Information processing device and program

Abstract

The purpose is to generate user persona extension information. [Solution] By executing a program, the processor of the information processing device 1 functions as: means 31 for receiving data of real users' utterances from a user terminal; means 23, 24 for calculating an empathy score indicating the degree of sympathy of the real user toward the virtual user and a reliability score indicating the degree of trust of the real user toward the virtual user based on the real user's utterances; means 28 for generating persona extension information for the real user by associating the changes in the empathy score and the reliability score with the content of the virtual user's utterances; means 30 for sending a request for generating a virtual user's utterance to a utterance generation device together with data on the utterance history of each of the real user and the virtual user and the persona extension information; means 31 for receiving data of the virtual user's utterances from the utterance generation device; and means 32 for sending the received data of the virtual user's utterances to the user terminal.

Description

[Technical field]

[0001] The present disclosure relates to an information processing device and a program. [Background technology]

[0002] Currently, EC sites (online sales) for apparel and other products are becoming common, and it is expected that in the future there will be more opportunities to open virtual stores in shopping malls in the virtual space (metaverse) on the Internet. In this situation, from the perspective of sales promotion, it is important for EC site staff to introduce recommended products through conversations with customers, to uncover customer needs and materialize the products that users want, and to provide one-on-one customer service to respond to various inquiries, just like in real stores.

[0003] Generally, the number of users who visit an EC site is far greater than the number of customers who visit a real store. Therefore, it is not realistic for real staff to serve each and every user who visits an EC site. Therefore, it is assumed that virtual staff (called virtual users or digital staff) are created on a computer, and the digital staff will serve customers (called real users) one-on-one instead of real staff.

[0004] It is expected that generative AI will be used to create utterances that digital staff can use to advance conversations with customers via chat, etc. The generative AI is sent the history of conversations between the digital staff and the customer along with a generation request. The generative AI analyzes the conversation history and creates the next utterance that the digital staff should make. Since it is used on an e-commerce site, the ultimate goal is sales, which requires high conversational skills and customer service abilities like those of excellent real staff. Excellent real staff are skilled at progressing conversations in a way that builds close relationships, whether consciously or unconsciously, while gaining the sympathy and trust of customers.

[0005] However, generating utterances based only on the input conversation history between the digital staff and the customer was not able to move away from a mechanical conversation. Summary of the Invention [Problem to be solved by the invention]

[0006] The purpose is to generate user persona extension information. [Means for solving the problem]

[0007] The information processing device according to this embodiment has a memory for storing a program, a processor for executing the program, and a communication interface for communicating with a user terminal and a statement generation device (generation AI) via an electrical public communication network, and a virtual user converses with a real user. By executing the program, the processor functions as a means for receiving data of a real user's statement from a user terminal, a means for calculating a sympathy score representing the degree of sympathy of the real user with respect to the virtual user and a reliability score representing the degree of trust of the real user with respect to the virtual user based on the real user's statement, a means for generating persona extension information related to the real user by associating the changes in the sympathy score and the reliability score with the content of the virtual user's statement, a means for transmitting a request for generating a statement of the virtual user to the statement generation device together with data on the statements of the real user and the virtual user and the persona extension information, a means for receiving data of the virtual user's statement from the statement generation device, and a means for transmitting the received data of the virtual user's statement to the user terminal. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a configuration diagram of an entire system including an information processing device according to this embodiment. [Diagram 2] FIG. 2 is a diagram showing a physical configuration of the information processing apparatus shown in FIG. [Diagram 3] FIG. 3 is a diagram illustrating a functional configuration of the processor in FIG. [Figure 4] FIG. 4 is a flowchart showing the flow of conversation processing of the information processing device of FIG. [Diagram 5] FIG. 5 is a flowchart showing the flow of the empathy level estimation process in FIG. [Figure 6] FIG. 6 is a flowchart showing the flow of the reliability estimation process in FIG. [Figure 7] FIG. 7 is a flowchart showing the flow of the intimacy degree estimation process in FIG. [Figure 8] FIG. 8 is a flow chart showing the flow of the wording conversion process by the information processing device of FIG. [Figure 9] FIG. 9 is a diagram showing an example of persona information of a digital staff member. [Figure 10] FIG. 10 is a diagram showing an example of changes in empathy score, trust score, intimacy score, and purchase likelihood score as well as changes in wording as the conversation progresses according to this embodiment. [Figure 11] FIG. 11 is a diagram showing an example of changes in the empathy score, the trust score, the intimacy score, and the purchase likelihood score, as well as changes in wording, following FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] An information processing device according to an embodiment of the present invention will be described below with reference to the drawings. The information processing device according to this embodiment is applied to a scene in which a virtual user (virtual user) constructed on a computer converses with a real user. Here, a scene in which a customer visiting an EC site converses with a virtual staff member (digital staff member) constructed on a computer in a chat room provided on the EC site will be described as an example. In this example, the customer corresponds to the real user, and the digital staff member corresponds to the virtual user. Of course, the present invention is not limited to a scene in which the real user corresponds to the customer, and the virtual user corresponds to the digital staff member.

[0010] As shown in FIG. 1, an information processing device 1 according to the present embodiment is connected to a statement generation device 2 functioning as a generation AI, an EC server 3, and a user terminal 4 via an electrical public communication network 5 such as an Internet line. The user terminal 4 is connected to the EC server 3, and a customer (real user) visits the EC site operated by the EC server 3. The EC site is provided with a chat room or the like, where the customer talks with a digital staff member (virtual user) and purchases a product. The information processing device 1 receives the customer's statement data input from the user terminal 4 directly from the user terminal 4 or indirectly via the EC server 3, creates a statement by the digital staff member in response to the customer's statement, and transmits the same to the user terminal 4 directly or indirectly via the EC server 3. By repeating this process, a conversation between the customer and the digital staff member progresses in the chat room. The information processing device 1 transmits a statement generation request to the statement generation device 2 together with the ordered customer's statements and the digital staff member's statements, that is, the history of the conversation between the customer and the digital staff member.

[0011] In this embodiment, a determination is made based on the customer's remarks as to whether the customer is positive or negative toward the digital staff's remarks, and a resonant degree score is calculated to estimate the degree to which the customer is sympathetic to the digital staff's remarks based on the determination result, and a reliability score is calculated to estimate the degree to which the customer is trustworthy toward the digital staff's remarks based on the determination result.Whether the determination result of whether the customer's remarks are positive or negative is applied to the resonant degree score or the reliability score is selected according to whether the digital staff's remarks are in line with the topic of the conversation. For example, if the EC site is an apparel shop, apparel is applied as the topic of the conversation.

[0012] When the customer is judged to have a positive reaction to the digital staff's comments, the trust score or sympathy score is increased, and when the customer is judged to have a negative reaction, the trust score or sympathy score is decreased. Of course, an increase in the trust score, sympathy score, and further the intimacy score described below means that the customer's trust, sympathy, and intimacy with the digital staff increases.

[0013] By repeating this process, it is possible to determine how the empathy score and trust score change as the conversation progresses, and further to identify what remarks made by the digital staff member caused these changes.

[0014] Furthermore, in real human relationships, as intimacy increases, the politeness of language tends to gradually decrease, and people gradually change from honorific language to friendly language. It is believed that intimacy can be deepened by decreasing the politeness of language and changing from honorific language to friendly language. In other words, if a customer uses friendly language but the digital staff continues to use polite language, the customer may lose a sense of intimacy.

[0015] Therefore, in this embodiment, the politeness level of the digital staff's speech generated by the speech generation device 2 is changed according to the situation so as to follow the change in the politeness level of the customer's speech. The politeness level of the customer's speech is identified, and the politeness level of the digital staff's speech is determined based on the change in the politeness level of the customer's speech identified. When the politeness level of the customer's speech decreases, the politeness level of the digital staff's speech is accordingly decreased by one level from the current politeness level. When the politeness level of the customer's speech increases, the politeness level of the digital staff's speech is also increased by one level from the current politeness level. In this way, in this embodiment, the politeness level of the digital staff's speech is changed to match the politeness level of the customer, thereby making it possible to increase the customer's sense of intimacy while suppressing a decrease in the sense of intimacy.

[0016] In this embodiment, the intimacy score of the customer with respect to the digital staff is calculated based on changes in the politeness of the language used in the customer's speech. When the politeness of the language used in the customer's speech decreases, the intimacy score of the customer with respect to the digital staff is increased. When the politeness of the language used in the customer's speech increases, the intimacy score of the customer with respect to the digital staff is decreased.

[0017] When a request for generating a message is sent to the message generation device 2, the history of the conversation between the customer and the digital staff as well as the changes in the empathy score, trust score, and intimacy score are sent, so that the digital staff's message can be generated in accordance with the flow of the conversation and in a way that increases the empathy score and trust score rather than decreasing them.

[0018] 2, the information processing device 1 as a conversation device has a processor 11 connected to a RAM 12, a ROM 13, a storage unit 14, an input device 15, a display 16, and a communication interface 17 via a system bus 10. The processor 11 is composed of, for example, a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). The processor 11 executes a conversation program loaded from the storage unit 14 and the ROM 13 to the RAM 12, and processes a conversation between a customer and a digital staff member.

[0019] The RAM 12 functions as a main memory, a work area, etc. for the processor 11. The ROM 13 or storage unit 14 stores a BIOS (Basic Input Output System) executed by the processor 11, an operating system program (OS), a conversation processing program according to this embodiment, programs for implementing various other functions, various data required for these processes, etc.

[0020] The input device 15 is composed of a keyboard (KB), a mouse, a touch panel, or other pointing device, etc. The display 19 is typically realized by an LCD (Liquid Crystal Display). In addition to the conversation processing program, the memory unit 14 stores data related to the conversation history between the customer and the digital staff, the customer's persona information and persona extension information exemplified in Fig. 9, etc.

[0021] As shown in Figure 3, by executing the conversation processing program, processor 11 functions as a control unit 21, a natural language analysis processing unit 22, an empathy estimation processing unit 23, a reliability estimation processing unit 24, an intimacy estimation processing unit 25, a purchase possibility estimation processing unit 26, a persona information update unit 27, a persona extension information update unit 28, a wording conversion processing unit 29, a statement generation request processing unit 30, a receiving processing unit 31, and a transmitting processing unit 32.

[0022] The data of the comment input by the customer from the user terminal 4 is received by the receiving processor 31 via the EC server 3. Note that the data of the comment input by the customer from the user terminal 4 may be received directly from the user terminal 4 to the receiving processor 31 without passing through the EC server 3.

[0023] The natural language analysis processing unit 22 performs natural language analysis on the data of customer utterances received via the reception processing unit 31, breaks it down into parts of speech, extracts character strings of nouns in particular, and performs syntactic analysis. The natural language analysis processing unit 22 also performs natural language analysis on the utterances of digital staff transmitted to the EC server 3 via the transmission processing unit 32, breaks it down into parts of speech, extracts character strings of nouns, and performs syntactic analysis. The parts of speech broken down by the natural language analysis processing unit 22, the extracted nouns, and the syntactic analysis results are used in the processes of the empathy estimation processing unit 23, the reliability estimation processing unit 24, the persona information update unit 27, and the persona extension information update unit 28.

[0024] The empathy estimation processing unit 23 calculates the empathy score when the content of the digital staff's statement is not in line with the topic of the conversation, and does not calculate the empathy score when the content of the digital staff's statement is in line with the topic of the conversation. For example, if the EC site is an apparel shop, the topic of the conversation is the apparel field. Specifically, if the noun included in the digital staff's statement is apparel-related, the content of the digital staff's statement is determined to be in line with the topic of the conversation, and if the noun included in the digital staff's statement is not apparel-related and unrelated to apparel, the content of the digital staff's statement is determined to be not in line with the topic of the conversation. This determination process may be performed by creating an apparel-related term dictionary in advance and storing it in the storage unit 14, and checking the noun included in the digital staff's statement against the dictionary to confirm whether it is listed in the apparel-related term dictionary, or by using a learned model that has been machine-learned to estimate the suitability of the noun and the topic of the conversation as teacher data.

[0025] When the customer is positive about the digital staff's remark, a predetermined value is added to the current value of the empathy score to increase the empathy score. When the customer is negative about the digital staff's remark, a predetermined value is subtracted from the current value of the empathy score to decrease the empathy score. Typically, the added value and the subtracted value are the same value, but the absolute value of the added value may be higher or lower than the absolute value of the subtracted value. The determination process for identifying whether the customer's remark is positive or negative in response to the digital staff's remark may be performed by creating a dictionary containing multiple positive expressions and multiple negative expressions in advance and storing it in the storage unit 14, and determining whether the customer's remark contains any positive or negative expression, or by using a learned model that has been machine-learned to estimate multiple expressions and the possibility of being positive or negative as teacher data.

[0026] The reliability estimation processing unit 24 calculates the reliability score when the content of the digital staff's statement is in line with the topic of the conversation, and does not calculate the reliability score when the content of the digital staff's statement is not in line with the topic of the conversation. When the customer makes a positive statement in response to the digital staff's statement, a predetermined value is added to the current value of the reliability score in order to increase the reliability score. When the customer sends a negative statement in response to the digital staff's statement, a predetermined value is subtracted from the current value of the reliability score in order to decrease the reliability score. Typically, the added value and the subtracted value are the same value, but the absolute value of the added value may be higher or lower than the absolute value of the subtracted value.

[0027] The intimacy estimation processor 25 compares the language used in the customer's speech with that used in the digital staff's speech, and calculates an intimacy score that indicates the degree of intimacy of the customer with the digital staff based on the comparison result. Specifically, when the politeness of the language used in the customer's speech decreases, a predetermined value is added to the current value of the intimacy score to increase the intimacy score. When the politeness of the language used in the customer's speech increases, a predetermined value is subtracted from the current value of the intimacy score to decrease the intimacy score. Typically, the added value and the subtracted value are the same value, but the absolute value of the added value may be higher or lower than the absolute value of the subtracted value.

[0028] The level of politeness of language is typically determined by creating in advance a dictionary of multiple ending expressions relating to multiple levels of politeness, from honorific language to friendly language, and storing the dictionary in memory unit 14.The ending expressions in the customer's speech are then compared with the dictionary to determine the level of politeness of the corresponding ending expressions.

[0029] It is assumed that the likelihood that a customer will purchase a product increases as the empathy score, the reliability score, and the intimacy score increase. Therefore, the purchase possibility estimation processing unit 26 calculates a purchase possibility score (overall score) that indicates the likelihood that a customer will purchase a product based on the calculated empathy score, reliability score, and intimacy score. For example, the purchase possibility score is calculated by adding the amount of change in the empathy score, the reliability score, and the intimacy score, or a value obtained by multiplying the amount of change by a predetermined coefficient such as 0.5, to the current value of the purchase possibility score. The purchase possibility score may be calculated by simply adding the empathy score, the reliability score, and the intimacy score, or may be calculated by weighted addition of the empathy score, the reliability score, and the intimacy score. The calculation method may be changed arbitrarily.

[0030] The persona information update unit 27 generates new persona information related to the hometown, preferences, etc. based on the parts of speech and syntax analysis results resolved by the natural language analysis processing unit 22, and adds it to the existing persona information related to the customer to update it. The persona information is initially user information registered on the EC site. As the conversation progresses, new persona information is generated and added. The persona information is stored in the storage unit 14.

[0031] The persona extension information update unit 28 generates persona extension information that associates the changes in the empathy score and the reliability score with the content of the conversation of the digital staff member that is thought to have caused the changes, here, nouns. In other words, the persona extension information is generated as information that can identify whether the nouns in the comments of the digital staff member lead to an increase or decrease in the empathy score and the reliability score. For example, as persona extension information for a certain customer, an increase in the reliability score is associated with "blue" and "jacket". The persona extension information is generated as appropriate as the conversation progresses, and is stored in the storage unit 14.

[0032] The language conversion processing unit 29 converts the language of the digital staff's message received from the message generation device 2 into language with a set level of politeness. Initially, the digital staff's messages received from the message generation device 2 are often mechanical, businesslike, and dry. As they are, they do not demonstrate a high level of intimacy. Therefore, a dictionary is created in advance to convert the language into multiple levels of politeness, from honorific language to friendly language, and is stored in the memory unit 14. The language conversion processing unit 29 converts the words in the digital staff's message received from the message generation device 2 into the set language using the dictionary.

[0033] As will be described in more detail below, the politeness of the customer's language is identified, and when that politeness changes in the course of the conversation, the politeness of the language used by the digital staff is changed to follow the change. For example, when the politeness of the customer's language decreases, that is, when the customer's intimacy with the digital staff increases, the politeness of the language used by the digital staff is decreased by one level from the current level. When the politeness of the customer's language increases, that is, when the customer's intimacy with the digital staff decreases, the politeness of the language used by the digital staff is increased by one level from the current level. By changing the politeness of the language used by the digital staff in this way to follow changes in the politeness of the customer's language, the customer's intimacy with the digital staff can be increased and a decrease in the customer's intimacy with the digital staff can be suppressed.

[0034] The statement generation request processing unit 30 transmits data on the conversation history between the customer and the digital staff member, including the customer's most recent statement, together with a request to generate a statement by the digital staff member in response to the customer's statement, to the statement generation device 2. Furthermore, in addition to the conversation history data, the statement generation request processing unit 30 transmits persona information, persona extension information, changes in empathy score, changes in trust score, changes in intimacy score, and changes in purchase possibility score to the statement generation device 2. By referring to this information, the statement generation device 2 can generate a statement by the digital staff member that follows the flow of the conversation and that increases the empathy score and trust score rather than decreasing them.

[0035] 4 shows a flow of conversation processing of the information processing device 1. When a conversation starts, under the control of the control unit 21, the empathy estimation processing unit 23, the reliability estimation processing unit 24, the intimacy estimation processing unit 25, and the purchase possibility estimation processing unit 26 initialize their respective scores (S1). For example, the empathy score, reliability score, and intimacy score are initialized to zero, and the purchase possibility score is initialized to 50 as a median value. When a conversation is repeated between the same customer and digital staff, the empathy score value, reliability score value, intimacy score value, and purchase possibility score value at the end of the previous conversation may be carried over.

[0036] Data on customer comments is received from the EC server 3 via the receiving processor 31 (S2). Data on digital staff comments is read from the memory unit 14. The customer comments and digital staff comments are subjected to natural language analysis processing (S3). Parts of speech are broken down, nouns are extracted, and syntactic analysis is performed.

[0037] Using the natural language analysis result, the empathy estimation processing unit 23 executes empathy estimation processing to calculate a empathy score (S4), the reliability estimation processing unit 24 executes reliability estimation processing to calculate a reliability score (S5), and the intimacy estimation processing unit 25 executes intimacy estimation processing to calculate an intimacy score (S6). Each process will be described later. The purchase possibility estimation processing unit 26 calculates a purchase possibility score representing the possibility that the customer will purchase a product based on the empathy score, reliability score, and intimacy score (S7). The method of calculating the purchase possibility score representing the possibility that the customer will purchase a product on the EC site from the empathy score, reliability score, and intimacy score is arbitrary, and for example, the purchase possibility score is calculated by adding the amount of change in each of the empathy score, reliability score, and intimacy score, or a value obtained by multiplying the amount of change by a predetermined coefficient such as 0.5, to the current value of the purchase possibility score.

[0038] Next, the persona extension information update unit 28 generates persona extension information that associates each change in the empathy score, the trust score, and the intimacy score with a noun in the statement made by the digital staff member immediately before the customer's statement indicating the change (S8). As a result, when a change occurs in the customer's empathy score, trust score, or intimacy, it is possible to recognize the noun in the statement made by the digital staff member that caused the change. For example, when a customer's trust score increases, it is possible to identify that the increase was caused by "blue" and "jackets," in other words, that the customer is interested in "blue" and "jackets," or to identify the customer's preferences.

[0039] In the next step S9, the message generation request processor 30 generates a message generation request together with the conversation history, persona information, persona extension information, empathy score transition, trust score transition, intimacy score transition, and purchase possibility score transition. The transmission processor 32 transmits the message generation request together with the data of the conversation history, persona information, persona extension information, empathy score transition, trust score transition, intimacy score transition, and purchase possibility score transition to the message generation device 2 via the communication interface 17 (S9).

[0040] The receiving processor 31 receives the data of the digital staff's statement from the statement generating device 2 via the communication interface 17 (S10). The language conversion processor 29 converts the language of the digital staff's statement to a preset level of politeness (S11). The sending processor 32 sends the data of the digital staff's statement, with the language converted, to the EC server 3 via the communication interface 17 (S12). If the conversation is continuing (S13, NO), the process returns to step S2 and waits for the next customer's statement. If the conversation has ended (S13, YES), the conversation process ends.

[0041] Fig. 5 shows a flow of the empathy estimation process in step S4 in Fig. 4. Please refer to examples of changes in empathy score, trust score, intimacy score, and purchase likelihood score, as well as changes in wording, shown in Figs. 10 and 11.

[0042] The empathy estimation processing unit 23 judges whether the content of the digital staff's statement is in line with the topic of the conversation based on the analysis result of the natural language analysis processing unit 22 (S21). For example, if the EC site is an apparel shop, the topic of the conversation is the apparel field. Typically, the nouns included in the digital staff's statement are checked against an apparel-related terminology dictionary to confirm whether they are listed in the apparel-related terminology dictionary. When the nouns included in the digital staff's statement match any of the terms listed in the apparel-related terminology dictionary, it is judged that the content of the digital staff's statement is in line with the topic of the conversation (S21, YES). When the nouns included in the digital staff's statement do not match any of the terms listed in the apparel-related terminology dictionary, it is judged that the content of the digital staff's statement is not in line with the topic of the conversation (S21, NO). When it is judged that the content of the digital staff's statement is in line with the topic of the conversation (S21, YES), the next statement by the digital staff is awaited. When it is determined that the content of the digital staff's remarks is not in line with the topic of the conversation (S21, NO), the process proceeds to the next step S22, and the calculation process of the empathy score begins. When it is determined that the digital staff's remarks are in line with the topic of the conversation, that is, that the digital staff's remarks are within the scope of the topic of the conversation, the empathy score is not calculated, but a reliability score is calculated. In other words, when the digital staff's remarks are casual conversation, the empathy score is calculated, and when the digital staff's remarks are not casual conversation but are actually in line with the topic of the consultation, a reliability score is calculated.

[0043] In step S22, the natural language analysis processing unit 22 analyzes the customer's statement, and in step S23, the empathy estimation processing unit 23 distinguishes whether the customer's statement is positive, negative, or neither of these in relation to the statement of the digital staff based on the analysis result of the natural language analysis processing unit 22. Typically, the predicate included in the customer's statement is compared with a dictionary that includes a plurality of positive expressions and a plurality of negative expressions created in advance, and whether or not it is included in the dictionary is checked to distinguish whether the customer's statement is positive, negative, or neither of these.

[0044] When the customer's comment is positive, the empathy estimation processing unit 23 adds a predetermined value I1 to the current value of the empathy score, thereby increasing the empathy score (S24). When the customer's comment is negative, the empathy estimation processing unit 23 subtracts a predetermined value D1 from the current value of the empathy score, thereby decreasing the empathy score (S25). When the customer's comment is neither positive nor negative, the empathy estimation processing unit 23 assumes that the customer is indifferent to the comments made by the digital staff, and subtracts a predetermined value D2, the absolute value of which is greater than the predetermined value D1, from the current value of the empathy score, thereby further decreasing the empathy score (S26).

[0045] Fig. 6 shows the flow of the reliability estimation process in step S5 in Fig. 4. The reliability estimation processing unit 24 judges whether the content of the digital staff's statement is in line with the topic of the conversation based on the analysis result of the natural language analysis processing unit 22 (S31). The judgment method is the same as in step S21. If it is judged that the content of the digital staff's statement is not in line with the topic of the conversation (S31, YES), the process waits for the next statement from the digital staff. If it is judged that the content of the digital staff's statement is in line with the topic of the conversation (S31, NO), the process proceeds to the next step S32 and starts the calculation process of the reliability score.

[0046] In step S32, the natural language analysis processing unit 22 analyzes the customer's comment, and in step S33, the reliability estimation processing unit 24 distinguishes whether the customer's comment is positive, negative, or neither in relation to the comment made by the digital staff member based on the analysis result of the natural language analysis processing unit 22. The method of this distinction is also the same as that of (S1) S23.

[0047] When the customer's comment is positive, the reliability estimation processing unit 24 adds a predetermined value I11 to the current value of the reliability score to increase the reliability score (S34). When the customer's comment is negative, the reliability estimation processing unit 24 subtracts a predetermined value D11 from the current value of the reliability score to decrease the reliability score (S35). When the customer's comment is neither positive nor negative, the reliability estimation processing unit 24 assumes that the customer is indifferent to the comments made by the digital staff and subtracts a predetermined value D12, the absolute value of which is greater than the predetermined value D11, from the current value of the reliability score to further decrease the reliability score (S36).

[0048] Fig. 7 shows a flow of the intimacy estimation process in step S6 in Fig. 4. The intimacy estimation processing unit 25 specifies the level of politeness of the language used in the customer's speech (S41). The level of politeness of the language is determined by checking the endings of the customer's speech against a dictionary of multiple ending expressions related to multiple levels of politeness, from honorific language to friendly language, and checking the level of politeness of the endings of the customer's speech that match the endings of the customer's speech and are listed in the dictionary.

[0049] The intimacy estimation processor 25 determines whether the politeness level of the specified customer's speech has changed from the politeness level of the previous customer's speech (S42). If the politeness level has not changed (S42, NO), the process returns to step S41 and waits for the next customer's speech. If the politeness level has changed (S42, YES), it is determined whether the politeness level of the specified customer's speech has decreased from the politeness level of the previous customer's speech (S43).

[0050] When it is determined that the politeness of the utterance of the specified customer has decreased compared to the politeness of the utterance of the previous customer (S43, YES), a predetermined value I21 is added to the current value of the intimacy score to increase the intimacy score (S44).When it is determined that the politeness of the utterance of the specified customer has increased compared to the politeness of the utterance of the previous customer (S43, NO), a predetermined value D21 is subtracted from the current value of the intimacy score to decrease the intimacy score (S45).

[0051] Fig. 8 shows the flow of the wording change process. The wording change process is executed in parallel with the conversation process shown in Fig. 4. Prior to the conversation, the wording conversion processing unit 29 first initially sets the wording of the digital staff member's speech to "5", which is the middle level of politeness out of 11 levels (S51). When receiving a customer's speech, the wording conversion processing unit 29 identifies the level of politeness of the customer's speech (S52). Then, the wording conversion processing unit 29 determines whether the level of politeness of the identified customer's speech has changed from the level of politeness of the previous customer's speech (S53).

[0052] If the politeness level has not changed (S53, NO), the process returns to step S52 and waits for the next customer's comment. If the politeness level has changed (S53, YES), a determination is made as to whether the politeness level of the specified customer's comment has decreased compared to the politeness level of the previous customer's comment (S54).

[0053] When it is determined that the level of politeness of the specified customer's speech has decreased compared to the level of politeness of the previous customer's speech (S54, YES), that is, when the customer's intimacy with the digital staff has increased, the digital staff's language is changed to a level of politeness one level lower accordingly (S55).When it is determined that the level of politeness of the specified customer's speech has increased compared to the level of politeness of the previous customer's speech (S54, NO), that is, when the customer's intimacy with the digital staff has decreased, the digital staff's language is changed to a level of politeness one level higher accordingly (S56).

[0054] If the conversation is continuing (S57, NO), the process returns to step S52 and waits for the next customer's comment. If the conversation has ended (S57, YES), the wording change process ends.

[0055] As described above, according to this embodiment, the customer's empathy score, trust score, and intimacy score towards the digital staff member in response to a comment made by the digital staff member are calculated, and the changes in these empathy scores and trust scores, as well as the conversation history, are sent to the comment generation device 2 along with a request to generate a comment.This enables the comment generation device 2 to generate comments by the digital staff member that increase the empathy and trust levels and prevent a decrease in empathy and trust levels.

[0056] Furthermore, according to this embodiment, by changing the politeness of the digital staff's language in response to changes in the politeness of the customer's speech, it is possible to increase the level of intimacy and prevent a decrease in that level of intimacy.

[0057] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents described in the claims, as well as in the scope and spirit of the invention. [Explanation of symbols]

[0058] 1...information processing device, 2...statement generation device (generation AI), 3...EC server, 4...user terminal, 11...processor, 10...system bus, 12...RAM, 13...ROM, 14...memory unit, 15...input device, 16...display, 17...communication interface, 21...control unit, 22...natural language analysis processing unit, 23...empathy estimation processing unit, 24...trustworthiness estimation processing unit, 25...intimacy estimation processing unit, 26...purchase possibility estimation processing unit, 27...persona information update unit, 28...persona extension information update unit, 29...language conversion processing unit, 30...statement generation request processing unit, 31...reception processing unit, 32...transmission processing unit.

Claims

1. An information processing device having a memory for storing a program, a processor for executing the program, and a communication interface for communicating with a user terminal and a message generation device (generation AI) via an electrical public communication network, the information processing device allowing a virtual user to converse with a real user, The processor executes the program, A means for receiving data of utterances of the real user from the user terminal; a score calculation means for calculating a sympathy score representing a degree of sympathy of the real user with the virtual user and a trust score representing a degree of trust of the real user with the virtual user based on the utterances of the real user; a persona extension information generating means for generating persona extension information for the real user by associating the changes in the empathy score and the reliability score with the content of the utterance of the virtual user; a means for transmitting a request for generating a utterance of the virtual user to the utterance generating device together with data on the utterance history of each of the real user and the virtual user and the persona extension information; means for receiving data of utterances of the virtual user from the utterance generation device; an information processing device that functions as a means for transmitting the received utterance data of the virtual user to the user terminal.

2. The information processing device according to claim 1 , wherein the persona expansion information generation means associates nouns contained in the utterances of the virtual user as factors for the changes in the empathy score and the reliability score, respectively, with the changes in the empathy score and the reliability score, respectively.

3. 2. The information processing device according to claim 1, wherein the score calculation means determines whether the virtual user's utterances are in line with the topic of the conversation, calculates the empathy score when the virtual user's utterances are not in line with the topic of the conversation, and calculates the reliability score when the virtual user's utterances are in line with the topic of the conversation.

4. The information processing device of claim 3, wherein the score calculation means determines whether the real user has a positive or negative opinion of the virtual user's comment based on the comment of the real user, and calculates the empathy score and the reliability score based on the result of the determination.

5. The information processing device of claim 4, wherein the score calculation means increases the empathy score or the reliability score when it is determined that the real user is positive about the virtual user's utterance, and decreases the empathy score or the reliability score when it is determined that the real user is negative about the virtual user's utterance.

6. An information processing device having a memory for storing a program, a processor for executing the program, and a communication interface for communicating with a first user terminal and a second user terminal via an electrical public communication network, The processor executes the program, means for receiving data of a utterance of a first user and data of a utterance of a second user from the first user terminal and the second user terminal, respectively; a score calculation means for calculating a sympathy score representing a degree of sympathy of the first user with the second user and a trust score representing a degree of trust of the first user with the second user based on the utterance of the first user; An information processing device that functions as a persona extension information generation means that associates the changes in the empathy score and the reliability score with the content of the second user's comments and generates them as persona extension information regarding the first user.

7. an information processing device having a communication interface for communicating with a user terminal and a statement generation device (generation AI) via an electrical public communication network, the information processing device allowing a virtual user to converse with a real user; A means for receiving data of utterances of the real user from the user terminal; a means for calculating a sympathy score representing a degree of sympathy of the real user with the virtual user and a trust score representing a degree of trust of the real user with the virtual user based on the utterances of the real user; A program that functions as a means for associating the changes in the empathy score and the confidence score with the content of the virtual user's utterances and generating them as persona extension information about the real user.

Images