JP7794405B2 - Scenario generation device, scenario generation method, and scenario generation program - Google Patents

Description

The present invention relates to a scenario generation device, a scenario generation method, and a scenario generation program.

Chatbots, communication robots that can automatically hold conversations with users over chat, have been known for some time. In recent years, efforts have been made to utilize chatbot technology to provide information useful for decision-making regarding the purchase of various products and services, and to carry out the procedures required for purchases through automated conversations.

Japanese Patent Application Laid-Open No. 2022-77779

In conventional technology, in order to use a chatbot to have a conversation with a user, a designer (communication designer) had to manually create in advance a scenario that specified a series of conversations with the user (see Patent Document 1). In such a scenario, the designer had to specify a conditional branching structure (tree-type structure) based on the expected content of the user's conversation, and specify a separate series of conversation flows for each condition, so that the next conversation the chatbot would respond to would be variable depending on the content of the user's conversation. Depending on the purpose of using the chatbot, scenarios created in this way could become extremely complex, which increased the workload of preparing the scenarios.

Meanwhile, there are known methods for automating responses to user inquiries by utilizing machine learning techniques such as Large Language Models (LLM). However, these conventional methods are only capable of responding to one-off inquiries from users, and are unable to generate scenarios that define a series of conversations with the user.

The present invention was made in consideration of these circumstances, and one of its objectives is to provide a scenario generation device, a scenario generation method, and a program that can reduce the effort required to prepare scenarios to be executed by chatbots.

One aspect of the present invention is a scenario generation device that generates a scenario that specifies a series of conversational content to be executed by a chatbot to provide a user with information about a product or service. The scenario generation device includes a scenario generation unit that inputs specific information that identifies a product or service specified by a scenario designer into a first learning model prepared by learning previously generated scenarios, and generates a scenario corresponding to the specific information.

One aspect of the present invention can reduce the effort required to prepare scenarios to be executed by a chatbot. Furthermore, by adopting a configuration in which scenarios are generated using a learning model generated by studying previously generated scenarios, the quality of the generated scenarios can be improved.

FIG. 1 is a diagram illustrating an example of a configuration for realizing a chatbot service according to a first embodiment. FIG. 1 is a functional block diagram showing an example of the configuration of a chatbot server 1 according to a first embodiment. A figure explaining input and output data of the first learning model (at the time of scenario generation) according to the first embodiment. A figure explaining input and output data of the second learning model (during auto-chat) related to the first embodiment. A figure explaining input and output data of the third learning model (individual offer generation) related to the first embodiment. FIG. 2 is a functional block diagram showing an example of the configuration of a user terminal device 100 according to the first embodiment. FIG. 2 is a diagram showing an example of a conversation screen P1 displayed on the user terminal device 100 according to the first embodiment. 10 is a flowchart showing an example of a sequence of steps in a scenario generation process performed by the chatbot server 1 according to the first embodiment. FIG. 10 is a diagram showing an example of a scenario condition setting screen P2 displayed on the operation terminal device 200 according to the first embodiment. A figure showing an example of a scenario confirmation screen (1) P3-1 displayed on the operation terminal device 200 according to the first embodiment. A figure showing an example of a scenario confirmation screen (2) P3-2 displayed on the operation terminal device 200 according to the first embodiment. A figure showing an example of a scenario confirmation screen (3) P3-3 displayed on the operation terminal device 200 according to the first embodiment. A figure showing an example of a scenario confirmation screen (4) P3-4 displayed on the operation terminal device 200 according to the first embodiment. FIG. 10 is a sequence diagram illustrating an example of a scenario display process according to the first embodiment. FIG. 10 is a diagram showing an example of response data RD according to the first embodiment. A figure showing an example of object type scenario data SD related to the second embodiment. FIG. 10 is a diagram illustrating an example of an object structure according to the second embodiment.

Embodiments of the scenario generation device, scenario generation method, and scenario generation program of the present invention will be described below with reference to the drawings. The scenario generation device of the embodiment uses a machine learning-based technique such as LangChain that uses a large-scale language model (LLM), thereby reducing the effort required to prepare a scenario that defines a series of conversational content to be executed by a chatbot to provide users with information about products or services.

[First embodiment]
<Chatbot service>
FIG. 1 illustrates an example of a configuration for implementing a chatbot service according to the first embodiment. The chatbot service uses a chatbot to automatically converse with users, such as businesses providing products or services (hereinafter referred to as "products, etc."), to learn about the user's hobbies, preferences, and personal background, and to suggest information on products, etc., tailored to the user. Examples of businesses include cosmetics retailers, food retailers, hospitals, human resources service providers, financial institutions, insurance companies, real estate agents, and telecommunications companies. The chatbot service is implemented primarily around a chatbot server 1. The chatbot server 1 communicates with, for example, one or more service servers 3, one or more message servers 5, one or more user terminal devices 100, and one or more operation terminal devices 200 via a communication network NW. The communication network NW includes, for example, the Internet, a local area network (LAN), a wireless base station, a provider device, and the like.

<Device configuration>
<Chatbot Server 1>
The chatbot server 1 provides a function for generating a scenario to a designer D who operates an operation terminal device 200. The chatbot server 1 also provides a chatbot service to a user U who operates a user terminal device 100. Figure 2 is a functional block diagram showing an example of the configuration of the chatbot server 1 according to the first embodiment. The chatbot server 1 includes, for example, a communication unit 11, a control unit 13, and a memory unit 15. The chatbot server 1 is an example of a "scenario generation device." The communication unit 11 is a communication interface for connecting to a communication network NW. The communication unit 11 is, for example, a network interface card.

The control unit 13 includes, for example, an acquisition unit 131, a conversation generation unit 132, a transmission unit 133, a scenario generation unit 134, an auto-chat unit 135, an individual offer generation unit 136, a display control unit 137, and a learning unit 138. The components of the control unit 13 are realized, for example, by a hardware processor such as a CPU executing a program (software). Some or all of these components may be realized by hardware such as an LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), or GPU (Graphics Processing Unit), or may be realized by a combination of software and hardware. The program may be stored in advance in a storage device (non-transitory computer-readable storage medium) such as an HDD (Hard Disk Drive) or flash memory, or may be stored in a removable storage medium (non-transitory computer-readable storage medium) such as a DVD or CD-ROM, and installed in the storage device by inserting the storage medium into a drive.

The acquisition unit 131 acquires various information (data) from the service server 3, message server 5, user terminal device 100, and operation terminal device 200 via the communication network NW. For example, the acquisition unit 131 acquires conversation information indicating the content of the conversation entered in response to the operation of user U from the user terminal device 100. Also, for example, the acquisition unit 131 acquires instruction information regarding scenario generation entered in response to the operation of designer D from the operation terminal device 200. The acquisition unit 131 is an example of an "acquisition unit."

The conversation generation unit 132 generates conversation text to be sent to the user terminal device 100 (conversation text to be presented to the user) based on conversation information acquired from the user terminal device 100 and scenario data SD for each provider pre-stored in the storage unit 15. The conversation generation unit 132 is an example of a "conversation generation unit."

The transmission unit 133 transmits conversation sentence data indicating the conversation sentence generated by the conversation generation unit 132 to the user terminal device 100. The transmission unit 133 is an example of a "transmission unit."

The scenario generation unit 134 generates a scenario using the first learning model AI1 pre-stored in the memory unit 15 based on instruction information related to scenario generation acquired from the operation terminal device 200, and stores the generated scenario as scenario data SD in the memory unit 15. The first learning model AI1 is a large-scale language model (LLM) generated by learning large amounts of data as learning data, such as past scenario data previously generated manually or by the scenario generation unit 134 and conversation history data with past users (past chat data). The first learning model AI1 is a model that outputs text in which words that are highly likely to follow specific words are predicted and arranged based on the input text data. Examples of the first learning model AI1 include ChatGPT (registered trademark), LaMDA, etc. Each conversational sentence included in the scenario may include emoticons.

Figure 3 is a diagram illustrating input and output data for the first learning model AI1 (during scenario generation) according to the first embodiment. As shown in Figure 3, when used for scenario generation, the first learning model AI1 receives input data as specific information identifying a product (e.g., the website address of the business providing the product, product information indicating the product details, etc.), and outputs the generated scenario data. By using the website address as input data, information about the company, brand, and product provided on the web is read via web scraping, and products that can be suggested in chat are selected as product information, and their prices and features are read. The first learning model AI1 may also receive input data such as purpose information indicating the purpose of providing users with information about the product (e.g., whether the information is intended to encourage purchases of the product (sales, conversion), or advertising the product), image information about the product or business, target information indicating the type of users to whom the generated scenario is to be provided (e.g., all members, men only, women only, etc.), the type of conversation application platform used for distribution (e.g., conversation application service A, conversation application service B, etc.), and the user U's entry path (e.g., pop-ups, trigger words, etc.). The purpose information is used to determine whether the overall purpose of the conversation is sales or something else. By storing a group of usable images, appropriate images are selected from the scenario and reflected in the scenario data. The input data for the first learning model AI1 may be obtained from information published on the web (e.g., information from an e-commerce store) via an API (Application Programming Interface). The scenario generation unit 134 is an example of a "scenario generation unit." That is, the scenario generation unit 134 inputs specific information identifying a product or service specified by the scenario designer D into the first learning model AI1, which is prepared by learning previously generated scenarios, to generate a scenario corresponding to the specific information. The scenario generation unit 134 also inputs purpose information specified by the operator into the first learning model AI1 to generate a scenario corresponding to the purpose information. The scenario generation unit 134 also inputs images related to a product or service specified by the designer D into the first learning model AI1 to generate a scenario in which images are incorporated into a series of conversational content. The scenario generation unit 134 also edits the generated scenario in response to editing instructions from designer D on the scenario confirmation screen displayed on the display unit.

Returning to FIG. 2, the auto-chat unit 135 inputs the user input text entered as free text by the user U when presenting the scenario to the second learning model AI2, thereby generating a response to the user input text (smart reply function). The second learning model AI2 is a large-scale language model (LLM) generated by learning large amounts of data, such as past conversation history data with users (past chat data), past scenario data generated manually or by the scenario generation unit 134, and other text data, as learning data. The second learning model AI2 is a model that predicts and arranges words that are likely to follow specific words based on the input text data. Examples of the second learning model AI2 include ChatGPT (registered trademark) and LaMDA. FIG. 4 is a diagram explaining the input and output data of the second learning model (during auto-chat) according to the first embodiment. As shown in FIG. 4, when used to generate auto-chat, the input data of the second learning model AI2 is the user input text entered by the user U, and the output data is a response. To prevent hallucination, the auto-chat unit 135 has a checking function that excludes answer sentences output from the second learning model AI2 if the answer sentences are not related to the sale or purchase of goods, etc. The auto-chat unit 135 is an example of an "auto-chat unit." That is, the auto-chat unit 135 generates answers to the free-text conversation sentences specified by the user U when the generated scenario is presented to the user U, by inputting these conversation sentences into the second learning model AI2.

Returning to FIG. 2 , the individual offer generation unit 136 uses the third learning model AI3 to output personalized product suggestions (hereinafter, "individual offers") for the user U. The third learning model AI3 is a machine learning model, such as a neural network, that is trained to output personalized products (products that the user U is likely to prefer and purchase) for the user U when the user data of the user U is input. The user data includes, for example, gender, age, and past purchase history. The user data is, for example, purchase data (e.g., name, address, email address, telephone number, identification ID (social handle), etc. based on past purchase history, also referred to as "first-party data") previously accumulated in a CRM (Customer Relationship Management) or CDP (Customer Data Platform) by the business operator managing the service server 3. The chatbot server 1 may import this purchase data accumulated in the service server 3, for example, via an API. The user data may be linked to an identification ID (social handle) issued in a conversation application to identify an individual. Instead of a neural network, the third learning model AI3 may be another machine learning model such as a support vector machine, a decision tree, a random forest, or a logistic regression. FIG. 5 is a diagram illustrating input and output data of the third learning model AI3 (individual offer generation) according to the first embodiment. As shown in FIG. 5, the input data of the third learning model AI3 is user data of a user U, and the output data is an individual offer. The input data may include a conversation history with the user U according to a scenario. The individual offer generation unit 136 is an example of an "individual offer generation unit." That is, the individual offer generation unit 136 generates a product or service proposal tailored to the user U by inputting the user data of the user U who presented the generated scenario to the third learning model AI3, which has been trained to output information about a product or service tailored to the specific user when the user data of the specific user is input.

The display control unit 137 generates information for displaying various operation screens for generating a scenario, transmits this information to the operation terminal device 200, and causes the operation terminal device 200 to display the various operation screens. The display control unit 137 is an example of a "display control unit." In other words, the display control unit 137 causes the display unit to display a confirmation screen for the scenario generated by the scenario generation unit 134.

The learning unit 138 performs learning using large amounts of data, such as past scenario data and past conversation history data generated in the past, as learning data, to generate a first learning model AI1, which is stored in the memory unit 15. The learning unit 138 also performs learning using large amounts of data, such as past conversation history data with users, past scenario data generated in the past manually or by the scenario generation unit 134, and other text data, to generate a second learning model AI2, which is stored in the memory unit 15. The learning unit 138 separates and tokenizes the text data included in the learning data into its smallest units, learns the context and meaning of words from the obtained tokens, and generates the first learning model AI1 and the second learning model AI2.

The learning unit 138 also generates a third learning model AI3 by learning learning data that is a combination of user data of users who have purchased products, etc. in the past and the products, etc. purchased by those users (purchase history data), and stores this in the memory unit 15. For example, the learning unit 138 learns from purchasing data (first-party data) that the business operator managing the service server 3 has previously accumulated in CRM or CDP as learning data.

The memory unit 15 is a hard disk drive (HDD), flash memory, RAM (Random Access Memory), etc. The memory unit 15 may be a NAS (Network Attached Storage) device accessible by the chatbot server 1 via a network. The memory unit 15 stores the first learning model AI1, the second learning model AI2, the third learning model AI3, scenario data SD, product data ID, brand data BD, user data UD, response data RD, etc. The product data ID is obtained from the service server 3 or another sales management server via the communication network NW and stored in the memory unit 15. The brand data BD is obtained from the service server 3 and stored in the memory unit 15. The brand data BD may also be stored in the memory unit 15 by collecting information about the business's brand from the business's homepage or a CVS-format list provided by the business. The user data UD is obtained from the service server 3 and stored in the memory unit 15. The response data RD is obtained from the user during a conversation with the user based on the scenario and stored in the memory unit 15.

<Service Server 3>
The service server 3 is managed by a business that provides products and the like. The service server 3 provides the user terminal device 100 with a site such as a homepage where the user can complete the purchase procedure for the product and the like. When the user U performs a preset trigger operation on this site, processing for providing the chatbot service is initiated. Examples of the trigger operation include pressing a button that starts the chatbot service displayed in a pop-up on the site, pressing an advertising banner displayed on the conversation application, or pressing an advertising display placed by a company that uses the conversation application.

<Message Server 5>
The message server 5 provides a message service (chat service) that realizes a conversation function by sending and receiving messages to the user terminal device 100, the chatbot server 1, and the service server 3. The message server 5 provides a conversation application that can be installed in the user terminal device 100 to the user terminal device 100, and the user terminal device 100 sends and receives messages with other user terminal devices 100 and various servers by executing this conversation application. The message server 5 also provides an API that enables use of the message service from an external device (server), and the chatbot server 1 and the service server 3 send and receive messages with the user terminal device 100 by using this API.

<User Terminal Device 100>
The user terminal device 100 is operated by a user U who purchases and uses products and the like provided by a business operator. The user terminal device 100 is, for example, a portable terminal device such as a smartphone or a tablet terminal. FIG. 6 is a functional block diagram showing an example of the configuration of the user terminal device 100 according to the first embodiment. The user terminal device 100 includes, for example, a communication unit 101, a display unit 103, an operation unit 105, a control unit 107, and a storage unit 109.

The communication unit 101 is a wireless communication module that performs wireless communication with a wireless base station connected to the communication network NW. The communication unit 101 communicates with the chatbot server 1, service server 3, message server 5, etc. via the wireless base station. The display unit 103 is a display device such as a liquid crystal display device. The operation unit 105 is an input interface that accepts operation instructions from the user U. Note that the display unit 103 and operation unit 105 may be configured as touch panel displays.

The control unit 107 controls the overall operation of the user terminal device 100. The functions of the control unit 107 are realized, for example, by a hardware processor such as a CPU executing a program (software). The control unit 107 launches a conversation application (program) AP stored in the memory unit 109 to realize a conversation function via a network. The conversation application may be an application dedicated to conversation, or it may also have other functions (such as a call function, an image upload function, and a message upload function). For example, the conversation application may be an application on the provider's side that is provided by a business that provides products, etc. and installed on the user terminal device 100 and that incorporates a conversation function. Alternatively, the conversation application may be realized using a general-purpose application program such as a web browser. For example, the conversation application may be realized using a chat function on the provider's homepage provided by a service server 3 managed by the business.

Figure 7 is a diagram showing an example of a conversation screen P1 displayed on the user terminal device 100 according to the first embodiment. The example in Figure 7 shows a conversation screen P1 in which a chatbot service provided by business operator A is started, and a scenario is presented in which a greeting message and questions are sent to user U. By conversing with the chatbot in accordance with this scenario, user U can receive information tailored to them and suggestions for products, etc.

<Operation Terminal Device 200>
The operation terminal device 200 is operated by a designer D who designs a scenario for a chatbot service. The operation terminal device 200 is, for example, a personal computer, a smartphone, a tablet terminal, etc. The operation terminal device 200 has at least a display function, an input function, a storage function, etc.

<Processing flow>
<Scenario generation process>
Next, various processes executed in the chatbot service will be described. First, the process on the operation side in which designer D generates a scenario will be described. FIG. 8 is a flowchart showing an example of a series of scenario generation processes by the chatbot server 1 according to the first embodiment. The process shown in FIG. 8 is started, for example, when designer D operates the operation terminal device 200 to input an instruction to start scenario generation, and the chatbot server 1 receives this start instruction.

(Step S101) First, the display control unit 137 of the chatbot server 1 displays a scenario condition setting screen on the operation terminal device 200. Then, the acquisition unit 131 acquires from the operation terminal device 200 the setting information entered by designer D into the scenario condition setting screen displayed on the operation terminal device 200.

9 is a diagram showing an example of a scenario condition setting screen P2 displayed on the operation terminal device 200 according to the first embodiment. The example scenario condition setting screen P2 shown in FIG. 9 includes a first column AR1 for accepting the URL of the business for which the scenario is to be generated, a second column AR2 for accepting the objective information of the scenario, a third column AR3 for accepting the target for whom the scenario is to be provided, a fourth column AR4 for accepting whether or not to accept free text input from the user, a fifth column AR5 for accepting the designation of products, etc. to be proposed to the user in the scenario, and a sixth column AR6 for accepting images related to products or services (images of businesses, products, etc.). Designer D can enter scenario conditions into these columns by operating the input interface (keyboard, mouse, etc.) of the operation terminal device 200. Other scenario conditions may include the number of characters in the conversation included in the scenario and the conversion rate. When the scenario generation button BT1 is pressed, the setting information entered by designer D is sent from the operation terminal device 200 to the chatbot server 1.

(Step S103) Next, the scenario generation unit 134 generates a scenario using the first learning model AI1 pre-stored in the memory unit 15 based on the setting information specified by the designer D transmitted from the operation terminal device 200. The scenario generation unit 134 generates a scenario by inputting the business's homepage address, product information, purpose information, image information, and target information contained in the setting information into the first learning model AI1. There are, for example, two types of scenarios generated here. The first scenario, also known as an initial conversation, is a conversation scenario that is displayed only once when the user U enters the chatbot. This conversation scenario collects basic information about the user U (the most important information for developing a subsequent sales conversation), followed by product suggestions and advice. The second scenario, also known as a push conversation, delivers additional information to the user U that has already accumulated in the chatbot's account. For example, delivery settings are made for each user U along a different time axis, such as a first delivery scenario on the first day after starting use of the conversation application, a second delivery scenario on the second day, and yet another delivery scenario on the fifth day (also known as "track push"). Alternatively, all users linked to a specific personal ID in the conversation application (e.g., registered as friends) are classified into some sort of segment, and a specific segment within that segment is delivered to everyone at the same time (also known as "shot push").

(Step S105) Next, the display control unit 137 causes the operation terminal device 200 to display a scenario confirmation screen that displays the scenario generated by the scenario generation unit 134. Figures 10 to 12 are diagrams showing the scenario confirmation screen P3 displayed on the operation terminal device 200 according to the first embodiment. Scenario confirmation screen (1) P3-1 in Figure 10 shows a conversation that specifies an initial conversation with the user, among a series of conversations included in the scenario. Scenario confirmation screen (2) P3-2 in Figure 11 shows a conversation that specifies a question requesting an answer from the user, among a series of conversations included in the scenario. Scenario confirmation screen (3) P3-3 in Figure 12 shows a conversation that specifies a proposal for a product or the like to the user, among a series of conversations included in the scenario. Scenario confirmation screen (1) P3-1, scenario confirmation screen (2) P3-2, and scenario confirmation screen (3) P3-3 are actually a single screen, and are displayed consecutively by operating the scroll bar provided within the screen. In a scenario specified in this way, the conversation progresses from top to bottom. Note that scenario confirmation screen (1) P3-1, scenario confirmation screen (2) P3-2, and scenario confirmation screen (3) P3-3 may be displayed as different screens that can be switched between.

Scenario confirmation screen (1) P3-1 in Figure 10 shows an example in which actions AC1-1 to AC1-5 are set in order as initial conversations with the user. Action AC1-1 is an action that indicates the settings of the provider's site from which the transition occurs. Action AC1-2 is an action that displays a greeting, and action AC1-3 is an action that displays image 1. Action AC1-4 is an action that displays a question. Action AC1-5 is an action that displays answer options for the question in action AC1-4.

The scenario confirmation screen (2) P3-2 in Figure 11 shows an example in which actions AC2-1 to AC2-5 are set in order as questions requesting an answer from the user. Action AC2-1 is an action that displays the first question. Action AC2-2 is an action that displays answer options for the question of action AC2-1. Action AC2-3 is an action that displays the second question. Action AC2-4 is an action that displays answer options for the question of action AC2-3. Action AC2-5 is an action that displays the third question.

The scenario confirmation screen (3) P3-3 in Figure 12 shows an example in which actions AC3-1 to AC3-2 are set in order as suggestions for products, etc. to the user. Action AC3-1 is an action that displays a message indicating that a product, etc. is being suggested. Action AC3-2 is an action that displays suggested products A to D corresponding to the message of action AC3-1. These suggestions are specified to be variable depending on the user U's response to the question included in the scenario confirmation screen (2) P3-2.

By operating the input interface of the operation terminal device 200, designer D can change the settings of each action displayed on the scenario confirmation screen P3 (correcting wording, adding or deleting options, adding or changing images, etc.). Designer D can also set a new action at any position by pressing the add action button BT2. Designer D can also decide on a scenario by pressing the decide scenario BT3.

In the examples shown in Figures 10 and 11, we have described the case where an action is set to display answer options for a question, but it is also possible to display a free input field that allows the user U to enter free text as a response to the question. Figure 13 is a diagram showing an example of a scenario confirmation screen (4) P3-4 displayed on the operation terminal device 200 according to the first embodiment. Compared to the scenario confirmation screen (2) P3-2 in Figure 11, in the scenario confirmation screen (4) P3-4 in Figure 13, action AC2-6 in the free input field is set as the action that accepts a response to the question in action AC2-1.

(Step S107) Returning to FIG. 8, the scenario generation unit 134 determines whether or not an editing instruction has been received from the operation terminal device 200. For example, if designer D performs some editing operation on the scenario confirmation screen P3 displayed on the operation terminal device 200 and receives an editing instruction from the operation terminal device 200, the scenario generation unit 134 determines that an editing instruction has been received (Step S107: YES). On the other hand, if designer D presses the scenario decision BT3 without performing any editing operation on the scenario confirmation screen P3 displayed on the operation terminal device 200 and receives a scenario decision instruction from the operation terminal device 200, the scenario generation unit 134 determines that an editing instruction has not been received (Step S107: NO).

(Step S111) If it is determined that an editing instruction has been received (Step S107: YES), the scenario generation unit 134 edits the scenario in accordance with the editing instruction. Then, the process returns to step S105, where the display control unit 137 causes the operation terminal device 200 to display a scenario confirmation screen displaying the edited scenario, and the subsequent processing is repeated.

(Step S109) If it is determined that an editing instruction has not been received (Step S107: NO), the scenario generation unit 134 determines (confirms) the scenario and stores it in the memory unit 15. This completes the processing of this flowchart.

<Scenario display process>
Next, the process on the service side where the user U experiences the chatbot service will be described. Fig. 14 is a sequence diagram showing an example of the scenario display process according to the first embodiment. Here, the process will be described taking as an example a situation where the user U operates the user terminal device 100 and browses a provider's website using a web browser.

(Step S201) First, when a trigger operation is performed on the site in response to an operation by user U, the user terminal device 100 transmits a trigger operation signal indicating that a trigger operation has been performed to the service server 3. An example of a trigger operation is pressing a button that starts a chatbot service displayed in a pop-up on the site.

(Step S203) Next, in response to the trigger operation signal received from the user terminal device 100, the service server 3 sends a conversation start instruction to the chatbot server 1 to start a series of conversations in accordance with a scenario in the chatbot service.

(Step S205) Next, when the conversation generation unit 132 of the chatbot server 1 receives a conversation start instruction from the service server 3, it generates a conversation based on the scenario data SD stored in the memory unit 15. Here, for example, the conversation generation unit 132 generates the first conversation (first conversation) set as the initial conversation sentence with the user in the scenario associated with the business operator.

(Step S207) Next, the sending unit 133 of the chatbot server 1 sends a request to send conversation data indicating the generated conversation to the message server 5. The chatbot server 1 sends the request to send conversation data to the message server 5, for example, by using an API provided by the message server 5. Having received the request to send conversation data from the chatbot server 1, the message server 5 sends this conversation data to the user terminal device 100 (Step S209).

(Step S211) Next, the user terminal device 100 launches the conversation application AP based on the conversation data received from the chatbot server 1 via the message server 5, and displays the conversational text based on the conversational data on the display unit 103. This allows the user U to check the conversational text automatically sent from the chatbot server 1.

(Step S213) Next, if the conversation data received from the chatbot server 1 via the message server 5 is a question requesting an answer, the user terminal device 100 transmits answer data indicating the answer determined in response to the user U's operation of the operation unit 105 to the message server 5. Note that if the conversation data received from the chatbot server 1 via the message server 5 is a simple message or image that does not request an answer, the user terminal device 100 may transmit receipt confirmation data indicating receipt to the message server 5. (Step S215) Having received the answer data from the user terminal device 100, the message server 5 transmits this answer data to the chatbot server 1.

(Step S217) Next, the acquisition unit 131 of the chatbot server 1 stores the response data received from the user terminal device 100 via the message server 5 in the response data RD of the storage unit 15. Figure 15 is a diagram showing an example of response data RD according to the first embodiment. The response data RD stores response data acquired from users, linked to a user ID that identifies each user.

If the answer data received from the user terminal device 100 via the message server 5 is entered as free text, the received answer data is stored in the memory unit 15, and the auto-chat unit 135 generates an answer sentence for the user input text by inputting the answer data (user input text) to the second learning model AI2, and the transmission unit 133 transmits the generated answer sentence to the user terminal device 100 via the message server 5. In addition, if the user U cannot find an appropriate answer option in response to a question from the chatbot server 1 for the user U and inputs a different answer as free text, the conversation generation unit 132 stores the content of that answer in the memory unit 15 and proceeds to the next scenario step. In this case, the answer from the user U entered as free text is ultimately used to determine the product to be proposed by the individual offer generation unit 136.

The series of processing QA from steps S205 to S217 above is repeatedly executed according to the scenario. That is, the transmission of conversations and reception of responses from objects set in order in the scenario are repeatedly executed. (Step S219) Finally, when the acquisition unit 131 of the chatbot server 1 receives response data from the user U in response to a conversation proposing a product, etc., the transmission unit 133 transmits purchase request data indicating a desire to purchase to the service server 3.

When the target scenario is set to make an individual offer based on instructions from designer D or the like, the individual offer generation unit 136 acquires user data for user U from user data UD pre-stored in the memory unit 15, and inputs the acquired user data into the third learning model AI3 to determine products, etc. personalized for user U (products, etc. that user U is likely to prefer and purchase). The conversation generation unit 132 then generates a conversation proposing products, etc. personalized for this user U, and the transmission unit 133 transmits conversation data indicating the generated conversation to the user terminal device 100 via the message server 5.

(Step S221) When the service server 3 receives the purchase request data from the chatbot server 1, it generates a purchase screen for performing payment processing, etc. (Step S223) Next, the service server 3 transmits purchase screen data for displaying the generated purchase screen to the user terminal device 100.

(Step S225) Next, the user terminal device 100 displays the purchase screen on the display unit 103 based on the purchase screen data received from the service server 3. (Step S227) Next, the user terminal device 100 transmits the purchase data entered by the user U in response to operation of the operation unit 105 to the service server 3.

(Step S229) Next, the service server 3 performs purchase processing, including payment processing and shipping processing, based on the purchase data received from the user terminal device 100. This completes the processing in this sequence diagram.

According to the first embodiment described above, it is possible to reduce the effort required to prepare a scenario using machine learning techniques. Furthermore, by adopting a configuration in which a scenario is generated using a learning model generated by studying previously generated scenarios, the quality of the generated scenario can be improved. Furthermore, by making it possible to respond to conversational text entered by the user in free text using the auto-chat function, it is possible to improve user convenience. Furthermore, by making it possible to suggest products, etc. based on user data, it is possible to suggest products, etc. that are more suitable for the user. By repeating this series of processes, data related to user information and preferences can be accumulated, and understanding of the user can be deepened.

Second Embodiment
Next, a second embodiment will be described. The second embodiment differs from the first embodiment in that a scenario is composed of multiple objects that divide a series of conversational content according to purpose. The following describes the chatbot server 1 of the second embodiment, focusing on the differences from the first embodiment.

In the second embodiment, a scenario is composed of multiple objects that divide (group) a series of conversational content according to purpose (function, role). For example, a scenario has a structure in which multiple objects are related to each other and arranged in a tree structure. Each of the multiple objects contains multiple conversational sentences in a specified order.

Figure 16 is a diagram showing an example of object-type scenario data SD according to the second embodiment. The object-type scenario data SD includes one or more scenarios SC defined for each business operator. Each scenario SC is composed of multiple objects. The objects include, for example, an entry object OB1, a question object OB2, an offer object OB3, and a retention object OB4. The multiple objects are not limited to these four objects, and may include other objects set according to other functions or purposes. The multiple objects include at least an entry object that specifies the initial conversation text with user U, a question object that specifies a question requesting an answer from user U, and an offer object that specifies a proposal for a product or service to user U.

The entry object OB1 defines the initial conversation text to be used when starting a conversation with a user. The entry object OB1 includes, for example, a greeting message or a question to determine the user's attributes. The question object OB2 defines the conversation text (question text) that requests an answer from the user to determine a product or other item to be proposed to the user. The offer object OB3 defines the content of the conversation text (proposal text) that proposes a product or other item to the user based on the user's answer collected by the question object OB2. The retention object OB4 defines the conversation text to bring a user who did not end up purchasing a product or other item back into the conversation. The retention object OB4 includes, for example, conversation text for push notifications set for each user.

Figure 17 is a diagram showing an example of an object structure according to the second embodiment. The object structure specifies the types of objects included in each scenario, the number of objects, how the objects are connected to each other, and so on. The example shown in Figure 17 shows an object structure including an entry object OB1, a question object (male) OB2-1 and a question object (female) OB2-2 each connected to the entry object OB1, an offer object (male) OB3-1 connected to the question object (male) OB2-1, and an offer object (female) OB3-2 connected to the question object (female) OB2-2. In a scenario, a conversation progresses from top to bottom through this object structure.

The scenario generation unit 134 generates an object-type scenario using the first learning model AI1 pre-stored in the storage unit 15 based on instruction information related to scenario generation acquired from the operation terminal device 200, and stores the generated object-type scenario as object-type scenario data SD in the storage unit 15. The first learning model AI1 is a large-scale language model (LLM) generated by learning using large amounts of data as learning data, such as past object-type scenario data generated in the past and conversation history data with past users (past chat data). In other words, the scenario generation unit 134 generates a scenario composed of multiple objects that divide a series of conversation contents according to purpose.

According to the second embodiment described above, the load of scenario generation can be reduced using machine learning techniques. Furthermore, by adopting a configuration in which scenarios are generated using a learning model generated by learning previously generated scenarios, the quality of the generated scenarios can be improved. Furthermore, by using object-type scenario data to enable learning and generation of dialogue sentences for each object, the accuracy of the scenario can be improved.

The above describes the form for carrying out the present invention using an embodiment, but the present invention is in no way limited to such an embodiment, and various modifications and substitutions can be made without departing from the spirit of the present invention.

1... Chatbot server, 11... Communication unit, 13... Control unit, 15... Memory unit, 131... Acquisition unit, 132... Conversation generation unit, 133... Transmission unit, 134... Scenario generation unit, 135... Auto chat unit, 136... Individual offer generation unit, 137... Display control unit, 138... Learning unit, 100... User terminal device, 101... Communication unit, 103... Display unit, 105... Operation unit, 107... Control unit, 109... Memory unit, 200... Operation terminal device, 3... Service server, 5... Message server, NW... Communication network

Claims (17)

A scenario generation device that generates a scenario that defines a series of conversation contents to be executed by a chatbot to provide a user with information about a product or service,
a display control unit that displays on a display unit a scenario condition setting screen that accepts specification of specific information that identifies a product or service specified by a scenario designer;
a scenario generation unit that generates a scenario corresponding to the specific information by inputting the specific information specified on the scenario condition setting screen into a first learning model that is provided in the scenario generation device and that is prepared by learning a scenario that has been generated in the past;
Scenario generation device. the scenario generation unit further inputs, into the first learning model, purpose information indicating a purpose of providing information about a product or service specified by the designer, thereby generating the scenario corresponding to the purpose information.
The scenario generation device according to claim 1 . The purpose indicated by the purpose information includes guiding the user to purchase a product or service.
The scenario generation device according to claim 2 . the scenario generation unit further inputs an image relating to a product or service designated by the designer into the first learning model, thereby generating the scenario in which the image is incorporated into the scenario.
The scenario generation device according to claim 1 . a display control unit that displays a confirmation screen of the scenario generated by the scenario generation unit on a display unit,
The scenario generation device according to any one of claims 1 to 4. the scenario generation unit edits the generated scenario in response to an editing instruction from the designer on the scenario confirmation screen displayed on the display unit.
The scenario generation device according to claim 5 . the scenario generation unit generates the scenario composed of a plurality of objects in which the series of conversation contents are divided according to purpose.
The scenario generation device according to any one of claims 1 to 4. The plurality of objects include at least an entry object that defines an initial conversation with the user, a question object that defines a question to which the user is requested to respond, and an offer object that defines a proposal for the product or service to the user.
The scenario generation device according to claim 7. The scenario generation device further includes an auto-chat unit that, when the generated scenario is presented to the user, inputs a free text conversational sentence designated by the user into a second learning model that is provided in the scenario generation device and that is trained to output a response sentence corresponding to the input text data using learning data including past conversation history data with the user and past scenario data, and generates a response sentence to the conversational sentence.
The scenario generation device according to any one of claims 1 to 4. The system further includes an auto-chat unit that, when the generated scenario is presented to the user, inputs a conversational sentence in a free text format designated by the user into a second learning model that is provided in a device accessible via a network and that is trained to output a response sentence corresponding to the input text data using learning data including conversation history data with the user and past scenario data, and generates a response sentence to the conversational sentence.
The scenario generation device according to any one of claims 1 to 4. The scenario generation device further includes an individual offer generation unit that generates a proposal for a product or service suited to a user by inputting user data of the user who presented the generated scenario into a third learning model that is provided in the scenario generation device and that is trained to output information on a product or service suited to a specific user when user data of the specific user is input, using learning data that is a set of user data of a user who has previously purchased a product or service and purchase history data of the user.
The scenario generation device according to any one of claims 1 to 4. The system further comprises an individual offer generation unit that generates a proposal for a product or service tailored to a user by inputting the user data of the user who presented the generated scenario into a third learning model that is provided in a device accessible via a network and that is trained to output information about a product or service tailored to a specific user when the user data of the specific user is input, using learning data that is a set of user data of a user who has previously purchased a product or service and purchase history data of the user.
The scenario generation device according to any one of claims 1 to 4. A scenario generation device that generates a scenario that defines a series of conversation contents to be executed by a chatbot to provide a user with information about a product or service,
a display control unit that displays on a display unit a scenario condition setting screen that accepts specification of specific information that identifies a product or service specified by a scenario designer;
a scenario generation unit that generates a scenario corresponding to the specific information by inputting the specific information specified on the scenario condition setting screen into a first learning model that is provided in a device accessible via a network and that is prepared by learning a scenario that has been generated in the past;
Scenario generation device. A scenario generation method for generating a scenario that defines a series of conversation contents to be executed by a chatbot to provide a user with information about a product or service, comprising:
The computer of the scenario generation device
displaying a scenario condition setting screen on the display unit for accepting specification of specific information for identifying a product or service specified by a scenario designer;
generating a scenario corresponding to the specific information by inputting the specific information specified on the scenario condition setting screen into a first learning model that is provided in the scenario generation device and that is prepared by learning a scenario that has been generated in the past;
Scenario generation method. A scenario generation method for generating a scenario that defines a series of conversation contents to be executed by a chatbot to provide a user with information about a product or service, comprising:
The computer of the scenario generation device
displaying a scenario condition setting screen on the display unit for accepting specification of specific information for identifying a product or service specified by a scenario designer;
generating a scenario corresponding to the specific information by inputting the specific information specified on the scenario condition setting screen into a first learning model that is provided in a device accessible via a network and that is prepared by learning a previously generated scenario;
Scenario generation method. A scenario generation program that generates a scenario that defines a series of conversation contents to be executed by a chatbot to provide a user with information about a product or service,
The computer of the scenario generation device
displaying a scenario condition setting screen on the display unit for accepting specification of specific information for identifying a product or service specified by a scenario designer;
generating a scenario corresponding to the specific information by inputting the specific information specified on the scenario condition setting screen into a first learning model that is provided in the scenario generation device and that is prepared by learning a scenario that has been generated in the past;
Scenario generator. A scenario generation program that generates a scenario that defines a series of conversation contents to be executed by a chatbot to provide a user with information about a product or service,
The computer of the scenario generation device
displaying a scenario condition setting screen on the display unit for accepting specification of specific information for identifying a product or service specified by a scenario designer;
a first learning model that is provided in a device accessible via a network and that is prepared by learning a previously generated scenario, and that generates a scenario corresponding to the specific information by inputting the specific information that has been specified and accepted on the scenario condition setting screen;
Scenario generator.