Information processing method, information processing device, and computer program

Abstract

The present invention involves: acquiring inquiry information regarding a molding machine from a user; and acquiring answer support information contributing to an answer to the inquiry by using the acquired inquiry information and a language model based on a database which stores information regarding the molding machine.

Description

Information processing method, information processing device, and computer program

[0001] The present disclosure relates to an information processing method, an information processing device, and a computer program.

[0002] As part of customer service in various businesses, including manufacturing and sales, call centers are established where operators handle telephone inquiries from customers. Call centers have introduced contact center systems that support responses to customer inquiries (see, for example, Patent Literature 1). The contact center system has functions such as a pop-up function that searches for customer information using an incoming telephone number and displays the searched information on the operator's computer screen, and a function that records the contents of the call.

[0003] JP 2016-39383 A

[0004] However, when responding to inquiries about molding machines such as injection molding machines and extruders, there are cases where the technical details need to be confirmed with the relevant department, which results in a time-consuming response. Inquiries from company employees to sales representatives also result in a time-consuming response. Even if the company had the technical information about molding machines, it was difficult to select the necessary information.

[0005] An object of the present disclosure is to provide an information processing method, an information processing device, and a computer program that can support responses to inquiries about molding machines and improve the satisfaction of users (customers or company employees).

[0006] An information processing method according to one aspect of the present disclosure acquires inquiry information from a user regarding a molding machine, and acquires response support information useful in responding to the inquiry using the acquired inquiry information and a language model that utilizes a database that stores information regarding the molding machine.

[0007] An information processing device according to one aspect of the present disclosure includes a processing unit for acquiring inquiry information regarding a molding machine from a user, and executing a process for acquiring response support information useful in responding to the inquiry using the acquired inquiry information and a language model that utilizes a database that stores information regarding the molding machine.

[0008] A computer program according to one aspect of the present disclosure acquires inquiry information regarding a molding machine from a user, and causes a computer to execute a process of acquiring response support information useful in answering the inquiry using the acquired inquiry information and a language model that utilizes a database that stores information regarding the molding machine.

[0009] According to the present disclosure, it is possible to assist in responding to inquiries about molding machines and improve user satisfaction.

[0010] FIG. 1 is a schematic diagram showing an example of the configuration of an inquiry support system according to embodiment 1. FIG. 2 is a block diagram showing an example of the configuration of an information processing device according to embodiment 1. FIG. 3 is a flowchart showing the processing procedure of the information processing device according to embodiment 1. FIG. 4 is a schematic diagram showing an example of the configuration of an inquiry support system according to embodiment 2. FIG. 5 is a schematic diagram showing an example of the configuration of a database according to embodiment 2. FIG. 6 is a schematic diagram showing an example of the configuration of a database according to embodiment 2. FIG. 7 is a schematic diagram showing an example of the configuration of a database according to embodiment 2. FIG. 8 is a flowchart showing the processing procedure of an information processing device according to embodiment 3. FIG. 9 is a block diagram showing an example of the configuration of an information processing device according to embodiment 4. FIG. 10 is a flowchart showing the processing procedure of an information processing device according to embodiment 4.

[0011] An information processing method, an information processing device, and a computer program according to embodiments of the present disclosure will be described below with reference to the drawings. Note that the present disclosure is not limited to these examples, but is defined by the claims, and is intended to include all modifications within the meaning and scope equivalent to the claims. Furthermore, at least some of the embodiments described below may be combined in any manner.

[0012] 1 is a schematic diagram showing an example of the configuration of an inquiry response support system according to embodiment 1. The inquiry response support system according to embodiment 1 includes a contact center system 10 installed in a call center and a user's telephone (hereinafter referred to as a user telephone 1). The contact center system 10 is communicatively connected to the user telephone 1 via a communication network such as a public switched telephone line 2 (hereinafter referred to as a "PSTN 2").

[0013] A user is someone who makes an inquiry about a molding machine. The main users are customers who purchase and use the molding machine, and inquiries are directed to sales representatives of manufacturers that produce and sell molding equipment. Users who make inquiries about molding machines also include employees of the manufacturer.

[0014] The molding machines in question include plastic molding machines and metal molding machines. Plastic molding machines include injection molding machines and extrusion machines, while metal molding machines include magnesium injection molding machines.

[0015] The contact center system 10 is a system that executes processes related to telephone calls between users and sales representatives and supports the sales representatives in responding to inquiries from users. The contact center system 10 includes a private branch exchange (PBX) 3, a telephone of the sales representative (hereinafter referred to as a sales representative telephone 4), a display operation terminal 5 for the sales representative, an internal communication network 6 such as a LAN, and an information processing device 7.

[0016] The sales representative's telephone 4 is an internal telephone such as an IP telephone installed in the contact center, and is equipped with a speaker and a microphone. The sales representative's telephone 4 enables calls by sending and receiving voice data between the user and the sales representative via IP communication via the PBX 3. The sales representative's telephone 4 is used by the sales representative to answer calls from users while operating the display operation terminal 5.

[0017] The display operation terminal 5 is a computer equipped with a CPU, memory, a hard disk drive, etc., and is connected to input devices such as a keyboard and a mouse, and an output device such as a display monitor. The display operation terminal 5 is a device for supporting responses to inquiries from users, and displays the caller's telephone number (hereinafter referred to as the incoming telephone number), the time of the call, user information, past response history, information necessary for responding to the inquiry, etc. In particular, in this embodiment 1, the information necessary for responding to the inquiry includes specialized information necessary for responding to technical inquiries regarding the molding device. The display operation terminal 5 can also accept input of response content from sales representatives.

[0018] The PBX 3 has a call control function, an automatic distribution function, a CTI (Computer Telephony Integration) function, etc. The call control function is a function that receives (incoming calls) from the user telephone 1 via the PSTN 2 and controls the connection with the sales representative telephone 4. The call control function enables voice communication between the user and the sales representative. The automatic distribution function is a function that, when a call is received from a user, distributes the call to the sales representative telephone 4 of an available sales representative and connects the call to the sales representative telephone 4. The CTI function is a function that identifies the called telephone number, which is the telephone number of the caller, from the incoming call and notifies the display operation terminal 5, the information processing device 7, etc. of CTI information including the identified called telephone number, the time of the call, and the response history ID.

[0019] The PBX 3 performs line switching and protocol conversion between the PSTN 2 and the private network 6, and transmits and receives information relating to calls between the user and the sales representative between the display / operation terminal 5 and the information processing device 7.

[0020] Specifically, the PBX 3 issues a response history ID for an incoming call. The PBX 3 transmits information such as CTI information linking the response history ID with the incoming telephone number, the incoming time, etc., the call voice data, an incoming call message notifying the call, and a call termination message notifying the call termination to the display operation terminal 5. The display operation terminal 5 receives the CTI information and the call voice data, and displays the CTI information.

[0021] The PBX 3 also transmits CTI information and call voice data in real time to the information processing device 7. The information processing device 7 stores the CTI information and call voice data transmitted from the PBX 3. Furthermore, the PBX 3 transfers response information that is created by a sales representative responding to a user and transmitted from the display operation terminal 5 to the information processing device 7. The information processing device 7 stores the response information as a call history.

[0022] The information processing device 7 is a server computer connected to the PBX 3 and the display operation terminal 5 via the in-house communication network 6, and is a device that executes processing related to CTI functions and assists in responding to inquiries from users regarding the molding device.

[0023] 2 is a block diagram showing an example of the configuration of the information processing device 7 according to embodiment 1. The information processing device 7 includes a processing unit 71, a storage unit 72, and a communication unit 73. Note that the information processing device 7 may be configured to perform distributed processing using multiple computers, may be realized by multiple virtual machines provided in a single server, or may be partially configured by a quantum computer.

[0024] The processing unit 71 is a processor having one or more arithmetic processing devices such as a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit), and implements the information processing method of this embodiment by reading and executing a computer program 72a stored in the memory unit 72.

[0025] The communication unit 73 includes a processing circuit for performing communication-related processing, a communication circuit, and the like, and transmits and receives information between the PBX 3 and the display operation terminal 5 .

[0026] The storage unit 72 is a storage device such as a hard disk, an EEPROM (Electrically Erasable Programmable Read-Only Memory), a flash memory, etc. The storage unit 72 stores a computer program 72 a, a language model 72 b, a customer DB 72 c, a call history DB 72 d, and a technical information DB 72 e.

[0027] The computer program 72a may be recorded in a computer-readable manner on the recording medium 70. The storage unit 72 stores the computer program 72a read from the recording medium 70 by a reading device (not shown). The recording medium 70 may be a semiconductor memory such as a flash memory, an optical disk, a magnetic disk, a magneto-optical disk, or the like. Alternatively, various programs according to this embodiment may be downloaded from an external server (not shown) and stored in the storage unit 72.

[0028] The language model 72b is, for example, a large-scale language model (LLM). The large-scale language model is a model based on a Transformer architecture, including GPT (Generative Pretrained Transformer, including GPT-1, GPT-2, GPT-3, GPT-4, etc.), BERT (Bidirectional Encoder Representations from Transformers), BART (Bidirectional and Auto-regressive Transformer), LLaMA (Large Language Model Meta AI), etc. Furthermore, the large-scale language model may employ Claude 3 Sonnet, which employs a rectified flow structure and a multimodal diffusion transformer. Note that, although a Transformer is used as an example of the architecture constituting the large-scale language model, the architecture of the language model is not particularly limited. While a general-purpose model may suffice as the large-scale language model, it is an LLM that is fine-tuned to output answer sentences or answer support information including information useful for answering in response to prompts including sentences based on at least user inquiry information. The inquiry information from the user is voice text data obtained by speech recognition of the user's call voice data.

[0029] In particular, the information processing device 7 according to the first embodiment is configured to employ RAG (Retrieval Augmented Generation) to generate prompts including sentences based on special technical information about the molding machine and inquiry information, and provide the prompts to the language model 72b, thereby executing a process of acquiring response support information for accurately answering inquiries about the molding machine. Details will be described later.

[0030] The large-scale language model is realized by the processing unit 71 executing information processing in accordance with the computer program 72a. The storage unit 72 stores data for realizing the large-scale language model. The large-scale language model may be configured by hardware. The information processing device 7 may also be configured to execute processing using an external large-scale language model. In other words, the language model 72b may be one provided by an external API. The large-scale language model may also be realized using multiple computers connected via a communication network, or may be realized using the cloud.

[0031] The customer DB 72c is a database that stores data on customers who have purchased molding machines. For example, the customer DB 72c stores information such as a customer ID, the customer's name or title, the customer's telephone number, the customer's address, the model of the purchased molding machine, and the purchase date, in association with each other.

[0032] The call history DB 72d is a database that stores CTI information, inquiry information including text data obtained by speech recognition of call voice data, response information, etc., in association with each other. The CTI information includes a response history ID, an incoming telephone number, and an incoming call time. The processing unit 71 of the information processing device 7 receives call voice data transmitted from the user telephone 1 in real time via the PBX 3. The processing unit 71 converts the received call voice data into text data by speech recognition, and stores the converted text data as inquiry information in the call history DB 72d in association with the response history ID. Note that the inquiry information, which is text data converted in real time, may be configured to be temporarily stored in the storage unit 72 until the call ends.

[0033] The technical information DB 72e is a database that stores information about molding machines purchased and used by users. For example, the technical information DB 72e stores past sales inquiry history information. The inquiry history information includes details of the inquiry, notes, and responses to the inquiry. The technical information DB 72e also stores molding machine instruction manuals, specifications, leaflets, catalog information, PR materials, learning materials for internal sales, and learning materials for purchasers on how to operate the machine. The technical information DB 72e may also store information such as general technical features, technical reports, and papers related to molding machines. The technical information DB 72e may also be configured to store specific specifications of molding machines delivered to users and other technical information in association with customer IDs. The technical information DB 72e associates technical information with vector data representing its characteristics to enable vector searches for technical information similar to user inquiry information.

[0034] 3 is a flowchart showing the processing procedure of the information processing device 7 according to the first embodiment. The processing unit 71 of the information processing device 7 acquires inquiry information from the call history DB 72d or the storage unit 72 (step S11) and extracts technical information related to the inquiry information from the technical information DB 72e (step S12). For example, the processing unit 71 performs a vector search for technical information similar to the inquiry information. That is, the processing unit 71 extracts technical information having a vector similar to a vector representing the characteristics of the inquiry information. The processing unit 71 may extract technical information related to the customer who made the inquiry using a customer ID as a key. The processing unit 71 can identify the customer ID from the CTI information.

[0035] The processing unit 71 may also be configured to extract keywords from the inquiry information by generating the following prompt and inputting it into the language model 72b, and to extract technical information using the extracted keywords: Prompt: "Please extract keywords in bullet points from <text of inquiry information>."

[0036] Next, the processing unit 71 generates a prompt based on the acquired and extracted inquiry information and technical information (step S13).

[0037] For example, the processing unit 71 generates a prompt including inquiry information and technical information. For example, the processing unit 71 may generate the following prompt:

[0038] "For <inquiry information text>, please generate an answer by referring to <technical information text>."

[0039] The "text of inquiry information" is information obtained by speech recognition and converting a user's inquiry into text, such as, for example, "Please tell me about the design concept, target resin, and usage history of part ABC that I purchased this time" in the case of a product inquiry. Of course, information added sequentially in a dialogue, such as, "Please tell me about part ABC that I purchased this time," or "Please tell me about the design concept, target resin, and usage history," can also be used as inquiry text. For inquiries about malfunctions, the text is information obtained by speech recognition and converting a user's inquiry into text, such as, "When I performed task A, the sensor value changed from B to C, and an alert occurred. Please tell me how to resolve this alert." or "I received an alert from the molding equipment website. What action is required?" The "text of technical information" is document documents such as the specifications and manuals for model A.

[0040] The processing unit 71 may also generate a prompt such as the following: "You are a professional in injection molding machines. Please respond to the following customer inquiry by referring to the search results for technical information. #Customer inquiry: <Text of inquiry information> #Search results for technical information: <Text of technical information>"

[0041] Furthermore, a prompt may be created that describes an instruction to create a summary of technical information related to the inquiry information, or an instruction to extract technical information related to the inquiry information.

[0042] Furthermore, the processing unit 71 may generate a prompt by Few-Shot prompting. That is, the processing unit 71 may generate a prompt including an example answer to a query. The processing unit 71 may also generate a prompt requesting multiple answer candidates, such as "Please provide three answers in descending order of priority."

[0043] Furthermore, the processing unit 71 may include in the prompt a phrase that improves the accuracy of the response support information, such as "step by step."

[0044] The information processing device 7 may be configured to be able to change the content of the prompt generated. For example, the information processing device 7 may receive a prompt created by free text via the display operation terminal 5. The information processing device 7 may also present templates for wording to be used in the prompt and receive selection of the wording to be used via the display operation terminal 5. By operating the display operation terminal 5, the sales representative can freely customize the content of the prompt generated by the information processing device 7 and adjust it so that desired response support information is obtained.

[0045] The processing unit 71 then obtains response support information by providing the prompt generated in step S13 to the language model 72b (step S14). For the above-mentioned product-related inquiry, for example, information such as "The design concept, target resin, and usage history of part ABC are XXX" can be obtained. For the above-mentioned defect consultation, information such as "Please operate according to the following troubleshooting procedure. Operating procedure YYY. If the alert is not resolved, part ZZZ must be replaced" and "For the alert from the molding device-specific website, please try performing the procedure Y" can be obtained.

[0046] The processing unit 71 determines whether or not a request for response support information has been made by the sales representative (step S15). The sales representative can perform a request operation for response support information by operating the display operation terminal 5. The display operation terminal 5 on which the request operation has been made transmits a response history ID for identifying the call currently being handled and request data requesting response support information to the information processing device 7. The information processing device 7 determines whether or not a request for response support information has been made based on whether or not the request data has been received.

[0047] The form of the request operation for response support information is not particularly limited, and the processing unit 71 may be configured to receive a request for response support information via a known interface.

[0048] If it is determined that there is no request for response support information (step S15: NO), the processing unit 71 returns the process to step S11. If it is determined that there is a request for response support information (step S15: YES), the processing unit 71 identifies the response support information acquired based on the inquiry information associated with the response history ID, and transmits the identified response support information to the requesting display operation terminal 5 (step S16).

[0049] In the above example, response support information is created at an appropriate time and then sent when requested by a sales representative. However, the system may be configured to create response support information when requested.

[0050] The information processing device 7 and the like according to the first embodiment can assist in responding to inquiries about molding machines and improve user satisfaction. Specifically, the information processing device 7 extracts technical information related to the inquiry information from the user from the technical information DB 72e and provides a prompt including the inquiry information and the extracted technical information to the language model 72b, thereby acquiring response support information. The information processing device 7 can transmit the response support information in response to a request from a sales representative responding to the user's inquiry. The sales representative can utilize the response support information to respond to the inquiry.

[0051] (Embodiment 2) The inquiry response support system according to embodiment 2 differs from embodiment 1 in that it can provide sales representatives with more specific response support information that takes into account information such as defect information and operation information of the molding device related to the inquiry, as well as external public web information. The other configurations of the inquiry response support system are the same as those of the inquiry response support system according to embodiment 1, and therefore similar parts are designated by the same reference numerals and detailed descriptions will be omitted.

[0052] 4 is a schematic diagram showing an example of the configuration of an inquiry support system according to embodiment 2. The inquiry support system according to embodiment 2 includes a contact center system 10, a user telephone 1, a molding device maintenance system 8, and an external web server 9.

[0053] The external web server 9 publishes all kinds of information as web information. The published web information may include the latest technical information on molding devices, information on the operation of molding devices, malfunctions of molding devices, and information on the distribution of parts for molding devices. The information processing device 7 can access the external web server 9 and obtain information related to the inquiry information.

[0054] The maintenance system 8 includes a defect information DB 81, a test information DB 82, a production information DB 83, and an operation information DB 84, which store various information for maintaining and managing the molding apparatus. The information stored in each database is data provided or uploaded to the maintenance system 8 by users. The maintenance system 8 is connected to a contact center system 10 so that the information processing device 7 can access the various databases of the maintenance system 8. The information processing device 7 is configured to be able to access the various databases of the maintenance system 8 and extract information related to inquiry information from users.

[0055] 5A to 5D are schematic diagrams showing examples of database configurations according to embodiment 2. 5A to 5D are conceptual diagrams showing record layouts of a defect information DB 81, a test information DB 82, a production information DB 83, and an operation information DB 84, respectively.

[0056] The defect information DB 81 stores, for example, a customer ID, the model of the molding device, the date and time when the defect occurred, the defect details, the cause of the defect, and a solution to the defect, all in association with each other. The defect details include, for example, the type of after-sales service or complaint. The defect details also include information on mechanical quality, electrical quality, mechanical and electrical quality, and delivery date as information categories. The defect details also include the response status and construction classification.

[0057] The test information DB 82 stores, in association with each other, for example, a customer ID, the model of the molding device, the date and time when the molding device was test run, the test run conditions, and the test run results. The test run conditions include manufacturing conditions such as cylinder temperature settings, the type of raw material, the raw material input locations, and if multiple materials are input, the feed ratios, and detailed information about the raw materials. The test run results include various actual measurements detected by sensors installed in the molding device, product status, and other information.

[0058] The production information DB 83 stores, for example, a customer ID, the model of the molding device, the date and time when the product was manufactured using the molding device, operation information at that date and time, and the yield, all in association with each other. The operation information includes various setting conditions set in the molding device, various actual measurements detected by sensors provided in the molding device, and the maximum and minimum values ​​of the detected actual measurements. The operation information may also be configured to include the type of resin used as the raw material and the grade of the resin.

[0059] The operation information DB 84 stores, in association with each other, for example, a customer ID, the model of the molding device, a date and time, sensor values ​​obtained at that date and time from various sensors provided to detect the state of the molding device, and the state of the molding device when the sensor values ​​were obtained. The sensors are, for example, temperature sensors, pressure sensors, etc. provided at multiple locations along the length of the cylinder that constitutes the molding device. The state of the molding device is obtained, for example, by the maintenance system 8 determining the state based on the sensor values.

[0060] 6 is a flowchart showing the processing procedure of the information processing device 7 according to embodiment 2. As in embodiment 1, the processing unit 71 of the information processing device 7 acquires inquiry information from the call history DB 72 d or the storage unit 72 (step S21), and extracts technical information related to the inquiry information from the technical information DB 72 e (step S22).

[0061] The processing unit 71 also accesses the maintenance system 8 and extracts defect information, test information, production information, and operation information related to the inquiry information (step S23). For example, the processing unit 71 extracts a customer ID from the customer DB 72c using the incoming telephone number included in the CTI information as a key. The processing unit 71 also searches for and extracts characters corresponding to the model from the inquiry information. The processing unit 71 can extract defect information, test information, production information, and operation information related to the customer who made the inquiry and the model from various databases of the maintenance system 8 using the customer ID and information related to the model as a key. Note that, similar to the technical information, the processing unit 71 may be configured to extract defect information, test information, production information, and operation information related to the inquiry information from the databases of the maintenance system 8 by vector search.

[0062] Furthermore, the processing unit 71 accesses the external web server 9 and extracts public web information related to the inquiry information (step S24). Note that the processing unit 71 may use a web crawler to collect public web page information related to the molding device, and extract the public web information related to the inquiry information from the collected public web pages.

[0063] The processing unit 71 then generates a prompt based on the acquired and extracted inquiry information, technical information, defect information, test information, production information, operation information, and public web information (step S25). The processing unit 71 then provides the prompt generated in step S25 to the language model 72b to acquire response support information (step S26). Thereafter, similar to steps S15 and S16 in the first embodiment, when a request for response support information is made, the processing unit 71 identifies the acquired response support information based on the inquiry information associated with the response history ID, and transmits the identified response support information to the requesting display operation terminal 5 (steps S27 and S28).

[0064] According to the information processing device 7 etc. of this embodiment 2, by providing the sales representative with more specific information about the user's molding machine, such as defect information, test information, production information, and operation information, as well as response support information that takes into account external public web information, it is possible to support inquiries and improve user satisfaction.

[0065] Although the maintenance system 8 has been described as including the defect information DB 81, the test information DB 82, the production information DB 83, and the operation information DB 84, the storage locations of the defect information, test information, production information, and operation information are not particularly limited. The above-mentioned various pieces of information may be uploaded by the user to the contact center system 10 or the information processing device 7 and stored in each device. The above-mentioned information may be stored in a predetermined folder in a database that the customer normally uses, and the information processing device 7 may be configured to access the information in the predetermined folder. Furthermore, if another external system stores the above-mentioned information, the information processing device 7 may be configured to search and acquire the information using an API or the like. To avoid storing such confidential information for a long period of time in the contact center system 10, cloud, or the like, a deletion function over time may be optionally selected.

[0066] Furthermore, the database of the security system 8 may be configured to separate, for each customer ID, information that can be used to create response support information from information that cannot be used. The information processing device 7 identifies the customer ID from the CTI information, extracts information related to the inquiry information from information that can be accessed using the sender's customer ID, and generates a prompt. This configuration makes it possible to create individual, specific response support information tailored to each user while managing confidential information.

[0067] Furthermore, a tag indicating the type of extruder or injection molding machine may be added to the information stored in the database of the maintenance system 8. The information processing device 7 can efficiently search and extract information related to the inquiry information by using the tag.

[0068] (Embodiment 3) The inquiry response support system according to embodiment 3 differs from embodiment 2 in that it can accept a selection of information related to the molding device related to the inquiry and provide response support information that takes the selected information into account to the sales representative. Since the other configurations of the inquiry response support system are the same as those of the inquiry response support system according to embodiment 2, the same reference numerals are used for the same parts and detailed description will be omitted.

[0069] 7 is a flowchart showing the processing procedure of the information processing device 7 according to the third embodiment. As in the first embodiment, the processing unit 71 of the information processing device 7 acquires inquiry information from the call history DB 72d or the storage unit 72 (step S31). Then, the processing unit 71 accepts, via the display operation terminal 5, a selection of information related to the inquiry information from the user among the technical information stored in the technical information DB 72e and the information stored in the database of the maintenance system 8 (step S32). For example, the processing unit 71 may accept the selection of information related to the inquiry information using a check box or a pull-down list on the application screen.

[0070] Next, the processing unit 71 extracts the technical information selected in step S32 from the technical information stored in the technical information DB 72e (step S33). The processing unit 71 also extracts the information selected in step S32 from the defect information, test information, production information, and operation information stored in the maintenance system 8 (step S34). Furthermore, the processing unit 71 accesses the external web server 9 and extracts public web information related to the inquiry information (step S35).

[0071] The processing unit 71 generates a prompt based on the inquiry information and the selectively extracted information (step S36).The processing unit 71 then provides the prompt generated in step S36 to the language model 72b to obtain response support information (step S37).Thereafter, similar to steps S27 and S28 in the second embodiment, when a request for response support information is made, the processing unit 71 identifies the obtained response support information based on the inquiry information associated with the response history ID and transmits the identified response support information to the requesting display operation terminal 5 (steps S38 and S39).

[0072] According to the information processing device 7 etc. of this embodiment 3, the user can select relevant information from more specific information about the molding machine, and by providing the sales representative with response support information that takes into account the selected information, it is possible to support inquiries and improve user satisfaction.

[0073] In the third embodiment, an example has been described in which a sales representative operates the display operation terminal 5 to select information related to inquiry information, but the information to be used to create response support information may be determined by a data structure. For example, in the database of the maintenance system 8, a folder directory may be used to distinguish between information to be used to create response support information and information that will not be used.

[0074] (Embodiment 4) An inquiry response support system according to embodiment 4 differs from embodiments 1 to 3 in that reply support information is generated using a plurality of language models 72b. Other configurations of the inquiry response support system are the same as those of the inquiry response support systems according to embodiments 1 to 3, and therefore similar parts are denoted by the same reference numerals and detailed description thereof will be omitted.

[0075] 8 is a block diagram showing an example of the configuration of an information processing device 7 according to embodiment 4. The information processing device 7 according to embodiment 4 stores, in a storage unit 72, a first language model 721a that summarizes a sentence and a second language model 722a that is different from the first language model 721a.

[0076] 9 is a flowchart showing the processing procedure of the information processing device 7 according to the fourth embodiment. The processing unit 71 of the information processing device 7 acquires inquiry information from the call history DB 72d or the storage unit 72 (step S41), and summarizes the inquiry information by providing a prompt including the inquiry information to the first language model 721a (step S42). Hereinafter, the summarized inquiry information will be referred to as summarized inquiry information.

[0077] Next, the processing unit 71 extracts technical information related to the summary inquiry information from the technical information DB 72e (step S43). For example, the processing unit 71 performs a vector search for technical information similar to the summary inquiry information. That is, the processing unit 71 extracts technical information having vectors similar to the vectors representing the characteristics of the summary inquiry information.

[0078] Next, the processing unit 71 generates a prompt based on the summary inquiry information and the technical information (step S44).Then, the processing unit 71 provides the prompt generated in step S44 to the second language model 722a to obtain response support information (step S45).

[0079] Hereinafter, similarly to steps S15 to S16 in embodiment 1, when a request for response support information is made, the processing unit 71 identifies the response support information acquired based on the inquiry information associated with the response history ID, and transmits the identified response support information to the display operation terminal 5 that made the request (steps S46 and S47).

[0080] According to the information processing device 7 etc. of this embodiment 4, even if the content of an inquiry from a user is redundant, the inquiry information can be summarized to generate response support information, thereby supporting the response to inquiries regarding the molding machine.

[0081] In the fourth embodiment, an example has been described in which the first language model 721a and the second language model 722a are used in series to create response support information. However, a configuration may also be adopted in which a sales representative can select a desired model from among multiple language models depending on the intended use. For example, the information processing device 7 is configured to store multiple language models, such as a language model for summarizing, a language model for extracting related information, and a language model for translating. The information processing device 7 then displays a list of multiple language models on the display operation terminal 5 and accepts the selection of a language model to be used. The information processing device 7 may also accept a selection of a language model's intended use, such as summarizing, requirement extraction, or translation, via the display operation terminal 5, and select a language model corresponding to the intended use. Furthermore, the information processing device 7 may also be configured to accept a selection of a language model's intended use, such as summarizing, requirement extraction, or translation, via the display operation terminal 5, display a list of multiple language models according to the intended use, and accept the selection of a language model. The information processing device 7 obtains response support information by providing a prompt including inquiry information and related technical information to the selected language model.

[0082] Furthermore, although the fourth embodiment describes a case in which multiple language models are used in series, multiple language models may be used in parallel, or multiple language models may be used in a series-parallel configuration. In a case in which multiple language models are used in parallel, a configuration is preferable in which, for example, the same prompt is given to each of multiple language models to obtain multiple pieces of answer support information, and answer support information having common content is transmitted to the display operation terminal 5 of the sales representative. Note that, when the content of multiple pieces of answer support information is a true / false judgment, a more likely judgment result may be identified by majority vote, and the identified judgment result may be transmitted to the display operation terminal 5 of the sales representative as answer support information. In a case in which multiple language models are used in a series-parallel configuration, a configuration is preferable in which, for example, the same prompt is given to each of multiple first-stage language models to obtain multiple pieces of answer support information, and the multiple pieces of answer support information are transmitted to the second-stage language model for summarization.

[0083] 1: User telephone 2: Public telephone line 4: Sales representative telephone 5: Display operation terminal 6: In-house communication network 7: Information processing device 8: Maintenance system 9: External web server 10: Contact center system 70: Recording medium 71: Processing unit 72: Storage unit 72a: Computer program 72b: Language model 73: Communication unit 721a: First language model 722a: Second language model 72c: Customer DB 72d: Call history DB 72e: Technical information DB 81: Defect information DB 82: Test information DB 83: Production information DB 84: Operation information DB

Claims

1. An information processing method that acquires inquiry information about a molding machine from a user, and acquires response support information that contributes to answering the inquiry using the acquired inquiry information and a language model that uses a database that stores information about the molding machine.

2. The information processing method according to claim 1, wherein the acquired reply support information is sent to a sales representative who handles the user's inquiry.

3. An information processing method according to claim 1 or claim 2, wherein the database stores technical information related to the molding machine, and based on the acquired inquiry information, the technical information related to the inquiry information is extracted from the database, and the extracted technical information and a prompt based on the inquiry information are provided to the language model, thereby acquiring the response support information.

4. An information processing method according to any one of claims 1 to 3, wherein the database stores at least one of past defect information, test information, and production information for the molding machine, and based on the acquired inquiry information, at least one of the defect information, test information, and production information related to the inquiry information is extracted from the database, and a prompt based on the extracted information and the inquiry information is provided to the language model, thereby acquiring the response support information.

5. An information processing method according to any one of claims 1 to 4, wherein the database stores operation information of the molding machine, and based on the acquired inquiry information, the operation information related to the inquiry information is extracted from the database, and a prompt based on the extracted operation information and the inquiry information is provided to the language model, thereby acquiring the response support information.

6. An information processing method according to any one of claims 1 to 5, wherein external public web information is extracted based on the acquired inquiry information, and the response support information is acquired by providing the extracted public web information and a prompt based on the inquiry information to the language model.

7. An information processing method according to any one of claims 1 to 6, comprising: accepting a selection of data related to the inquiry information; and providing a prompt based on the selected data and the inquiry information to the language model, thereby obtaining the response support information.

8. The information processing method according to any one of claims 1 to 7, wherein the response support information related to a response to an inquiry from the user is acquired using a plurality of types of language models.

9. An information processing method according to any one of claims 1 to 8, wherein the multiple types of language models include a first language model that summarizes a sentence and a second language model different from the first language model, wherein acquired inquiry information is provided to the first language model to obtain summary information of the inquiry information, and wherein a prompt based on the acquired summary information and information stored in the database is provided to the second language model to obtain the response support information related to the response to the inquiry from the user.

10. An information processing device having a processing unit for executing a process of acquiring inquiry information regarding a molding machine from a user, and acquiring response support information that contributes to answering the inquiry using the acquired inquiry information and a language model that uses a database that stores information regarding the molding machine.

11. A computer program for causing a computer to execute a process of acquiring inquiry information about a molding machine from a user, and acquiring response support information that contributes to answering the inquiry, using the acquired inquiry information and a language model that utilizes a database that stores information about the molding machine.

Images