Ai income
The AI-based intercom system addresses the challenge of inadequate customer service in retail stores by using an AI-based intercom system that connects multiple client terminals through an information communication network, allowing voice input of questions and transmitting audio data between servers to provide accurate answers, utilizing machine learning for improved response extraction and provision.
Patent Information
- Application Number
- JP2025033868
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-03-04
- Publication Date
- 2025-05-13
AI Technical Summary
Retail stores face challenges in providing adequate customer service due to a decline in staff numbers and the difficulty in accessing product information, leading to frustrated customers and increased staff turnover.
An AI-based intercom system that connects multiple client terminals through an information communication network, allowing voice input of questions and transmitting audio data between servers to provide accurate answers, utilizing machine learning for improved response extraction and provision.
The system enables efficient answering of customer inquiries, reduces staff training time, and enhances customer service quality, leading to increased customer satisfaction and reduced staff turnover.
Smart Images

Figure 2025074287000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an intercom system having an automatic response function using AI. [Background technology]
[0002] Recently, while retail stores are becoming larger, the number of sales clerks in the stores is decreasing, and customers who want to purchase products have a problem of not being able to find a sales clerk to serve them, even if they visit the store. Also, even if you find a sales clerk in the store and ask questions about the product, it does not mean that all of your problems will be solved. This is because many retail stores are operated under a system where different sales staff are assigned to different sales areas for each product group, and sales clerks and staff cannot answer questions about sales areas and products other than those they are in charge of. This problem is particularly noticeable in stores with large sales floors and large-scale retail stores.
[0003] As shown in Figure 8, retail staff training consists of (1) training that should be given before they start working in the store, and (2) training that should continue even after they start working in the store. In other words, before working in the store, staff must learn the basics, such as "products," "sales," "accounting," "company basics," and "customer service," before they can step out onto the counter. Depending on the industry, other knowledge, such as delivery, must also be learned in advance.
[0004] On the other hand, even after they are in the store, they need to supplement their product knowledge with new products that are released on a regular basis, such as every week or month. When a product is featured on television or social media, they need to remember the connection between the program and the product featured, and when there are changes to in-store operations, they need to remember those as well.
[0005] In other words, the staff and clerks working on the sales floor have a lot to remember and are overwhelmed, and if they cannot answer questions or inquiries from customers, they will end up upsetting them and receiving complaints. Since it is difficult to remember all product information in the first place, one method is to search for information on an information terminal, but there is also the problem that it is difficult to find the information you need even if you search.
[0006] Retail business managers face the problem that, in addition to the difficulty of hiring staff and salespeople, they must bear the costs of training them, and if they do not improve the situation of having few or no staff and salespeople on the sales floor, customers will leave their companies. Furthermore, because customer service is largely dependent on the experience and ability of individuals, raising the level of customer service throughout the chain is an extremely difficult task.
[0007] As a management issue, if the burden on staff and clerks is too great, they may quit. If staff quits, the company may not be able to recover the expenses it has spent on training, which may result in losses. [Prior art documents] [Patent documents]
[0008] [Patent Document 1] International Publication WO2020 / 116531 Summary of the Invention [Problem to be solved by the invention]
[0009] An object of the present invention is to provide an intercom system that can be used in retail stores in order to solve the above-mentioned problems. [Means for solving the problem]
[0010] This problem is solved by an intercom system consisting of a plurality of client terminals and an information processing server connected to each other through an information and communication network, characterized in that the intercom system has a client terminal that enables voice input of question information, a first information processing server that receives the input question voice data and enables it to be transmitted as voice data to other client terminals and / or a second information processing server, and a second information processing server that receives the voice data from the first information processing server and, in cooperation with a third information processing server that accumulates answer information, enables the extraction and provision of answer information, and transmits the extracted and available answer information to the first information processing server as voice data.
[0011] In an advantageous development, the second information processing server may have a machine learning unit, and may extract and provide answer information to question information by machine learning.
[0012] In a further developed version, a fourth information processing server is provided, which is connected to the third information processing server via an information and communication network, and advance information is stored in this fourth information processing server, and the advance information can be made transmittable to the third information server. [Brief description of the drawings]
[0013] [Figure 1] Overall view of the system according to the present invention [Diagram 2] A diagram showing the hardware configuration of each element constituting the present invention. [Diagram 3] FIG. 2 is a diagram showing an example of a software configuration of a first information processing server. [Figure 4] FIG. 1 is a diagram showing an example of a software configuration of a second information processing server. [Diagram 5] FIG. 13 is a diagram showing an example of a software configuration of a third information processing server. [Figure 6] FIG. 13 is a diagram showing an example of a software configuration of a fourth information processing server. [Figure 7] FIG. 1 is a diagram showing an example of a software configuration of a client terminal. [Figure 8] A diagram showing an example of the training that retail store staff should receive DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0014] FIG. 1 shows an overall view of an intercom system using AI according to the present invention (hereinafter referred to as an AI intercom system). As can be seen in FIG. 1, the AI intercom system according to the present invention has multiple client terminals and first to fourth information processing servers. Each client terminal and the information processing server are connected wired and / or wirelessly via a general information communication network (such as the Internet communication network). The hardware and software configurations and functions of the client terminals and each information processing server are described below.
[0015] [Client terminal hardware configuration] As shown in Fig. 1, the AI intercom system according to the present invention includes multiple client terminals. These client terminals are intended for use by users of the AI intercom system, primarily store clerks at retail stores. The client terminals can be, for example, information and communication terminals such as smartphones, tablet terminals, and personal computers.
[0016] As shown in Figure 2, the client terminal consists of a CPU for information processing, an SSD / HDD as a storage device, RAM as temporary storage memory, ROM as non-volatile storage memory, an input device, an output device (display device), an external interface, and a communication interface.
[0017] The CPU is responsible for reading programs and data from storage devices such as ROM and SSD / HDD and executing the processes. It is a computing device that controls the entire terminal (computer) shown in Figure 2 and realizes the functions of a computer or smartphone. The program that realizes the intercom function according to the present invention is executed by the CPU.
[0018] An SSD / HDD is a non-volatile storage device that stores programs and data. The programs stored include the OS (operating system), which is the basic software, and application software that realizes various functions on the OS. A program for realizing the intercom function of the present invention is stored in the SSD / HDD. Note that the SSD / HDD is merely an example of a storage device, and does not prevent the adoption of alternative technologies to replace it.
[0019] RAM is a volatile semiconductor memory (storage device) that temporarily stores programs and data. The program for realizing the income function of the present invention is read from an SSD / HDD, etc., and temporarily written into the RAM, which is then executed by the CPU and used to temporarily store data required for the execution of the program of the present invention.
[0020] ROM is a non-volatile semiconductor memory (storage device) that can retain programs and data even when the power is turned off for computers that realize server terminals, client terminals, etc. ROM can store the BIOS, OS settings, various programs, data, etc. that are executed when the computer in Figure 2 starts up.
[0021] There are two types of input devices: a voice input device (microphone) and an instruction input device. This is a device for inputting instructions from a user to the computer / smartphone in FIG. 2 or for enabling input of the user's voice. Examples of input devices for inputting instructions include a keyboard, a mouse, a controller, and a touch panel. When implementing the program according to the present invention, the input device has a function for enabling input of various instructions from the user. The voice input device can be incorporated in the device in FIG. 2 (a small computer, a smartphone, etc.), or can be an external input device configured as a separate device such as a headset that can be connected to the computer / smartphone by wire or wirelessly.
[0022] There are two types of output devices: audio output devices (speakers) and display devices. Display devices are so-called displays or monitors. When the computer in FIG. 2 is realized as a notebook PC or desktop PC, it can be formed as a display for display output, and when it is realized as a tablet terminal, smartphone, or the like, the display device can be formed as, for example, a liquid crystal display or an organic EL display integrated with the input device. The audio output device (speaker) can be built into the computer or smartphone in FIG. 2, or can be an external output device configured as a separate device, such as a headset that can be connected to the computer or smartphone via wired or wireless connection.
[0023] An external interface is an interface for connecting to an external device. An example of an external device is a storage medium. In addition to the common SSD / HDD, memory cards and USB memory are also possible storage media. Such storage media are used when the capacity of the storage media inside the computer is insufficient, or when exchanging programs or data with the outside.
[0024] The communication interface is an interface for connecting to an information and communication network (such as the Internet) in the AI intercom system according to the present invention. The computer, smartphone, etc. in FIG. 2 are configured to be able to perform data communication and information communication via this communication interface.
[0025] [Hardware configuration of information processing server] As shown in Fig. 1, the AI intercom system according to the present invention includes multiple information processing servers. In the embodiment of Fig. 1, first to fourth information processing servers are shown. Each information processing server can be implemented as a computer capable of performing currently available general information processing and information communication, and can have the same or similar configuration as the above-mentioned client terminal, for example, the configuration shown in Fig. 2.
[0026] [Functions and roles of each information processing server] 1, the first information processing server has a role of realizing a so-called intercom function in cooperation with the client terminals. For this reason, the first information processing server is connected to a plurality of client terminals via an information and communication network, and is configured to receive voice data input through a client terminal and transmit the data to another client terminal.
[0027] In addition to the functions described above, the first information processing server has the function and role of passing (transmitting) the information and voice data (inquiry data) received from each client terminal to the second information processing server, and in response, transmitting the information and voice data (answer data) received from the second information processing server to the appropriate client terminal.
[0028] The second information processing server has the function and role of working as an answer extraction and provision server that cooperates with the first information processing server to extract answers requested by each client terminal and provide them to each client terminal. The second information processing server is configured to extract and provide answers using so-called AI machine learning. In addition, the second information processing server can also have a function of converting voice data (conversation data) received from the first information processing server into text and storing it. The storage can be performed in the second information processing server and / or in a third information processing server described below.
[0029] The third information processing server has the function and role of an information accumulation server. The second information processing server (answer extraction providing server) described above is connected to this information accumulation server (third information processing server) and is configured to be able to extract appropriate answers from the information accumulated in the third information processing server. The information to be accumulated in the third information processing server (information accumulation server) is configured to be able to be uploaded from the fourth information processing server described below.
[0030] The fourth information processing server has a function as a pre-information accumulation server. It is a server for pre-saving and storing information to be accumulated in the above-mentioned third information processing server, i.e., the information accumulation server. This can include, for example, manuals, product specifications, and other information as text information. It can also include inventory information, new product information, and other information as data information.
[0031] [Software configuration of each information processing server] Next, the software configuration of the first information processing server will be described with reference to Fig. 3. As described above, the first information processing server has a role of realizing a so-called intercom function in cooperation with a client terminal. For this reason, the first information processing server has a voice information transmitting / receiving unit for receiving voice information (voice data) from a client terminal and transmitting the received voice information (voice data) to another client terminal.
[0032] The first information processing server also has an ID information transmitting / receiving unit. The ID information transmitting / receiving unit is configured to be able to receive ID information transmitted from a client terminal. As will be described later, the ID information includes information specific to each client terminal, and is useful for determining which client terminal transmitted the above-mentioned voice information (voice data).
[0033] The first information processing server further has a call channel forming unit. The call channel formed by the call channel forming unit is configured to allow only users belonging to a specific group to join. As a result, the intercom function realized by the first information processing server is configured to be able to expand and reduce the area in which calls can be made in stages and hierarchically, such as calls only within the store, calls within a block, calls within an area, and nationwide calls. It is of course also conceivable to form a call channel that allows calls only to specific people within the store.
[0034] A call content determination unit is provided in the first information processing server. The call content determination unit is provided to determine and process the content of the voice information transmitted from the client terminal and received by the voice information transmission / reception unit. For example, the call content determination unit is configured to be able to check whether a specific voice string (character string) is present in the voice information. For example, when a user using a client terminal utters a specific keyword, "AI intercom," the call content determination unit can be configured to issue an instruction to the call channel formation unit. In this case, the call channel formation unit forms a channel dedicated to AI intercom, which will be described later, in accordance with the instruction and assigns the user to the channel.
[0035] It is also possible to configure the system so that multiple keywords can be set as keywords to be judged and processed by the call content judgement unit. For example, it is possible to set "question," "thank you," etc. as keywords. In this case, it is possible to set the system so that a specific action is performed when the user utters one of these keywords during the above-mentioned AI intercom dedicated channel. The specific action will be described later.
[0036] The first information processing server is connected to a second information processing server, which will be described later, via an information communication network, and is configured to be able to transmit voice information (voice data) to the second information processing server through a voice information transmission / reception unit.
[0037] Next, the second information processing server will be described with reference to Fig. 4. As shown in Fig. 4, the second information processing server has a voice information receiving unit, an answer extracting unit, a voice information transmitting unit, a machine learning unit, and a learning information transmitting unit. As can be seen in Fig. 4, the second information processing server is connected to the first information processing server and the third information processing server via an information and communication network.
[0038] The voice information receiving unit of the second information processing server is configured to receive voice information (voice data) from the first information processing server. The voice information receiving unit of the second information processing server is configured to transfer the received voice information to an answer extracting unit also provided in the second information processing server. The answer extracting unit is configured to determine the content of the received voice information. If the determined voice information includes a question, the answer extracting unit of the second information processing server crawls the text information in a third information processing server described later and searches the text information for suitable information as an answer to the question. When suitable information as an answer is found, the answer extracting unit extracts it as an answer, converts it into voice data, and transfers it to the voice information transmitting unit. The voice information transmitting unit transmits the received voice data to the first information processing server, and the first information processing server transmits the voice data to an appropriate client terminal. The first information processing server transmits the voice information (answer information) to the client terminal through the voice information transmitting and receiving unit of the first information processing server.
[0039] The second information processing server further has a machine learning unit. The machine learning unit is connected to the answer extraction unit in the second information processing server, and is configured to perform machine learning according to the manner of answer extraction by the answer extraction unit. For this reason, the machine learning unit can be configured to have a function of converting voice data (conversation data) into text. The textual information is configured to be delivered to the learning information transmission unit as learning information. The learning information transmission unit is configured to be able to transmit the received textual learning information to the third information processing server. In the third information processing server, such textual learning information, particularly information regarding cases that could not be answered, is configured to be able to be listed and stored.
[0040] Next, the third information processing server will be described with reference to Fig. 5. As described above and as can be seen in Fig. 5, the third information processing server has the function and role of an information accumulation server. The second information processing server crawls the third information processing server and is configured to be able to search for necessary answer information, thereby realizing the AI income function according to the present invention.
[0041] As can be seen in Fig. 5, the third information processing server has a prior information receiving unit, an API linking unit, a learning information receiving unit, an information processing unit, an information storage unit, and an information transmitting unit. As can be seen in Fig. 5, the third information processing server is connected to the second information processing server and the fourth information processing server in a wired and wireless manner via an information communication network.
[0042] The advance information receiving unit is configured to be able to receive advance information stored in a fourth information processing server described below. The advance information stored in the fourth information processing server is so-called text information such as product manuals, product specifications, and other information. Having received the advance information from the fourth information processing server, the advance information receiving unit passes the advance information to the information storage unit, and the information storage unit is configured to store the advance information appropriately.
[0043] The third information processing server includes an API linking unit that can be configured to automatically receive so-called data information, such as product inventory information, new product information, and other information, that exists in the third information processing server by linking with the fourth information processing server.
[0044] The third information processing server further includes a learning information receiving unit configured to receive the learning information generated by the machine learning unit of the second information processing server.
[0045] The third information processing server further includes an information processing unit configured to transfer the information received by the advance information receiving unit, the API linking unit, and the learning information receiving unit to the information storage unit as appropriate and store the information therein.
[0046] The third information processing server further includes an information transmission unit configured to transmit information in the information storage unit as required in response to a request from the second information processing server.
[0047] The fourth information processing server, which is the last information processing server, will be described with reference to FIG. 6. As described above, the fourth information processing server has the function of an advance information accumulation server. It is a server for preserving and storing information to be accumulated in the above-mentioned third information processing server, i.e., the information accumulation server. For this reason, the fourth information processing server is provided with an advance information accumulation unit, an advance information transmission unit, and an API linking unit.
[0048] The information stored in the advance information storage unit may include, as text information, manuals, product specifications, and other information, and may also include, as data information, inventory information, new product information, and other information.
[0049] The fourth information processing server further includes a prior information transmitting unit configured to transmit the prior information stored in the prior information storing unit, more specifically, so-called text information such as product manuals, product specifications, and other information, to the prior information receiving unit of the third information processing server.
[0050] The fourth information processing server further includes an API linking unit configured to automatically transmit the advance information stored in the advance information storage unit, more specifically, so-called data information such as product inventory information, new product information, and other information, to the API linking unit of the third information processing server.
[0051] Finally, the client terminal in the AI intercom system according to the present invention will be described with reference to Fig. 7. A plurality of client terminals are provided in the AI intercom system according to the present invention, and each client terminal is configured to be able to communicate with each other via the first information processing server as described above. For this reason, each client terminal has a voice information input unit, a voice information transmitting unit, a voice information receiving unit, and a voice information output unit. Furthermore, each client terminal has an ID information generating unit.
[0052] The voice information input unit is a part that realizes inputting conversation, question, and answer information (signals) issued by a user as voice data in cooperation with an input device (voice input device, microphone) of the client terminal. ID information generated by an ID information generation unit is added to the voice data input by the voice information input unit. The voice data with the ID information added is passed from the voice information input unit to the voice information transmission unit. The voice information transmission unit transmits the voice data to the voice information transmission / reception unit of the first information processing server and the ID information to the ID information transmission / reception unit via an information and communications network.
[0053] On the other hand, each client terminal has a voice information receiving unit so as to be able to receive voice information and voice data from the first information processing server. This voice information receiving unit receives voice data sent from the first information processing server via an information and communication network. The voice information receiving unit passes the received voice data to a voice information output unit. The voice information output unit is configured to output the voice data as an acoustic signal that can be heard by the user through the output device in cooperation with the output device of the client terminal.
[0054] [System-wide operation] The operation of the entire system according to the present invention will be described with reference to FIG. 1. Three store staff members are illustrated in FIG. 1 as an example. Among the store staff members, the female staff member illustrated at the bottom is serving customers. Suppose that the store staff member serving customers receives an inquiry or question from a customer, such as, "Can I use this plate in the oven?" The staff member serving customers does not have an answer to this question. Therefore, the staff member serving customers inputs the question as voice data through the client terminal he or she owns. The input voice data (question data) is sent to the first information processing server. When the call content determination unit in the first information processing server confirms that the voice data is question data, the first information processing server transmits the voice data to the second information processing server. The second information processing server crawls the third information processing server to search for an answer to the question. When an appropriate answer is found, the second information processing server converts the answer information into voice data (answer data) and transmits it as voice information to the first information processing server, which transmits it to the client terminal. This allows staff members serving customers to instantly receive the appropriate response.
[0055] [Other features / intercom functions] Next, individual functions achieved by the entire system according to the present invention will be described. The AI intercom system according to the present invention allows staff at one store to provide answers to questions received by staff at another store, that is, allows staff at other stores to serve customers at other stores through the AI intercom system. As described above, each client terminal is connected to the first information communication server through an information communication network, for example, the Internet communication network, and is configured to be able to transmit and receive voice information (voice data). The received voice information (voice data) can be heard by the user through a voice output device (hardware) and a voice information output unit (software). Since the first information processing server is provided with a channel forming unit, it is possible to connect to staff at other stores by setting an appropriate channel.
[0056] [Other functions: store display function] Furthermore, it is also conceivable that the system according to the present invention is provided with an image output device such as an external display or digital signage as an external output device. It is also conceivable that the system according to the present invention is provided with an audio input device such as a microphone installed in the sales floor as an external input device. The image output device and audio input device can be configured as an output device and an input device connected to, for example, one client terminal. In this case, it is conceivable that the client terminal is provided with an image output unit as a software configuration. It is conceivable that, for example, a sales floor of a specific product in a store can be displayed through the image output device and the image output unit. For this reason, the third information processing server is configured to store such sales floor information as image information in the information storage unit. Through such a configuration, a customer visiting a store can directly input a question such as "Where is the interior sales floor?" into the audio input device (such as a microphone installed in the sales floor), and in response to this, the AI intercom system according to the present invention can display an image that guides the route from the current location to the interior sales floor that has received the question through an image output device such as digital signage. For this purpose, it is necessary to store information such as the sales floor name and section (location) in the third information processing server.
[0057] [Other features: Staff self-learning function] By using the system according to the present invention, store staff can not only answer questions from customers, but also learn by themselves. This is because by inputting not only the information and questions that customers want to know, but also the information and questions that they themselves want to know, appropriate answers to the questions can be obtained. This allows, for example, store staff to (1) organize products in the store while (2) acquiring product knowledge in the store. Conventionally, these were separate tasks. Organizing products was, for example, the task of folding clothes and arranging them for display, and acquiring product knowledge was often done by looking at documents on paper or a PC in the office. However, the system according to the present invention makes it possible to perform these separate tasks simultaneously. There is also an advantage in that the time spent working in the office is reduced, which increases the time spent on the sales floor.
[0058] [Other features: Conversation text conversion function] As mentioned above, all conversations through the AI intercom system of the present invention are converted into text. The converted text information can be used for the following two purposes: (1) to create a list (database) of questions that were entered but for which an appropriate answer was not extracted or provided, in order to prepare for subsequent responses (such as preparing answers by the person in charge), and (2) to use it to extract issues throughout the store. By converting conversations in a store where no one thinks there are any issues into text, it becomes possible to propose solutions to improve the issues at that store.
[0059] [Other functions: Cost reduction by shortening customer service time] By using the AI intercom system of the present invention, it is possible to (1) shorten the time spent serving customers in the store, (2) shorten the time required for pre-training store staff, and (3) shorten the time required for store staff to learn the latest update information. This leads to significant cost reductions for the store as a whole.
[0060] [Other features: Improved response rate to reduce lost opportunities] By using the AI income system of the present invention, the probability of providing appropriate answers to customer questions is greatly improved. This makes it possible to significantly reduce the so-called opportunity loss, which occurs when a customer leaves a store without being able to provide an answer to their question.
[0061] [Other functions: Improving customer service levels at each store] The quality of customer service is often left to the experience and ability of the individual. By using the AI intercom system of this invention, even part-time employees who have just joined the company or started working at the store today will be able to provide a certain level of customer service.
[0062] [Other functions / use of ID information] Furthermore, by utilizing ID information (employee number, etc.), it is possible to identify staff members who use the system according to the present invention. This makes it possible to obtain various qualitative and quantitative information for each individual. In addition, by utilizing ID information, it is possible to visualize the efforts of individuals based on the usage status of the self-learning function mentioned above.
Claims
1. An intercom system comprising a plurality of client terminals and an information processing server connected to each other through an information and communication network, An intercom system characterized by having a client terminal that enables the audio input of question information, a first information processing server that receives the input question audio data and enables it to be transmitted as audio data to other client terminals and / or a second information processing server, and a second information processing server that receives the audio data from the first information processing server and, in cooperation with a third information processing server that accumulates answer information, enables the extraction and provision of answer information and transmits the extracted and available answer information to the first information processing server as audio data.
2. The intercom system according to claim 1, characterized in that the second information processing server has a machine learning unit, and extracts and provides answer information to question information by machine learning.
3. The intercom system according to claim 2, further comprising a fourth information processing server connected to the third information processing server via an information communication network, the fourth information processing server storing advance information, and the advance information being capable of being transmitted to the third information server.
Citation Information
Patent Citations
Conversation analysis system, method, and program
WO2020116531A1