversation content for at least one of multiple virtual agents interacting with a user, and acquires the conversation content generated by the task processing device and outputs it to the user, Regarding the multiple individual notifications to the user, each of the multiple virtual agents being used as an individual notification source, the notification unit summarizes and combines the individual notification sources and content of the multiple individual notifications based on priority according to the user's interest information, and notifies the user of an aggregated notification, with the main agent among the multiple virtual agents acting as the notification entity. A conversation provision system equipped with the following features.

2. The interest information includes the user's intimacy with each of the virtual agents, The conversation provision system according to claim 1.

3. The interest information includes the importance of each virtual agent to the user, The conversation provision system according to claim 1.

4. The interest information includes the urgency for the user of each individual notification from each virtual agent, The conversation provision system according to claim 1.

5. The interest information includes the importance to the user of each individual notification from each virtual agent, The conversation provision system according to claim 1.

6. The interest information includes the user's level of interest in each individual notification from each virtual agent, The conversation provision system according to claim 1.

7. The aggregated notification provides a detailed summary of the individual notification content of the high-priority individual notification, including the factors that contributed to the high priority. The conversation provision system according to claim 1.

8. A task processing device that processes a task using artificial intelligence in response to a task processing instruction, comprising a conversation unit that instructs the generation of conversation content for at least one of a plurality of virtual agents to a user, and acquires the conversation content generated by the task processing device and outputs it to the user, Regarding the multiple individual notifications to the user, each of the multiple virtual agents being the individual notification source, a notification unit is provided that notifies the user of an aggregated notification, which summarizes and combines the individual notification sources and content of the multiple individual notifications, with the main agent among the multiple virtual agents acting as the notification entity. It is equipped with, The system further comprises a modification unit that, in response to a user request based on the aggregated notification notified by the notification unit, changes the virtual agent that acts as the user's conversation partner from the main agent to one of the virtual agents that are the source of the individual notifications included in the aggregated notification. A conversation provision system.

9. A task processing device that processes a task using artificial intelligence in response to a task processing instruction, comprising a conversation output unit that instructs the generation of conversation content for at least one of a plurality of virtual agents to a user, and acquires and outputs the conversation content generated by the task processing device, A notification output unit outputs an aggregated notification, which summarizes and combines the individual notifications sent to a user, with the individual notification source and content of each individual notification being prioritized according to the user's interest information, with the main agent among the multiple virtual agents acting as the notification entity. A conversation-providing device equipped with the following features.

Images