Support device, support method, and support program

Abstract

To improve a technique for automatically determining whether to make an automatic response or a response by an operator.SOLUTION: A support device includes a reception part for receiving a request related to asset management from a support target person, and a response control part for determining whether to make an operator respond to the request or to make an automatic response according to the content of the request. In addition, according to the support device, it is possible to support decision making of a subject who manages assets.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an assistance device, an assistance method, and an assistance program. [Background technology]

[0002] Asset management requires not only basic knowledge about asset management, but also individual circumstances such as the manager's life stage and investment funds, and decisions on where to invest must be made taking into consideration these. For this reason, it is not easy for many people to manage their assets on their own, and many people consult with asset management experts such as banks or financial planners when starting asset management. Alternatively, an easier way to get started is to use a service that provides information on asset management via an automated response system.

[0003] However, automatic responses have drawbacks such as difficulty in responding to unexpected questions and difficulty in responding according to the person making the inquiry, so it is desirable to be able to switch from automatic responses to responses by an operator as needed. An example of prior art related to switching from automatic responses to responses by an operator is the multi-cloud chat service providing device (hereinafter abbreviated as "providing device") described in Patent Document 1 below.

[0004] The provision device described in Patent Document 1 can send a question message received from a user to an AI (Artificial Intelligence) chat cloud service and generate a reply message. Furthermore, this provision device sends the question message from the user to an operator terminal and has the operator input the reply, on the condition that the user has performed an operation to switch to operator support or the AI ​​has not been able to provide a reply. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2019-128737 Summary of the Invention [Problem to be solved by the invention]

[0006] The method for automatically determining whether to use an automated response or a human operator in the provision device described in Patent Document 1 leaves room for improvement. For example, in this provision device, if an AI is able to provide a response, the system does not switch to human operation regardless of the content of the response. This makes it inevitable that the AI ​​will provide an inappropriate response. Furthermore, even if a user's question could potentially lead to a large-value contract, such as a question about the management of retirement benefits, if the AI ​​is able to provide a response, the system does not switch to human operation. This could result in missed business opportunities due to the AI ​​providing a uniform response.

[0007] The present disclosure has been made in view of such problems, and an exemplary purpose thereof is to improve the technology for automatically determining whether to provide an automatic response or a response by an operator. [Means for solving the problem]

[0008] An exemplary aspect of the present disclosure is a support device that supports asset management, and includes a reception means for receiving requests from persons to be supported, and a response control means for determining, depending on the content of the request, whether to have an operator respond to the request or to have an automatic response.

[0009] An exemplary aspect of the present disclosure is a support method for supporting asset management, in which at least one processor executes a reception process for receiving a request for asset management from a person to be supported, and a response control process for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

[0010] An exemplary aspect of the present disclosure is a support program that causes a computer to function as a support device that supports asset management, and causes the computer to function as a reception means that receives requests regarding asset management from persons to be supported, and a response control means that determines whether to have an operator respond to the request or to respond automatically, depending on the content of the request. [Effects of the Invention]

[0011] According to an exemplary aspect of the present disclosure, an exemplary effect is achieved in that it is possible to improve the technology for automatically determining whether to provide an automatic response or a response by an operator. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram illustrating a configuration of an assistance device according to the present disclosure. [Figure 2] FIG. 1 is a flow diagram showing the flow of a support method according to the present disclosure. [Figure 3] FIG. 10 is a block diagram showing the configuration of another support device according to the present disclosure. [Figure 4] 4 is a diagram showing an example of assistance provided by the assistance device shown in FIG. 3. FIG. [Figure 5] 4 is a flowchart showing the flow of processing executed by the support device shown in FIG. 3. [Figure 6] FIG. 1 is a block diagram illustrating a configuration of a computer that functions as a support device according to the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0013] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining the technologies (part or all of the products or methods) employed in the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technologies employed in the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, the effects mentioned in the exemplary embodiments shown below are examples of effects expected in the exemplary embodiments, and do not define the scope of the present invention. In other words, embodiments that do not exhibit the effects mentioned in the exemplary embodiments shown below may also be included in the scope of the present invention.

[0014] First Exemplary Embodiment A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is the basic form of each exemplary embodiment described later. The scope of application of each technology employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise. Furthermore, each technology shown in the drawings referenced to explain this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise.

[0015] (Configuration of support device 1) The configuration of the support device 1 according to this exemplary embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the configuration of the support device 1. The support device 1 is a device that supports asset management, and as shown in Fig. 1, includes a reception unit 101 and a response control unit 102. The support device 1 also has the function of executing various information processes described below, and can also be referred to as an information processing device or a computer.

[0016] The reception unit 101 receives a request for asset management from a target of support. The "target" is a user of the support device 1, and is typically a person who is considering asset management. The "request" refers to the items that the target inputs into the support device 1 in order to receive support for asset management. For example, a question about asset management, a request for information on asset management, etc. are typical examples of the "request." In addition to these, questions and confirmation messages about items indirectly related to asset management (e.g., stock market trends, etc.) also fall within the category of the "request."

[0017] The data format of the "request" is arbitrary. For example, when asset management support is provided in the form of text chat, the reception unit 101 may acquire the "request" input as text data. When asset management support is provided via voice communication, the reception unit 101 may acquire the "request" input as voice data. In this case, the reception unit 101 may convert the acquired voice data into text data by voice recognition.

[0018] The "request" may include data in multiple formats, such as image (still image or video) data and text data. In this case, the receiving unit 101 may convert non-text elements, such as images, into text format. For example, the receiving unit 101 can perform such conversion by using a generative model trained by machine learning to generate text indicating the content of an image. Note that such conversion and voice recognition may be performed by a device other than the assistance device 1.

[0019] The response control unit 102 determines whether to have an operator respond to the request or to use an automatic response, depending on the content of the request received by the receiving unit 101. The format of the response by the operator is arbitrary. For example, the response by the operator may be in the form of a text chat or a voice call. Furthermore, the voice call may be a video call accompanied by video, or a response may be a combination of a voice call and a text chat. The format of the response is also arbitrary for an automatic response. Furthermore, the algorithm for the automatic response is also arbitrary. For example, the automatic response may be performed using a rule base, or a language model.

[0020] Here, the language model is a model that has been machine-learned to learn the arrangement of components (such as words) in a sentence or the arrangement of sentences in a document. For example, a Generative Pre-Trained Transformer (GPT) can be used as a language model for automatic responses. The GPT predicts a character string that is likely to follow an input character string, thereby outputting a sentence containing the input character string. In addition to the above, other models that can be used to classify subjects include a Text-to-Text Transfer Transformer (T5), a Bidirectional Encoder Representations from Transformers (BERT), a Robustly optimized BERT approach (RoBERTa), and an Efficiently Learning an Encoder that Classifies Token Replacements Accurately (ELECTRA).

[0021] After the response control unit 102 determines whether to have an operator respond to the request of the subject or to have an automatic response, the device that responds to the subject may be the assistance device 1 or another device. For example, if the response control unit 102 determines to have an operator respond, the response to the subject may be made via a terminal device used by the operator. Furthermore, if the response control unit 102 determines to have an automatic response, the response control unit 102 may transfer the request of the subject to an external information processing device equipped with an automatic response function and have the information processing device respond to the request.

[0022] As described above, the support device 1 is a support device that supports asset management, and includes a reception unit 101 that receives requests regarding asset management from the recipient of support, and a response control unit 102 that determines whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

[0023] According to the above configuration, it is determined whether to have an operator respond to the request or to have an automatic response depending on the content of the received request. Therefore, it is possible to provide an appropriate response depending on the content of the received request. In this way, the support device 1 has the effect of making it possible to improve the technology for automatically determining whether to have an automatic response or a response by an operator. Furthermore, the support device 1 can support the subject in making decisions regarding asset management by responding appropriately to the subject's request.

[0024] (Support Program) The functions of the support device 1 described above can also be realized by a program. The support program according to this exemplary embodiment is a support program that causes a computer to function as a support device that supports asset management, and causes the computer to function as a reception unit that receives requests regarding asset management from support recipients, and a response control unit that determines, depending on the content of the request, whether to have an operator respond to the request or to have an automatic response. This support program has the effect of enabling an improvement in the technology for automatically determining whether to have an automatic response or an operator respond.

[0025] (Flow of support methods) The flow of the support method according to this exemplary embodiment will be described with reference to Fig. 2. Fig. 2 is a flow diagram showing the flow of the support method. Note that the execution entity of each step in this support method may be a processor provided in the support device 1, or a processor provided in another device, or each step may be executed by a processor provided in a different device.

[0026] In S1 (reception process), at least one processor receives a request for asset management from a target of support.

[0027] In S2 (response control process), at least one processor determines whether to have an operator respond to the request or to have an automatic response, depending on the content of the request accepted in S1.

[0028] Note that, before the processing of S2, it may be determined whether to have an operator respond to the received request or to use an automatic response. In this case, S2 may determine whether to switch the response mode for the request received in S1. In other words, if it was determined before the processing of S2 that an operator will respond to the received request, S2 may determine whether to switch the response to the request received in S1 to an automatic response. On the other hand, if it was determined before the processing of S2 that an automatic response will be used for the received request, S2 may determine whether to switch the response to the request received in S1 to an operator response. Even when making such a determination, it is essentially "determining whether to have an operator respond to the request or to use an automatic response."

[0029] As described above, the support method according to this exemplary embodiment is a support method for supporting asset management, in which at least one processor executes a reception process for receiving a request regarding asset management from a support recipient, and a response control process for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request. Therefore, the support method according to this embodiment has the effect of enabling an improvement in the technology for automatically determining whether to have an automatic response or to have an operator respond.

[0030] Second Exemplary Embodiment A second exemplary embodiment, which is one example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same functions as those described in the above exemplary embodiment will be assigned the same reference numerals, and their description will be omitted as appropriate. The scope of application of each technology employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical hindrance occurs. Furthermore, each technology shown in each drawing referenced to explain this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical hindrance occurs.

[0031] (Configuration of support device 1A) The configuration of the support device 1A according to this exemplary embodiment will be described with reference to FIG. 3. FIG. 3 is a block diagram showing the configuration of the support device 1A. The support device 1A is a device having a function of supporting asset management. Note that the support device 1A may be a device whose main function is to support asset management, or may be a general-purpose device having other functions as well. Furthermore, the support device 1A may be a stationary device or a portable device.

[0032] 3, the support device 1A includes a control unit 10A that controls the various units of the support device 1A and a storage unit 11A that stores various data used by the support device 1A. The support device 1A also includes a communication unit 12A that enables the support device 1A to communicate with other devices, an input unit 13A that accepts input to the support device 1A, and an output unit 14A that enables the support device 1A to output data. The control unit 10A includes a reception unit 101A, a response control unit 102A, a related information acquisition unit 103A, an index value calculation unit 104A, an automatic response unit 105A, and a response evaluation unit 106A. The index value calculation unit 104A and the response evaluation unit 106A will be described in detail below.

[0033] The reception unit 101A receives a request for asset management from a support recipient, similar to the reception unit 101 described in the first exemplary embodiment. In this exemplary embodiment, an example of providing asset management support via voice communication will be described. For this purpose, the reception unit 101A receives input of voice data indicating the content of the support recipient's request for asset management, and performs voice recognition on the voice data to convert it into a text format. Note that the processing when providing asset management support via text exchanges such as text chat is similar to the processing for support via voice communication, except that the request from the support recipient is received in text and the response to the support recipient is displayed as text.

[0034] Similar to the response control unit 102 described in the first exemplary embodiment, the response control unit 102A determines whether to have an operator respond to a request or to use an automatic response depending on the content of the request received by the receiving unit 101A. As described above, in this exemplary embodiment, an example will be described in which asset management support is provided via voice communication. Therefore, when the response control unit 102A determines that an operator will respond, it transfers the incoming call from the target person to the operator. On the other hand, when the response control unit 102A determines that an automatic response will be provided, it causes the automatic response unit 105A to respond. The response control unit 102A may use related information described below when determining whether to have an operator respond or to use an automatic response.

[0035] The related information acquisition unit 103A acquires related information used when determining whether to have an operator respond or to have an automatic response. The related information may be any information that is useful in determining whether to have an operator respond or to have an automatic response. For example, the related information acquisition unit 103A may acquire related information that indicates the attributes of the subject. Examples of the attributes of the subject include personal characteristics (age, gender, occupation, assets held, past asset management history, etc.) that the subject has registered in advance.

[0036] Furthermore, for example, the related information acquiring unit 103A may acquire, as related information, images (which may be still images or moving images) of the subject, audio data of the subject's speech, and measurement data of various sensors (e.g., acceleration sensors, temperature sensors, heart rate sensors, sweat sensors, etc.) worn by the subject. In this way, the related information acquiring unit 103A may acquire data of multiple modalities (data formats) as related information.

[0037] The related information acquisition unit 103A can detect the subject's current emotions and changes in emotions by analyzing these data related to the subject's emotions. The response control unit 102A can then use the analysis results to determine whether to have an operator respond or to use an automatic response, taking into account the subject's emotions and changes in those emotions. For example, the response control unit 102A can switch to an operator response if the subject's emotions worsen immediately after presenting an automatic response. The data analysis may be performed by the related information acquisition unit 103A, or the results of analysis by another device may be acquired as related information.

[0038] The automatic response unit 105A realizes an automatic response to a request from the subject. As explained in the first exemplary embodiment, the algorithm for the automatic response is arbitrary. In this exemplary embodiment, an example will be described in which the automatic response unit 105A performs an automatic response using a language model. In this case, the automatic response unit 105A inputs text indicating the content of the subject's request into the language model and generates a response to the request. The automatic response unit 105A then presents the generated response to the subject. In this exemplary embodiment, since asset management support is provided via voice communication, the automatic response unit 105A converts the generated response into voice data and transmits this voice data to the subject's terminal device via a communication line. Note that the conversion to voice data may be performed by a device other than the support device 1A.

[0039] As described above, the support device 1A includes the reception unit 101A that receives a request for asset management from a support recipient, and the response control unit 102A that determines whether to have an operator respond to the request or to have an automatic response depending on the content of the request received by the reception unit 101A. Therefore, similar to the support device 1 of the exemplary embodiment 1, it is possible to obtain an effect that it is possible to improve the technology for automatically determining whether to have an automatic response or an operator respond.

[0040] (Examples of support) Fig. 4 is a diagram showing examples of assistance provided by the assistance device 1A. More specifically, Fig. 4 shows two examples, Ex1 and Ex2. In Ex1 and Ex2, the recipients of assistance are different, U1 and U2, and the requests made by recipients U1 and U2 to the assistance device 1A are different.

[0041] In Ex1, the target person U1 is on a call with the support device 1A and makes an utterance Q1 requesting advice on what to do first to start asset management. The reception unit 101A of the support device 1A receives this utterance Q1 as a request for asset management from the target person U1. The response control unit 102A then determines whether to have an operator respond to the request or to have an automatic response, depending on the content of the request.

[0042] Here, the content of utterance Q1 in Ex1 is a request for general information required to begin asset management, and it is considered highly likely that automatic response unit 105A will be able to provide a response that satisfies target U1. For this reason, in Ex1, response control unit 102A decides to use an automatic response, and answer A1 generated by automatic response unit 105A is presented to target U1.

[0043] A specific method for the response control unit 102A to make such a determination will be described later. Also, in Fig. 4, the answer A1 is shown in a speech bubble associated with the support device 1A, but in reality, the answer A1 is output from the audio output unit of the receiver held by the subject U1. The same applies to the answer A2 in Ex2 described below.

[0044] In Ex2, the target person U2 is on the phone with the support device 1A and makes an utterance Q2 requesting consultation about the management of his / her retirement benefits. The reception unit 101A of the support device 1A receives this utterance Q2 as a request from the target person U2 regarding asset management. The response control unit 102A then determines whether to have an operator respond to the request or to have an automatic response, depending on the content of the request.

[0045] The content of utterance Q2 in Ex2 is that the person wants to consult about the management of his / her retirement benefits, and since a relatively large profit is expected if the management of the retirement benefits can be entrusted, it is considered desirable for the person to be answered by an operator. Therefore, in Ex2, the response control unit 102A decides to have the person U2 respond by an operator, and presents answer A2 to the person U2 indicating that the call will be transferred to the operator. Although not shown in the figure, after this, the person U2 and the operator are in a conversation, and the person U2 can consult with the operator about the management of the retirement benefits.

[0046] (Example 1 of Determination by Response Control Unit 102A: Determination Based on Index Value) 3, the support device 1A includes an index value calculation unit 104A. The index value calculation unit 104A calculates an index value for determining whether or not to switch to a response by an operator, based on the content of the request of the subject during the period in which the automatic response is applied. The response control unit 102A then determines whether or not to switch the response to the subject to a response by an operator, based on the index value calculated by the index value calculation unit 104A. This provides the effect of being able to switch from an automatic response to a response by an operator at an appropriate time, in addition to the effect provided by the support device 1.

[0047] Furthermore, when a new request is received from the subject, the index value calculation unit 104A may update the most recently calculated index value based on the content of the new request. Alternatively, when a new request is received from the subject, the index value calculation unit 104A may recalculate the index value based on the content of the new request and the content of the series of requests received up to the time the new request was received. These configurations, in addition to the effects of the support device 1, provide the effect of being able to switch from automatic response to response by an operator at an appropriate timing, taking into account multiple requests received during the period in which automatic response is applied.

[0048] The method for calculating the index value is not particularly limited. For example, the index value calculation unit 104A may calculate the index value using an index value calculation model generated by machine learning the relationship between a request from a target of asset management support and an index value indicating the need for an operator to respond to the request. This provides, in addition to the effects of the support device 1, an effect of being able to switch from automatic response to response by an operator at an appropriate timing based on the results of machine learning.

[0049] Alternatively, a language model may be used instead of the index calculation model. In this case, the index calculation unit 104A generates a prompt that includes a request from the asset management support recipient and instructs the output of an index indicating the need for an operator to respond to the request, and inputs the generated prompt into the language model. This causes the index to be output from the language model. When using a language model, a prompt may be used that asks whether the response should be made by an operator or an automatic response. In this case, the response control unit 102A determines whether the response should be made by an operator or an automatic response according to the output of the language model in response to the prompt. It is not necessary to input the entire request of the recipient into the language model; a summary of the request or words extracted from the request may be input into the language model. This also applies when the index calculation model described above is used.

[0050] Alternatively, for example, multiple keywords or phrases that determine whether to have an operator respond or an automated response may be preset, and an evaluation value may be preset for each keyword or phrase. For example, keywords such as "operator" and "person in charge" may be set with an evaluation value indicating that an operator should respond. Furthermore, negative keywords or phrases that describe the current response, such as "difficult to understand" or "not that," may also be set with an evaluation value indicating that an operator should respond. By setting keywords or phrases and their evaluation values ​​in this manner, the index value calculation unit 104A can detect the above keywords from the target user's request and calculate the index value of the request by adding up the evaluation values ​​set for the detected keywords.

[0051] Furthermore, in asset management support by the support device 1A, an operator may provide asset management advice based on information collected during an automatic response (e.g., a period during which the automatic response unit 105A and the subject are allowed to freely converse). In this case, the index value calculation unit 104A may calculate an index value indicating the collection status of information necessary for the operator's advice. In this case, the response control unit 102A may switch to an operator's response when the calculated index value reaches or exceeds a predetermined threshold. For example, the information necessary for advice may be listed in advance, and the index value may be calculated by comparing the information collected during the automatic response with the list.

[0052] Furthermore, when considering asset management, it is common to first decide on a portfolio and then consider specific investment destinations, such as selecting financial products, in line with that portfolio. In asset management, the term "portfolio" refers to asset allocation. For example, suppose you have assets of 10 million yen and invest them equally in domestic stocks, domestic bonds, foreign stocks, and foreign bonds. In this case, the portfolio would allocate 25% of assets each to domestic stocks, domestic bonds, foreign stocks, and foreign bonds. The allocation destinations are optional, and could include financial assets such as deposits and insurance, as well as physical assets such as real estate and precious metals.

[0053] For this reason, the automatic response unit 105A may first present the subject with questions to elicit information necessary for determining a portfolio, and obtain the subject's answers to those questions. Then, the response control unit 102A may switch to responses by an operator once collection of the information necessary for determining a portfolio has been completed. This allows the operator to create a portfolio based on the collected information and consult with the subject about asset management based on that portfolio.

[0054] The support device 1A may also generate a portfolio. For example, by predetermining rules for generating a portfolio according to the answers of the subject, it is possible to automatically generate a portfolio according to the rules. Also, for example, it is possible to input the answers of the subject into a language model and generate a portfolio that suits the subject as understood from the answers. Furthermore, instead of generating a portfolio, it is also possible to prepare multiple patterns of portfolios in advance and select and present one of the portfolios that suits the subject based on the answers of the subject.

[0055] (Example 2 of Determination by Response Control Unit 102A: Determination Based on Evaluation Results of Expected Profit) Furthermore, the response control unit 102A may determine whether to have an operator respond to the subject or to have an automatic response based on at least one of the evaluation results of the benefit expected to be obtained by having an operator respond to the subject's request and the evaluation results of the benefit expected to be obtained by having an automatic response to the subject's request. This provides an effect that, in addition to the effect achieved by the support device 1, it is possible to increase the benefit obtained by responding to the subject.

[0056] The method for evaluating the expected profit is arbitrary. For example, if the request from the target person includes the amount of money the target person intends to invest or content suggesting the amount of money to be invested, the response control unit 102A may evaluate the expected profit from such content. To give a specific example, if the request from the target person includes keywords such as "retirement benefits" or "inheritance," the response control unit 102A may evaluate that the expected profit from having an operator respond is large and decide to have an operator respond (see Ex2 in FIG. 4). On the other hand, if the request from the target person includes the keyword "inexperienced," the response control unit 102A may evaluate that the expected profit from having an operator respond is small and decide to use an automatic response (see Ex1 in FIG. 4).

[0057] Furthermore, for example, the response control unit 102A may predict the expected profit using a prediction model (machine learning of the relationship between the request and the expected profit) that predicts the expected profit, or a language model. The evaluation of the expected profit may be performed separately from the calculation of the index value described above, or an index value including the expected profit may be calculated. In the former case, the response control unit 102A may determine whether to have an operator or an automatic response respond to the target person based on both the evaluation result of the expected profit and the index value. In the latter case, the index value may be calculated taking into account factors related to the expected profit (for example, keywords such as "retirement benefits" included in the request, and the target person's age, income, etc. identified from the target person's related information).

[0058] (Another Example of Determination by Response Control Unit 102A) In some countries and regions, there are regulations that prohibit the act of recommending financial products using an automated response system. Therefore, when the content of a request from a target person is a request for a proposal of a financial product to invest in, the response control unit 102A may decide to have an operator respond to the request. This provides the effect of using the automated response system while switching to an operator response as needed, thereby enabling compliance with the above-mentioned regulations.

[0059] The method for determining whether the content of the target person's request is a request for a proposal of a financial product to invest in is arbitrary. For example, if the target person's request contains a predetermined keyword or phrase such as the name of a financial product, the response control unit 102A may determine that the request may be a request for a proposal of a financial product and have an operator respond to the request. Alternatively, for example, the response control unit 102A may input the target person's request into a language model to determine whether the request is a request for a proposal of a financial product.

[0060] Furthermore, in the provision of asset management services, there are actions that are permitted only to those with certain qualifications. Therefore, when the content of a request from a target person indicates that a response from a qualified person with certain qualifications is required for the request, the response control unit 102A preferably determines to have a qualified operator respond to the request. This provides the effect of not only utilizing the effect of the support device 1, but also enabling a qualified person to respond to requests that require a response from a qualified person, while utilizing an automatic response.

[0061] The method for determining whether a response from a qualified person is required may also be arbitrary. For example, the response control unit 102A may determine whether a response from a qualified person is required based on whether a predetermined keyword or phrase that is often included in requests that require a response from a qualified person is included in the request of the target person. Alternatively, for example, the response control unit 102A may input the request of the target person into a language model to determine whether a response to the request from a qualified person is required.

[0062] In addition, there may be a case where there are multiple operators, and the multiple operators may include both qualified and unqualified ones. In such a case, the response control unit 102A causes an operator among the multiple operators who has the necessary qualifications to respond to the request of the target person to respond to the request. Note that if there is no operator with the necessary qualifications to respond to the target person's request, or if such an operator is unable to handle the target person, the response control unit 102A may present the target person with a message indicating that it is difficult to respond.

[0063] Furthermore, when there are multiple operators, it is preferable that the response control unit 102A have an operator who has a good compatibility with the target respond to the target's request. The method of estimating the compatibility between the target and the operator is arbitrary. For example, it is also possible to estimate an operator who has a good compatibility with the target from the content of the target's request or related information about the target.

[0064] Furthermore, multiple language models for automatic responses may be prepared. In this case, the response control unit 102A can respond to the request of the subject by applying, from among the multiple language models, a language model that is compatible with the subject or a language model that is suited to the content of the subject's request. For example, a language model A for consultations on life planning (e.g., a general-purpose language model fine-tuned for consultation cases on life planning) and a language model B for consultations on financial products (e.g., a general-purpose language model fine-tuned for consultation cases on financial products) may be prepared in advance. In this case, if the subject's request is related to life planning, the response control unit 102A causes the automatic response unit 105A to automatically respond using the language model A. On the other hand, if the subject's request is related to consultations on financial products, the response control unit 102A causes the automatic response unit 105A to automatically respond using the language model B. Note that a similar response can also be achieved by modifying the prompt. For example, it is possible to generate responses suitable for life planning by including in the prompt a phrase such as "Please answer as a life planning expert" or by including pairs of requests and model answers in life planning as answer examples.

[0065] (Response Evaluation) Since the operator is a human being, it is difficult for the operator to always respond appropriately to the subject, and there is a possibility that the operator may sometimes respond inappropriately. From this perspective, the support device 1A is provided with a response evaluation unit 106A (see FIG. 3).

[0066] The response evaluation unit 106A evaluates the content of the response by the operator. The support device 1A including the response evaluation unit 106A has the effect of being able to improve the quality of the response by the operator, in addition to the effect achieved by the support device 1. For example, the evaluation result by the response evaluation unit 106A may be notified to the operator himself / herself and / or a manager such as his / her superior, to encourage improvement in the quality of the response.

[0067] Furthermore, for example, the response control unit 102A may determine whether or not an operator needs to be changed while the operator is responding to the target person, based on the evaluation result of the operator by the response evaluation unit 106A. This makes it possible to prevent the target person from feeling bad about the operator, who is unable to respond appropriately to the target person, from continuing to respond.

[0068] The evaluation criteria for the response content may be determined in advance. Furthermore, any method according to the evaluation criteria may be applied as the method for evaluating the response content. For example, the evaluation criteria may be based on whether or not there is a compliance problem in the operator's response content. In this case, for example, the response evaluation unit 106A may summarize the operator's response content (and the request content of the subject) using a language model and determine whether or not there is a compliance problem in the summarized response content. The presence or absence of a problem may also be determined using a language model. Furthermore, the response evaluation unit 106A may determine whether or not there is a problem by presenting a summary to a predetermined judge who determines whether or not there is a compliance problem and having the judge input the judge's judgment result into the assistance device 1A. In this case, the judgment result input by the judge becomes the evaluation result of the response content by the response evaluation unit 106A.

[0069] The response control unit 102A can also evaluate the content of the response of the automatic response in a similar manner. However, in order to evaluate the content of the response of the automatic response, it is necessary to record not only the content of the response of the automatic response during the period in which the automatic response is applied, but also the requests of the target person during the same period. Furthermore, recording in this manner may hinder the target person's free speech, which is based on the assumption that the response is automatic (i.e., that it will not be heard by others). For this reason, the response control unit 102A may record and evaluate the content of the response by the operator (and the content of the request of the target person), but not record or evaluate the content of the response of the automatic response (and the content of the request of the target person).

[0070] (Processing flow) The flow of processing executed by the support device 1A will be described with reference to Fig. 5. Fig. 5 is a flow diagram showing an example of processing executed by the support device 1A. Fig. 5 includes each process of the support method according to this exemplary embodiment. Fig. 5 also shows an example in which an automatic response is used as the basic response method, and is switched to an operator's response as needed.

[0071] In S11, the related information acquisition unit 103A acquires related information that is used when determining whether to have an operator respond or to have an automatic response. The related information can be acquired by any method. For example, the related information acquisition unit 103A may acquire related information input by the person receiving support via the communication unit 12A or the input unit 13A, or may acquire related information from a predetermined source (for example, a database that pre-records related information about the person receiving support).

[0072] In S12 (reception process), the reception unit 101A receives a request for asset management from a support recipient. When providing support for asset management via voice communication, the reception unit 101A receives a request from the support recipient (specifically, input of voice data indicating the content of the request) via a call with the support recipient. The reception unit 101A then performs voice recognition on the voice data and converts it into text format. As described above, automatic response is the basic response method here, and therefore, automatic response is applied at the time the process of S12 is performed.

[0073] In S13, the index value calculation unit 104A calculates an index value for determining whether to switch to an operator response based on the content of the target person's request accepted in S12. As described above, the response control unit 102A can also determine whether to have an operator respond or to use an automatic response without using an index value. For example, the response control unit 102A can also determine whether to have an operator respond or to use an automatic response based on the evaluation result of expected profit. In this case, the evaluation result of expected profit is performed in S13. The response control unit 102A may also determine whether to have an operator respond or to use an automatic response based on whether the target person's request contains a predetermined keyword or phrase, in which case the keyword or phrase is extracted in S13.

[0074] In S14 (response control processing), the response control unit 102A determines whether to switch the response to the target person to a response by an operator based on the index value calculated in S13. In other words, the response control unit 102A determines whether to have an operator respond to the request or to have an automatic response, depending on the content of the request accepted in S12. Note that in S14, the response control unit 102A may also make the determination using the related information acquired in S11. If the determination in S13 is YES (to have an operator respond), the process proceeds to S15. On the other hand, if the determination in S13 is NO (to maintain the automatic response), the process proceeds to S18.

[0075] In S15, the response control unit 102A transfers the call from the target person to an operator. At this time, the response control unit 102A may present the target person with a message indicating that the call will be transferred to an operator, as in Ex2 of FIG. 4. The response control unit 102A may also transmit the request (specifically, voice data) accepted in S12 to the operator. When the transfer of the call is complete, the operator begins to respond.

[0076] In S16 after the operator starts responding, the response evaluation unit 106A starts recording the content of the operator's response. When asset management support is provided via voice communication, in S16 the response evaluation unit 106A records the content of the call between the operator and the target person. Recording of the content of the response continues until the operator finishes responding.

[0077] In S17 after the operator has finished responding, the response evaluation unit 106A evaluates the content of the operator's response based on the data recorded while the operator was responding. As described above, the response evaluation unit 106A may provide feedback on the evaluation result to the operator or their superior, etc. This completes the processing in FIG. 5. It is also possible to switch to an automatic response while the operator is responding.

[0078] In S18, the automatic response unit 105A causes the language model to generate a response to the request accepted in S12. Note that in S18, the automatic response unit 105A may also use the related information acquired in S11 to generate the response.

[0079] In S19, the automatic response unit 105A presents the answer generated in S18 to the subject. Specifically, the automatic response unit 105A converts the answer generated in S18 into voice data and transmits it to the subject, thereby presenting the answer to the subject.

[0080] In S20, the reception unit 101A determines whether to end the reception of the request. When asset management support is provided via voice communication, the reception unit 101A ends the reception of the request when the target person ends the call. If the determination in S20 is YES, the processing in FIG. 5 ends.

[0081] On the other hand, if the determination in S20 is NO, the process returns to S12, where a new request from the subject is accepted. Then, in S13, the index value calculation unit 104A recalculates the index value based on the new request from the subject and the request accepted in the first S12. For example, the index value calculation unit 104A may recalculate the index value by inputting the new request from the subject and the request accepted in the first S12 into an index value calculation model. The same applies to the third and subsequent S13, where the index value calculation unit 104A recalculates the index value based on the new request from the subject and each request accepted up to that point. As described above, instead of recalculating the index value every time a new request is accepted, the index value may be updated based on the new request every time a new request is accepted.

[0082] [Modification] The execution entity of each process described in the above exemplary embodiment is arbitrary and is not limited to the above example. For example, a system having the same functions as the support devices 1 and 1A can be constructed using multiple devices that can communicate with each other. Also, the execution entity of each process shown in the flowchart of FIG. 5 may be one device (which can also be called a processor) or multiple devices (which can also be called processors).

[0083] [Software implementation example] Some or all of the functions of the support devices 1 and 1A may be realized by hardware such as an integrated circuit (IC chip), or may be realized by software.

[0084] In the latter case, the support device 1, 1A is realized by, for example, a computer that executes instructions of a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as computer C) is shown in Fig. 6. Fig. 6 is a block diagram showing the hardware configuration of computer C that functions as support device 1 or 1A.

[0085] The computer C includes at least one processor C1 and at least one memory C2. The memory C2 stores a program P for causing the computer C to function as the support device 1 or 1A. In the computer C, the processor C1 reads and executes the program P from the memory C2, thereby realizing the functions of the support device 1 or 1A.

[0086] The processor C1 may be, for example, a central processing unit (CPU), a graphic processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a tensor processing unit (TPU), a quantum processor, a microcontroller, or a combination thereof. The memory C2 may be, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or a combination thereof.

[0087] The computer C may further include a RAM (Random Access Memory) for expanding the program P during execution and for temporarily storing various data. The computer C may also include a communication interface for transmitting and receiving data to and from other devices. The computer C may also include an input / output interface for connecting input / output devices such as a keyboard, mouse, display, and printer.

[0088] Furthermore, the program P can be recorded on a non-transitory tangible recording medium M that can be read by the computer C. Such a recording medium M can be, for example, a tape, a disk, a card, a semiconductor memory, or a programmable logic circuit. The computer C can acquire the program P via such a recording medium M. The program P can also be transmitted via a transmission medium. Such a transmission medium can be, for example, a communication network or broadcast waves. The computer C can also acquire the program P via such a transmission medium.

[0089] Furthermore, each of the above functions of the support device 1 or 1A may be realized by a single processor provided in a single computer, by multiple processors provided in a single computer working together, or by multiple processors provided in each of multiple computers working together. Also, the program for causing the support device 1 or 1A to realize each of the above functions may be stored in a single memory provided in a single computer, or may be distributed and stored in multiple memories provided in a single computer, or may be distributed and stored in multiple memories provided in each of multiple computers.

[0090] [Additional Notes] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0091] (Appendix A1) A support device that supports asset management, comprising: a reception means for receiving requests regarding asset management from a person to be supported; and a response control means for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

[0092] (Appendix A2) The assistance device according to appendix A1, further comprising an index value calculation means for calculating an index value for determining whether or not to switch to a response by an operator based on the content of the request of the subject during the period in which the automatic response is applied, and the response control means determines whether or not to switch the response to the subject to a response by an operator based on the index value.

[0093] (Appendix A3) The support device described in Appendix A2, wherein, when a new request is accepted from the subject, the index value calculation means updates the most recently calculated index value based on the content of the new request, or recalculates the index value based on the content of the new request and the content of a series of requests accepted up to the time the new request is accepted.

[0094] (Appendix A4) The support device described in Appendix A2 or A3, wherein the index value calculation means calculates the index value using an index value calculation model generated by machine learning the relationship between a request from a target of asset management support and an index value indicating the need for an operator to respond to the request.

[0095] (Appendix A5) The support device described in any of Appendices A1 to A4, wherein the response control means determines whether to have an operator respond to the request of the subject or to have an automatic response based on at least one of an evaluation result of the benefit expected to be obtained by having an operator respond to the request of the subject and an evaluation result of the benefit expected to be obtained by having an automatic response to the request of the subject.

[0096] (Appendix A6) An assistance device described in any of Appendices A1 to A5, wherein the response control means decides to have an operator respond to the request when the content of the request from the target person is a request for a proposal of a financial product to invest in.

[0097] (Appendix A7) The support device according to any one of Appendices A1 to A6, wherein the response control means determines that the request should be responded to by an operator who is a qualified person when the content of the request of the subject person indicates that the request requires a response from a qualified person who has specified qualifications.

[0098] (Appendix A8) The assistance device according to any one of appendices A1 to A7, further comprising a response evaluation means for evaluating the validity of the content of the response by the operator.

[0099] (Appendix B1) A method for supporting asset management, in which at least one processor executes a reception process for receiving a request for asset management from a person to be supported, and a response control process for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

[0100] (Appendix B2) The support method described in Appendix B1, further comprising an index value calculation process in which the at least one processor calculates an index value for determining whether or not to switch to a response by an operator based on the content of the subject's request during the period in which the automatic response is applied, and in the response control process, the at least one processor determines whether or not to switch the response to the subject to a response by an operator based on the index value.

[0101] (Appendix B3) The support method described in Appendix B2, wherein, when a new request is received from the subject, the at least one processor updates the most recently calculated index value based on the content of the new request, or recalculates the index value based on the content of the new request and the content of a series of requests received up to the time the new request is received.

[0102] (Appendix B4) The support method described in Appendix B2 or B3, wherein the at least one processor calculates the index value using an index value calculation model generated by machine learning the relationship between a request from a target of asset management support and an index value indicating the need for an operator to respond to the request.

[0103] (Appendix B5) An assistance method described in any of Appendices B1 to B4, wherein in the response control processing, the at least one processor determines whether to have an operator respond to the subject's request or to have an automatic response based on at least one of an evaluation result of the benefit expected to be obtained by having an operator respond to the subject's request and an evaluation result of the benefit expected to be obtained by having an automatic response to the subject's request.

[0104] (Appendix B6) A support method described in any of Appendices B1 to B5, wherein in the response control processing, the at least one processor decides to have an operator respond to the request if the content of the request from the target person is a request for a proposal of a financial product to invest in.

[0105] (Appendix B7) An assistance method described in any of Appendices B1 to B6, wherein in the response control processing, the at least one processor decides to have a qualified operator respond to the request if the content of the request of the subject indicates that the request requires a response from a qualified person with specified qualifications.

[0106] (Appendix B8) The support method according to any one of appendices B1 to B7, further comprising a response evaluation process in which the at least one processor evaluates the validity of the content of the response by the operator.

[0107] (Appendix C1) A support program that causes a computer to function as a support device that supports asset management, and causes the computer to function as a reception means that receives requests regarding asset management from those who are the recipients of support, and a response control means that determines whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

[0108] (Appendix C2) An assistance program as described in Appendix C1, which causes the computer to function as an index value calculation means for calculating an index value for determining whether or not to switch to a response by an operator based on the content of the request of the subject during the period in which the automatic response is applied, and the response control means determines whether or not to switch the response to the subject to a response by an operator based on the index value.

[0109] (Appendix C3) The assistance program described in Appendix C2, wherein the index value calculation means, when a new request is received from the subject, updates the most recently calculated index value based on the content of the new request, or recalculates the index value based on the content of the new request and the content of a series of requests received up to the time the new request is received.

[0110] (Appendix C4) The support program described in Appendix C2 or C3, wherein the index value calculation means calculates the index value using an index value calculation model generated by machine learning the relationship between a request from a target of asset management support and an index value indicating the need for an operator to respond to the request.

[0111] (Appendix C5) The assistance program described in any of Appendices C1 to C4, wherein the response control means determines whether to have an operator respond to the subject's request or to have an automatic response based on at least one of an evaluation result of the benefit expected to be obtained by having an operator respond to the subject's request and an evaluation result of the benefit expected to be obtained by having an automatic response to the subject's request.

[0112] (Appendix C6) An assistance program described in any of Appendices C1 to C5, wherein the response control means decides to have an operator respond to the request when the content of the request from the target person is a request for a proposal for a financial product to invest in.

[0113] (Appendix C7) An assistance program described in any of Appendices C1 to C6, wherein the response control means determines that, when the content of the request of the subject indicates that a response from a qualified person with specified qualifications is required for the request, the qualified operator will respond to the request.

[0114] (Appendix C8) The assistance program according to any one of appendices C1 to C7, which causes the computer to function as a response evaluation means for evaluating the validity of the content of the response by the operator.

[0115] (Appendix D1) A support device that supports asset management, comprising at least one processor, the at least one processor executing a reception process for receiving a request regarding asset management from a person to be supported, and a response control process for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

[0116] The support device may further include a memory, and the memory may store a program for causing the at least one processor to execute each of the processes.

[0117] (Appendix D2) The assistance device described in Appendix D1 executes an index value calculation process that calculates an index value for determining whether or not to switch to a response by an operator based on the content of the request of the subject during the period in which the automatic response is applied, and in the response control process, the at least one processor determines whether or not to switch the response to the subject to a response by an operator based on the index value.

[0118] (Appendix D3) The assistance device described in Appendix D2, wherein, when a new request is accepted from the subject, the at least one processor updates the most recently calculated index value based on the content of the new request, or recalculates the index value based on the content of the new request and the content of a series of requests accepted up to the time the new request is accepted.

[0119] (Appendix D4) The support device described in Appendix D2 or D3, wherein the at least one processor calculates the index value using an index value calculation model generated by machine learning the relationship between a request from a target of asset management support and an index value indicating the need for an operator to respond to the request.

[0120] (Appendix D5) The assistance device according to any one of appendices D1 to D4, wherein in the response control process, the at least one processor determines whether to have an operator respond to the request of the subject or to have an automatic response based on at least one of an evaluation result of the benefit expected to be obtained by having an operator respond to the request of the subject and an evaluation result of the benefit expected to be obtained by having an automatic response to the request of the subject.

[0121] (Appendix D6) An assistance device described in any of Appendices D1 to D5, wherein in the response control processing, the at least one processor decides to have an operator respond to the request if the content of the request from the target person is a request for a proposal of a financial product to invest in.

[0122] (Appendix D7) An assistance device described in any of Appendices D1 to D6, wherein in the response control processing, the at least one processor decides to have a qualified operator respond to the request when the content of the request of the subject indicates that the request requires a response from a qualified person with specified qualifications.

[0123] (Appendix D8) The assistance device according to any one of appendices D1 to D7, wherein the at least one processor executes a response evaluation process that evaluates the validity of the content of the response by the operator.

[0124] (Appendix E) A non-transient recording medium on which is recorded a program that causes a computer to function as a support device for supporting asset management, the program causing the computer to execute a reception process for receiving requests regarding asset management from those who are the recipients of support, and a response control process for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request. [Explanation of symbols]

[0125] 1 Support equipment 101 Reception section (reception means) 102 Response control unit (response control means) 1A Support device 101A Reception unit (reception means) 102A Response control unit (response control means) 104A index value calculation unit (index value calculation means) 106A Response evaluation unit (response evaluation means)

Claims

1. A support device for supporting asset management, a receiving means for receiving requests regarding asset management from the recipient of support; and a response control means for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request.

2. an index value calculation means for calculating an index value for determining whether to switch to a response by an operator based on the content of the request of the subject during the period in which the automatic response is applied; The assistance device according to claim 1 , wherein the response control means determines whether or not to switch the response to the subject to a response by an operator based on the index value.

3. 3. The support device according to claim 2, wherein when a new request is received from the subject, the index value calculation means updates the most recently calculated index value based on the content of the new request, or recalculates the index value based on the content of the new request and the content of a series of requests received up until the new request is received.

4. The support device described in claim 2 or 3, wherein the index value calculation means calculates the index value using an index value calculation model generated by machine learning the relationship between requests from recipients of asset management support and index values ​​indicating the need for an operator to respond to the requests.

5. 4. The assistance device according to claim 1, wherein the response control means determines whether to have an operator respond to the request of the subject or to have an automatic response based on at least one of an evaluation result of the benefit expected to be obtained by having an operator respond to the request of the subject and an evaluation result of the benefit expected to be obtained by having an automatic response to the request of the subject.

6. The support device described in any one of claims 1 to 3, wherein the response control means decides to have an operator respond to the request when the content of the request from the target person is a request for a proposal of a financial product to invest in.

7. 4. The assistance device according to claim 1, wherein the response control means determines that the request should be responded to by an operator who is a qualified person when the content of the request of the subject indicates that the request requires a response from a qualified person who has predetermined qualifications.

8. 4. The assistance device according to claim 1, further comprising a response evaluation unit that evaluates the validity of the content of the response by the operator.

9. A method for supporting asset management, comprising: At least one processor A reception process for receiving requests regarding asset management from the recipient of support; a response control process for determining whether to have an operator respond to the request or to have an automatic response depending on the content of the request;

10. A support program that causes a computer to function as a support device that supports asset management, The computer A means for receiving requests regarding asset management from recipients of support; and a support program that functions as a response control means that determines whether to have an operator respond to the request or to have an automatic response, depending on the content of the request;

Images