Store support system, store support method, and program

Abstract

Support store operations. [Solution] The store support system (1)'s estimated store performance information acquisition unit (102) acquires estimated store performance information relating to estimated store performance, which is the performance of the estimated store. The input unit (103) inputs the estimated store performance information to an AI (Artificial Intelligence) that has learned the relationship between training store performance information relating to training store performance, which is the performance of the training store, and training store factor information relating to training store factors, which are the factors of said training store performance. The estimated store factor information acquisition unit (105) acquires estimated store factor information relating to estimated store factors, which are the factors of estimated store performance, and is output from the AI ​​that has been input with estimated store performance information. The support unit (106) supports the operations of the estimated store based on the estimated store factor information.

Description

Technical Field

[0001] The present disclosure relates to a store support system, a store support method, and a program.

Background Art

[0002] Conventionally, technologies for supporting store operations have been known. For example, in Patent Document 1, based on user information regarding users who use a store, as a settlement tendency of the users, the relationship between the discount rate set in a discount coupon and the usage rate at which the user uses the discount coupon according to the discount rate is estimated for each physical store, and by generating a discount coupon in which the discount rate at which the usage rate is equal to or higher than a threshold value is set, it is described that the store's business of generating discount coupons is supported.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, since store operations are diverse, simply generating discount coupons that are estimated to be highly likely to be used by users, as in the technology of Patent Document 1, cannot sufficiently support store operations. This is the same for technologies other than the technology of Patent Document 1. Conventional technologies have not been able to sufficiently support store operations.

[0005] One object of the present disclosure is to support store operations.

Means for Solving the Problems

[0006] The store support system relating to this disclosure includes: an estimated store performance information acquisition unit that acquires estimated store performance information relating to estimated store performance, which is the performance of an estimated store; an input unit that inputs the estimated store performance information to an AI (Artificial Intelligence) that has learned the relationship between training store performance information relating to training store performance, which is the performance of a training store, and training store factor information relating to training store factors, which are the factors of said training store performance; an estimated store factor information acquisition unit that acquires estimated store factor information relating to estimated store factors, which are the factors of said estimated store performance, and which is output from the AI ​​into which the estimated store performance information has been input; and a support unit that supports the operations of the estimated store based on the estimated store factor information. [Effects of the Invention]

[0007] This disclosure can support the operations of stores. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows an example of the hardware configuration of a store support system. [Figure 2] This figure shows an example of a screen displayed on a store terminal. [Figure 3] This figure shows an example of a screen displayed on a store terminal. [Figure 4] This diagram shows an example of the functions implemented in the store support system. [Figure 5] This figure shows an example of a training database. [Figure 6] This figure shows an example of a store database. [Figure 7] This figure shows an example of AI input and output. [Figure 8] This figure shows an example of the process performed by the store support system. [Figure 9] This figure shows an example of a function that can be achieved through modification. [Figure 10] This figure shows an example of the input and output of the AI ​​in Modification Example 1. [Figure 11]This figure shows an example of the input and output of the AI ​​in Modification Example 3. [Figure 12] This figure shows an example of the estimated result screen for Modification Example 3. [Figure 13] This figure shows an example of the input and output of the AI ​​in Modification 4. [Figure 14] This figure shows an example of the input and output of the AI ​​in variation 6. [Figure 15] This figure shows an example of the screen displayed on the store terminal in variation 11. [Figure 16] This figure shows an example of AI input and output in modified example 13. [Figure 17] This figure shows an example of the input and output of the AI ​​in variation 15. [Modes for carrying out the invention]

[0009] [1. Hardware configuration of the store support system] An example of an embodiment of the store support system, store support method, and program related to this disclosure will be described. Figure 1 is a diagram showing an example of the hardware configuration of the store support system. For example, store support system 1 includes a server 10, a store terminal 20, and a customer terminal 30. Each of the server 10, store terminal 20, and customer terminal 30 can be connected to a network N such as the Internet or a LAN. Note that there may be multiple units of at least one of the server 10, store terminal 20, and customer terminal 30.

[0010] Server 10 is a server computer. For example, Server 10 includes a control unit 11, a storage unit 12, and a communication unit 13. The control unit 11 includes at least one processor. The storage unit 12 includes at least one of volatile memory such as RAM and non-volatile memory such as flash memory. The communication unit 13 includes at least one of a communication interface for wired communication and a communication interface for wireless communication.

[0011] The store terminal 20 is a computer of the store. For example, the store terminal 20 is a personal computer, a smartphone, a tablet, or a POS terminal. The store terminal 20 includes a control unit 21, a storage unit 22, a communication unit 23, an operation unit 24, and a display unit 25. The hardware configurations of the control unit 21, the storage unit 22, and the communication unit 23 may be the same as those of the control unit 11, the storage unit 12, and the communication unit 13, respectively. The operation unit 24 is an input device such as a touch panel or a mouse. The display unit 25 is a display such as a liquid crystal or an organic EL.

[0012] The customer terminal 30 is a computer of the customer. For example, the customer terminal 30 is a personal computer, a smartphone, a tablet, or a wearable terminal. The customer terminal 30 includes a control unit 31, a storage unit 32, a communication unit 33, an operation unit 34, and a display unit 35. The hardware configurations of the control unit 31, the storage unit 32, the communication unit 33, the operation unit 34, and the display unit 35 may be the same as those of the control unit 21, the storage unit 22, the communication unit 23, the operation unit 24, and the display unit 25, respectively.

[0013] Note that the programs stored in the storage units 12, 22, and 32 may be supplied to the server 10, the store terminal 20, or the customer terminal 30 via the network N. Also, at least one of a reading unit (e.g., a memory card slot) for reading a computer-readable information storage medium and an input / output unit (e.g., a USB port) for inputting / outputting data with an external device may be included in at least one of the server 10, the store terminal 20, and the customer terminal 30. For example, a program stored in the information storage medium may be supplied to at least one of the server 10, the store terminal 20, and the customer terminal 30 via at least one of the reading unit and the input / output unit.

[0014] Furthermore, the store support system 1 may include at least one computer. The computers included in the store support system 1 are not limited to the example in Figure 1. For example, the store support system 1 may include the server 10 and the store terminal 20, but not the customer terminal 30. In this case, the customer terminal 30 is located outside the store support system 1. The store support system 1 may include the server 10, but not the store terminal 20 and the customer terminal 30. In this case, the store terminal 20 and the customer terminal 30 are located outside the store support system 1. The store support system 1 may include the server 10 and other computers not shown in Figure 1.

[0015] [2. Overview of the Store Support System] In this embodiment, we take as an example the case in which the store support system 1 supports the operations of stores participating in the payment service. The payment service is a service that provides users with electronic payments (cashless payments). The payment methods available to users through the payment service may be any method. For example, the payment methods may be various cards such as credit cards, electronic money, balances not called electronic money, points, bank accounts, accounts other than bank accounts, crypto assets, wallets, or other payment methods.

[0016] Furthermore, Store Support System 1 may support the operations of stores that participate in services other than payment services. Store Support System 1 can support the operations of any store. For example, a store may be a retail store, restaurant, beauty salon, esthetic salon, or other type of store. Store Support System 1 may support the operations of merchants participating in e-commerce services, travel booking services, beauty services, telecommunications services, financial services, or other types of stores. Store Support System 1 may also support the operations of stores that do not participate in any specific service.

[0017] In this embodiment, a customer visiting a store uses a payment service from a payment application installed on the customer terminal 30 to make a payment at the store. The payment may be made by a known method. The customer may also make a payment by means other than the payment application. For example, the other means may be another application other than the payment application, a browser, the IC chip on the customer terminal 30, a physical card, or cash. For example, the store support system 1 supports the store's operations by analyzing the store's performance obtained from the payment status.

[0018] Figures 2 and 3 show examples of screens displayed on the store terminal 20. For example, a store employee operates the store terminal 20 to log in to the payment service. In this embodiment, an example is given where a business support service is provided to the store as one of the services attached to the payment service. When a store employee performs an operation on the store terminal 20 to display a sales summary, the store terminal 20 displays a sales summary screen SC1 showing a summary of the store's sales on the display unit 25, as shown in Figure 2. Note that someone other than the store employee may also display the sales summary screen SC1. For example, an employee at the head office may operate a terminal at the head office to display the sales summary screen SC1.

[0019] For example, the sales summary screen SC1 displays a sales report graphing store sales, number of visitors, average customer spending, recent trends, and sales trends, or a combination thereof. Server 10 may acquire data necessary for displaying the sales summary screen SC1 based on the customer's usage history in the payment service. Server 10 may also acquire data necessary for displaying the sales summary screen SC1 from services other than the payment service. Server 10 may receive and store data showing daily performance from the store terminal 20, and based on the stored data, may display the sales summary screen SC1 on the store terminal 20.

[0020] In this embodiment, the server 10 uses AI (Artificial Intelligence) to estimate at least one of the problems in the store and the factors affecting the store's performance. Details of the AI ​​will be described later. For example, when a store employee performs an operation on the store terminal 20 to display the estimation results by the AI, the store terminal 20 displays the estimation result screen SC2, which shows the estimation results by the AI, on the display unit 25, as shown in Figure 3. The estimation result screen SC2 shows the problems and factors estimated by the AI. The estimation result screen SC2 may also allow the store employee to input questions to the AI.

[0021] In the example shown in Figure 3, the estimation result screen SC2 displays both the problem and the factors, but the estimation result screen SC2 may display only the problem or only the factors. In other words, the AI ​​may estimate only the problem or only the factors. As described above, the store support system 1 can support store operations by having the AI ​​estimate at least one of the problem and the factors. The details of the store support system 1 will be explained below.

[0022] [3. Functions provided by the store support system] Figure 4 shows an example of the functions implemented by the store support system 1. The various components implemented by the store support system 1 can be configured by combining them into a single device or by further distributing them among multiple devices. In Figure 4, the functions implemented by the server 10 are shown among the functions implemented by the store support system 1.

[0023] For example, server 10 includes a data storage unit 100, a learning unit 101, an estimated store performance information acquisition unit 102, an input unit 103, an estimated store problem information acquisition unit 104, an estimated store factor information acquisition unit 105, and a support unit 106. The data storage unit 100 is implemented by a storage unit 12. Each of the learning unit 101, the estimated store performance information acquisition unit 102, the input unit 103, the estimated store problem information acquisition unit 104, the estimated store factor information acquisition unit 105, and the support unit 106 is implemented by a control unit 11.

[0024] [3-1. Data Storage Unit] The data storage unit 100 stores various types of data related to the payment service. For example, the data storage unit 100 stores the training database DB1 and the store database DB2.

[0025] Figure 5 shows an example of the training database DB1. The training database DB1 is a database that stores multiple training data sets for training the AI. The training data includes an input portion that is input to the AI ​​during training, and an output portion that indicates the content that should be output by the AI ​​during training. The output portion can also be called the correct answer during training. The training data is prepared by the administrator who manages the store support system 1. The pairs of input and output portions of the training data can be any pairs that are correlated with each other.

[0026] In this embodiment, the input portion is assumed to be information in the same format as the input information provided to the AI ​​during estimation. Similarly, the output portion is assumed to be information in the same format as the output information provided by the AI ​​during estimation. However, the input portion of the training data may differ slightly in format from the input information provided to the AI ​​during estimation. Similarly, the output portion of the training data may differ slightly in format from the output information provided by the AI ​​during estimation.

[0027] In this embodiment, the input portion of the training data is training store performance information, which is the performance of the training store. The training store is a store used for training (learning). That is, the training store is a store where training performance occurred. The training store may be a real store or a fictional store. The training store performance information is information that indicates whether the training store performance is good or bad. The training store performance information may be any information that indicates some training store performance and may include any items. The training store performance information is expressed by numbers, letters, other symbols other than numbers and letters, or a combination thereof that indicates the training store performance.

[0028] For example, training store performance information may include sales, profits, number of visitors, average customer spending, number of sales, product or service offerings (lineup), prices of products or services, discount rates, discount amounts, number of employees, training store shifts, changes in these, or combinations thereof. Training store performance information may show training store performance for the entire past period, or for a specific period (e.g., daily, weekly, monthly, or yearly). Training store performance information may show training store performance at a specific point in time. Training store performance information may show future training store performance rather than past training store performance; that is, training store performance information may show projected future training store performance.

[0029] In this embodiment, the output portion of the training data is training store factor information relating to training store factors, which are factors influencing the performance of the training store. Training store factors are the factors that contributed to the training store's performance. Training store factors may also be factors that caused training store problems estimated from the training store's performance. Training store factors may be any factors that are related in some way to the training store's performance. Training store factor information may be any information that indicates some training store factor and may include any items. For example, training store factor information may be natural language text that indicates the training store factor. Training store factor information may be represented by letters, numbers, other symbols other than letters and numbers, or a combination thereof.

[0030] For example, training store factor information may include factors related to the training store itself, factors related to other stores, factors attributable to events in the vicinity of the training store, weather factors, political factors, or other factors. Factors related to the training store itself may include employee job duties, number of employees, store shifts, product or service lineup, price, discount rate, discount amount, inventory level, or other factors. Other stores may or may not be competitors of the training store. Factors related to other stores may include employee job duties, number of employees, store shifts, product or service lineup, price, inventory level, or other factors. Training store factor information may indicate future training store factors rather than past training store factors. That is, training store factor information may indicate training store factors that are predicted to occur in the future.

[0031] The output portion of the training data may contain multiple pieces of information. In the example in Figure 5, the output portion of the training data includes training store problem information, which concerns the training store. A training store problem is a problem that the training store has. A training store problem is a problem that affects the performance of the training store, or a problem with operations at the training store. The training store problem information can be any information that indicates some kind of training store problem, and may include any items. The training store problem information may be represented by letters, numbers, other symbols other than letters and numbers, or a combination thereof.

[0032] For example, training store problem information may include changes in sales, profits, number of visitors, average customer spending, sales volume, employee duties, number of employees, training store shifts, product or service lineup, prices, inventory levels, or other factors. Training store problem information may not only represent past training store problems but also future training store problems. That is, training store problem information may represent training store problems that are predicted to occur in the future. The output portion of the training data does not have to include training store factor information. The output portion of the training data may consist only of training store problem information.

[0033] Figure 6 shows an example of a store database DB2. The store database DB2 is a database that stores various information about stores. In this embodiment, we give an example where various information about estimated stores is stored in the store database DB2, but various information about training stores may also be stored in the store database DB2. Estimated stores are stores that are the target of estimation by AI. Hereafter, unless there is a particular distinction between training stores and estimated stores, they will simply be referred to as stores. For example, the store database DB2 stores store identification information, estimated store basic information, and estimated store performance information. Other information may also be stored in the store database DB2. For example, the store database DB2 may store passwords that stores use to log in to payment services.

[0034] Store identification information is information that can identify a store. For example, store identification information may include an ID assigned to the store, a number assigned to the store, or the store name. Estimated store basic information is basic information about the estimated store. For example, estimated store basic information may include the store name, address, telephone number, email address, type of business, information about the products or services handled by the estimated store, or other information.

[0035] Estimated store performance information can be any information that indicates some kind of estimated store performance and may include any items. An estimated store can also be defined as a store indicated by the estimated store performance. Estimated store performance information is information that indicates whether the estimated store performance is good or bad. Estimated store performance information is expressed by numbers, letters, other symbols other than numbers and letters that indicate the estimated store performance, or a combination thereof. For example, server 10 may acquire estimated store performance information for an estimated store based on the execution status of a payment performed at that store and store it in the store database DB2. Server 10 may also acquire estimated store performance information for an estimated store from the store terminal 20 of that store and store it in the store database DB2.

[0036] For example, estimated store performance information may include estimated store sales, profits, number of visitors, average customer spending, number of sales, product or service lineup, price of product or service, discount rate, discount amount, number of employees, estimated store shifts, changes in these, or combinations thereof. Estimated store performance information may show estimated store performance for the entire past period, or for a specific period (e.g., daily, weekly, monthly, or yearly). Estimated store performance information may show future estimated store performance rather than past estimated store performance. That is, estimated store performance information may show estimated store performance that is projected to occur in the future.

[0037] The data stored in the data storage unit 100 is not limited to the above example. The data storage unit 100 may store actual AI data. There are various definitions of AI, but the AI ​​in this embodiment may be AI in various known definitions. For example, AI may be AI developed not only by machine learning methods, but also by other methods other than machine learning. Similar to the definition of AI, there are various definitions of machine learning, but machine learning in this embodiment includes various known definitions. For example, deep learning is also included in machine learning.

[0038] For example, AI may be a large-scale language model, a machine learning model not classified as a large-scale language model (a model trained using machine learning techniques), or another model. AI may be a transformer-based model such as GPT (Generative Pre-trained Transformer) or BERT (Bidirectional Encoder Representations from Transformers), or it may be another model not classified as a transformer, such as a neural network or support vector machine. AI may also be so-called generative AI.

[0039] For example, the AI ​​includes a program that processes input information given to it, and parameters referenced by that program. The parameters of the AI ​​may be the same as known parameters. For example, the parameters of the AI ​​may be weights and biases. The AI ​​may include multiple layers, such as an input layer, an intermediate layer, and an output layer. The AI's learning method may be the same as known machine learning techniques. The actual data of the AI ​​may be stored in an external computer that interacts with the server 10. In this case, the data storage unit 100 does not need to store the actual data of the AI. The server 10 uses the AI ​​stored in the external computer.

[0040] [3-2. Learning Department] The learning unit 101 trains the AI ​​using training data stored in the training database DB1. The algorithm used for AI training may be the same as a known algorithm. AI training may utilize either a supervised learning algorithm or an unsupervised learning algorithm. For example, the learning unit 101 may train the AI ​​using training data based on a known algorithm such as backpropagation or gradient descent.

[0041] For example, the learning unit 101 trains the AI ​​so that when the input portion of the training data is input to the AI, the output portion of the training data is output by the AI. The learning unit 101 inputs the input portion of the training data to the AI ​​and obtains output information, which is the information actually output by the AI. The learning unit 101 calculates the loss based on the output information output by the AI, the output portion of the training data, and a known loss function. The learning unit 101 adjusts the parameters of the AI ​​so that the loss is minimized. Once the AI ​​training is complete, the learning unit 101 records the trained AI in the data storage unit 100.

[0042] Furthermore, AI training may be performed by an external computer that interacts with server 10. In this case, the external computer stores the training database DB1. The external computer has the same functions as the learning unit 101 and trains the AI ​​with training data. In addition, various documents may be used as training data for the AI, rather than the training data prepared for the store support system 1. The data storage unit 100 may store the trained AI that has been trained by the external computer.

[0043] [3-3. Estimated Store Performance Information Acquisition Department] The estimated store performance information acquisition unit 102 acquires estimated store performance information relating to the estimated store's performance. In this embodiment, since the estimated store performance information is stored in the store database DB2, the estimated store performance information acquisition unit 102 acquires the estimated store performance information from the store database DB2. Since the store database DB2 stores estimated store performance information for various stores that could be estimated stores, the estimated store performance information acquisition unit 102 acquires the estimated store performance information for the estimated store from the store database DB2.

[0044] For example, the estimated store performance information acquisition unit 102 acquires estimated store performance information from the store database DB2 for estimated stores that are logged into the store's payment service. The estimated store performance information may also be stored in a database other than the store database DB2, a computer other than the server 10 (for example, a store terminal 20), or an information storage medium. The estimated store performance information acquisition unit 102 may acquire estimated store performance information from a database, a computer, or an information storage medium.

[0045] [3-4. Input Section] The input unit 103 inputs estimated store performance information to an AI that has learned the relationship between training store performance information, which is the performance of training stores, and training store factor information, which is the factor in said training store performance. The relationship between training store performance information and training store factor information is shown in the training data, so an AI that has learned the relationship between training store performance information and training store factor information is an AI that has already learned the training data.

[0046] Figure 7 shows an example of AI input and output. As shown in Figure 7, the input unit 103 inputs the estimated store performance information acquired by the estimated store performance information acquisition unit 102 to the AI. In this embodiment, we take the case where the actual data of the AI ​​is stored in the data storage unit 100 as an example, so the input unit 103 inputs the estimated store performance information to the AI ​​stored in the data storage unit 100. That is, the input unit 103 inputs the estimated store performance information as input information to the AI.

[0047] Furthermore, if the AI's actual data is stored in an external computer linked to the store support system 1, the input unit 103 inputs the estimated store performance information to the AI ​​by transmitting the estimated store performance information to the external computer. In other words, the input unit 103 transmitting the estimated store performance information to the external computer is equivalent to the input unit 103 inputting the estimated store performance information to the AI.

[0048] For example, if the AI ​​is a large-scale language model, the input unit 103 may input a prompt indicating instructions for the AI ​​along with the estimated store performance information. The prompt is an instruction for the AI. The prompt indicates the task that the AI ​​should perform. In other words, the prompt indicates what input information should be given to the AI ​​and what output information the AI ​​should output. The prompt may be written in natural language. The input unit 103 may also embed the estimated store performance information into the prompt and then input the prompt with the embedded estimated store performance information to the AI.

[0049] The prompt may be any sentence. For example, the prompt may be a sentence indicating that estimated store factor information should be output based on estimated store performance information. The prompt may also be a sentence such as, "You are an AI that estimates estimated store factors based on estimated store performance information. Please estimate the estimated store factors based on the estimated store performance information entered into you." The prompt may be prepared by the administrator of the store support system 1. The prompt may be stored in advance in the data storage unit 100. The prompt may be entered by the store staff.

[0050] In this embodiment, when the AI ​​estimates not only estimated store factors but also estimated store problems, the prompt may be a sentence indicating that estimated store factor information and estimated store problems should be output based on estimated store performance information. When the AI ​​estimates estimated store problems without estimating estimated store factors, the prompt may be a sentence indicating that estimated store problems should be output based on estimated store performance information. The prompt may be any sentence indicating what kind of input information is being input to the AI ​​and what kind of output information the AI ​​should output based on the input information.

[0051] [3-5. Estimated Store Problem Information Acquisition Section] The Estimated Store Problem Information Acquisition Unit 104 acquires Estimated Store Problem Information relating to Estimated Store Problems, which are problems of the Estimated Store, and which is output from the AI ​​that has Estimated Store Performance Information as input. In other words, the Estimated Store Problem Information Acquisition Unit 104 acquires Estimated Store Problem Information output by the AI ​​in response to the input of Estimated Store Performance Information. Estimated Store Problems are problems that the Estimated Store has. Estimated Store Problems are problems that affect the performance of the Estimated Store, or problems related to operations at the Estimated Store.

[0052] Estimated store problem information can be any information that indicates some kind of estimated store problem and may include any items. Estimated store problem information may be expressed by letters, numbers, other symbols other than letters and numbers, or a combination thereof. For example, estimated store problem information may be changes in the estimated store's sales, changes in profits, changes in the number of visitors, changes in the average customer spending, changes in the number of items sold, employee job descriptions, number of employees, estimated store shifts, product or service lineup, prices, inventory levels, or other factors. Estimated store problem information may indicate future estimated store problems rather than past ones. That is, estimated store problem information may indicate estimated store problems that are predicted to occur in the future.

[0053] For example, when estimated store performance information is input to the AI, the AI ​​calculates an embedding representation of the estimated store performance information based on its own parameters. The embedding representation is information that describes the characteristics of the input information. The embedding representation can be in any format. For example, the embedding representation may be in vector format, array format, matrix format, combination of multiple numbers, single number, or other format. The AI ​​may divide the estimated store performance information into units called tokens and calculate an embedding representation for each token. Based on its own parameters, the AI ​​outputs estimated store factor information corresponding to the embedding representation. The estimated store problem information acquisition unit 104 acquires the estimated store factor information output by the AI.

[0054] In this embodiment, we take the case where the actual AI data is stored in the data storage unit 100 as an example, so the estimated store problem information acquisition unit 104 acquires the estimated store factor information output from the AI ​​stored in the data storage unit 100. If the actual AI data is stored in an external computer, the estimated store problem information acquisition unit 104 only needs to acquire the estimated store factor information from the external computer. When the AI ​​outputs the estimated store factor information, the external computer sends the estimated store factor information to the server 10. The estimated store problem information acquisition unit 104 acquires the estimated store factor information from the external computer.

[0055] [3-6. Estimated Store Factor Information Acquisition Unit] The Estimated Store Factor Information Acquisition Unit 105 acquires Estimated Store Factor Information relating to Estimated Store Factors, which are factors in Estimated Store Performance, and which is output from the AI ​​that has Estimated Store Performance Information as input. In other words, the Estimated Store Factor Information Acquisition Unit 105 acquires Estimated Store Factor Information output by the AI ​​in response to the input of Estimated Store Performance Information. Estimated Store Factors are the factors that led to the Estimated Store Performance. Estimated Store Factors may also be the factors that caused Estimated Store Problems estimated from Estimated Store Performance.

[0056] Estimated store factors can be any factors that are related in some way to estimated store performance. Estimated store factor information can be any information that indicates some estimated store factors and may include any items. For example, estimated store factor information may be natural language text that indicates the estimated store factors. Estimated store factor information may be expressed by letters, numbers, other symbols other than letters and numbers, or a combination thereof. In the example in Figure 3, not only factors 1 and 2 but also problem 1 is some kind of factor and may be treated as an estimated store factor. For example, estimated store factor information may indicate problem 1. Estimated store factor information may indicate at least one of the main factors and more detailed factors of the main factors. The estimated store solutions, which will be explained in the modified examples below, may be proposed for the main factors or for the detailed factors.

[0057] For example, estimated store factor information may include factors related to the estimated store itself, factors related to other stores, factors attributable to events in the vicinity of the estimated store, meteorological factors, political factors, or other factors. Factors related to the estimated store itself may include employee job duties, number of employees, store shifts, product or service lineup, price, discount rate, discount amount, inventory level, or other factors. Other stores may or may not be competitors of the estimated store. Factors related to other stores may include employee job duties, number of employees, store shifts, product or service lineup, price, inventory level, or other factors. Estimated store factor information may indicate future estimated store factors rather than past estimated store factors. That is, estimated store factor information may indicate estimated store factors that are predicted to occur in the future.

[0058] For example, when estimated store performance information is input to the AI, the AI ​​calculates an embedded representation of the estimated store performance information based on its own parameters. The AI ​​may divide the estimated store performance information into units called tokens and calculate an embedded representation for each token. Based on its own parameters, the AI ​​outputs estimated store factor information corresponding to the embedded representation. In this embodiment, the AI ​​outputs estimated store factor information and estimated store problem information corresponding to the embedded representation. The estimated store factor information acquisition unit 105 acquires the estimated store factor information output by the AI. Although only one piece of estimated store factor information is shown in Figure 7, multiple pieces of estimated store factor information (for example, factors 1 and 2 in Figure 3) may be output by the AI.

[0059] In this embodiment, we take the case where the actual AI data is stored in the data storage unit 100 as an example, so the estimated store factor information acquisition unit 105 acquires the estimated store factor information output from the AI ​​stored in the data storage unit 100. If the actual AI data is stored in an external computer, the estimated store factor information acquisition unit 105 only needs to acquire the estimated store factor information from the external computer. When the AI ​​outputs the estimated store factor information, the external computer transmits the estimated store factor information to the server 10. The estimated store factor information acquisition unit 105 acquires the estimated store factor information from the external computer.

[0060] [3-7. Support Department] The support unit 106 supports the operations of the estimated store based on the estimated store factor information. Supporting the operations of the estimated store by the support unit 106 means that the support unit 106 performs information processing related to supporting the operations. In this embodiment, an example is given where the support unit 106 providing estimated store factor information to the store terminal 20 corresponds to the support unit 106 supporting the operations of the estimated store. The support method by the support unit 106 is not limited to the example of this embodiment. For example, the support unit 106 may support the operations of the estimated store by sending an email to the estimated store based on the estimated store factor information. The support unit 106 may also support the operations of the estimated store by sending a message to the estimated store using a communication tool other than email, based on the estimated store factor information.

[0061] For example, the support unit 106 generates display data for the estimation result screen SC2 based on the estimated store factor information and transmits the display data to the store terminal 20. The display data can be any data that allows any screen to be displayed on the store terminal 20. The display data may be the data for the entire estimation result screen SC2, or it may be a part of the data for the estimation result screen SC2. The display data may be in any format. For example, the display data may be in HTML format or image format. The estimation result screen SC2 includes the estimated store factors indicated by the estimated store factor information (factors 1 and 2 in the example in Figure 3). If future estimated store performance information is obtained, the support unit 106 may display information on the estimation result screen SC2 indicating what factors were used to estimate the future estimated store performance information (the basis for the AI ​​estimation).

[0062] In this embodiment, the AI ​​has learned the relationship between training store performance information and training store problem information, which is related to the training store's problems. When estimated store performance information is input to the AI, it outputs estimated store problem information. Therefore, the support unit 106 supports the operations of the estimated store based on the estimated store problem information. For example, the support unit 106 generates display data for the estimation result screen SC2 based on the estimated store problem information and transmits this display data to the store terminal 20.

[0063] In this embodiment, we take as an example a case in which the support unit 106 supports the operations of an estimated store based on estimated store factor information and estimated store problem information. However, the support unit 106 may support the operations of an estimated store based on estimated store problem information, without relying on estimated store factor information. In this case, the store support system 1 does not need to include the estimated store factor information acquisition unit 105. Even if the store support system 1 includes an estimated store problem information acquisition unit 104, an input unit 103, an estimated store problem information acquisition unit 104, and a support unit 106, and does not include the estimated store factor information acquisition unit 105, the operations of the estimated store can be supported by the estimated store problem information, thus solving the problems of this disclosure. Such embodiments are also within the scope of this disclosure.

[0064] Furthermore, the support unit 106 may support the operations of the estimated store based on estimated store factor information, rather than on estimated store problem information. In this case, the store support system 1 does not need to include the estimated store problem information acquisition unit 104. Even if the store support system 1 includes the estimated store factor information acquisition unit 105, the input unit 103, the estimated store problem information acquisition unit 104, and the support unit 106, and does not include the estimated store problem information acquisition unit 104, it can still support the operations of the estimated store using the estimated store factor information, thus solving the issues of this disclosure. Such an embodiment is also within the scope of this disclosure.

[0065] [4. Processes executed by the store support system] Figure 8 shows an example of processing performed by the store support system 1. The control units 11 and 21 execute the programs stored in the memory units 12 and 22, respectively, thereby executing the processing shown in Figure 8. The processing shown in Figure 8 is an example of a store support method. It is assumed that the AI ​​has completed its training before the processing shown in Figure 8 is executed.

[0066] As shown in Figure 8, when the person in charge of the estimated store launches the payment service tool (e.g., the store's payment app), the server 10 performs a login process with the store terminal 20 for the estimated store to log in to the payment service (S1). When the person in charge of the estimated store performs an operation to cause the AI ​​to estimate estimated store problem information and estimated store factor information, the server 10 obtains the estimated store performance information of the logged-in estimated store from the store database DB2 (S2). The estimated store performance information may be obtained at times other than those performed by the person in charge. For example, when a predetermined time (e.g., the end of the month) arrives, the estimated store performance information of the logged-in estimated store may be obtained from the store database DB2 and periodically sent to the estimated store that has applied for support services to assist with the operations of the estimated store.

[0067] Server 10 inputs estimated store performance information to the AI ​​(S3). When the estimated store performance information is input to the AI ​​in S3, the AI ​​calculates an embedded representation of the estimated store performance information based on parameters adjusted in previous training, and outputs estimated store problem information and estimated store factor information corresponding to the calculated embedded representation. Server 10 acquires the estimated store problem information and estimated store factor information output from the AI ​​(S4). Based on the estimated store problem information and estimated store factor information, Server 10 supports the operations of the estimated store by executing a process to display the estimation result screen SC2 with the store terminal 20 (S5), and this process ends.

[0068] [5. Summary of Embodiments] The store support system 1 of this embodiment acquires estimated store performance information. The store support system 1 inputs the estimated store performance information into an AI that has learned the relationship between training store performance information and training store factor information. The store support system 1 acquires the estimated store factor information output from the AI ​​that has been input with the estimated store performance information. The store support system 1 supports the operations of the estimated store based on the estimated store factor information. As a result, the person in charge of the estimated store can know the estimated store factors estimated by the AI, and the store support system 1 can support the operations of the estimated store using the estimated store factor information. For example, the person in charge of the estimated store can know the factors that make the estimated store performance good or bad, making it easier to consider countermeasures. Even if the person in charge of the estimated store has little knowledge of marketing, etc., the store support system 1 can use the AI ​​to provide information to compensate for the person's lack of knowledge.

[0069] Furthermore, the AI ​​has learned the relationship between training store performance information and training store problem information. Store support system 1 acquires estimated store problem information output from the AI, which has been input with estimated store performance information. Store support system 1 supports the operations of the estimated store based on the estimated store problem information. As a result, the person in charge of the estimated store can learn about the estimated store's problems, and store support system 1 can support the operations of the estimated store. For example, the person in charge of the estimated store can learn about the estimated store problems that may be occurring at the estimated store, making it easier to consider solutions to those problems.

[0070] [6. Variant] This disclosure is not limited to the embodiments described above. This disclosure may be modified as appropriate without departing from the spirit of this disclosure.

[0071] Figure 9 shows an example of a function implemented in a modified version. For example, the server 10 includes an estimated competitor performance information acquisition unit 107, an estimated store solution information acquisition unit 108, an estimated competitor solution information acquisition unit 109, a question information acquisition unit 110, an answer information generation unit 111, a question information provision unit 112, a modification unit 113, and an estimated store trend information acquisition unit 114. Each of the estimated competitor performance information acquisition unit 107, estimated store solution information acquisition unit 108, estimated competitor solution information acquisition unit 109, a question information acquisition unit 110, an answer information generation unit 111, a question information provision unit 112, a modification unit 113, and an estimated store trend information acquisition unit 114 is implemented by the control unit 11.

[0072] [6-1. Variation 1] For example, not only factors related to the estimated store itself, but also the competition of the estimated store may influence its performance. The competition of the estimated store is other stores in the same industry as the estimated store. It can also be other stores that handle the same goods or services as the estimated store. The competition of the estimated store may be other stores in the vicinity of the estimated store, or other stores that are not particularly nearby. Example 1 explains a case in which factors related to the competition of the estimated store are estimated as estimated store factors.

[0073] The AI ​​in Modification 1 has learned the relationship between training competitor performance information (the performance of training competitors, which are competitors of the training store) and training store factor information. The training store's competitors are other stores in the same industry as the training store. The training store's competitors may be other stores in the vicinity of the training store, or other stores that are not particularly nearby. The training store's competitors may be real stores or fictional stores.

[0074] Figure 10 shows an example of the input and output of the AI ​​in Modification Example 1. Figure 10 also shows the contents of the training data for Modification Example 1. As shown in Figure 10, in Modification Example 1, the input portion of the training data includes training competition performance information. Training competition performance information can be any information that influences the performance of the training competitors in some way. Training competition performance information may be information that directly indicates the performance of the training competitors, or it may be information that indirectly influences the performance without directly indicating the performance of the training competitors.

[0075] For example, training competition performance information is information that indicates the quality of training competition performance. Training competition performance information can be any information that indicates some form of training competition performance and may include any items. Training competition performance information is represented by numbers, letters, other symbols other than numbers and letters, or a combination thereof that indicates training competition performance. In Modification 1, the learning unit 101 trains the AI ​​with training data that includes training competition performance information in its input portion.

[0076] For example, competitive training performance information may include the competitive training company's sales, profits, number of visitors, average transaction value, number of sales, product or service offerings (lineup), prices, discount rates, discount amounts, number of employees, the competitive training company's shifts, changes in these factors, or combinations thereof. Alternatively, competitive training performance information may not directly represent performance but may include general information that indirectly affects performance, such as the opening of new stores, the closing of existing stores, changes in business hours, the holding of events, or combinations thereof. These factors can also affect performance and therefore qualify as competitive training performance information.

[0077] Furthermore, the training competition performance information may show training competition performance for the entire past period, or it may show training competition performance for a specific period (e.g., daily, weekly, monthly, or yearly). The training competition performance information may show training competition performance at a specific point in time. The training competition performance information may show future training competition performance rather than past training competition performance. That is, the training competition performance information may show training competition performance that is projected to occur in the future.

[0078] In Modification 1, the training store factor information indicates that the suspected competitor is the factor. For example, the training store factor information may indicate that the suspected competitor has offered discounts, expanded its product or service lineup, run a campaign, distributed coupons, advertised, or other factors. Similarly, the training store problem information indicates problems caused by the suspected competitor. For example, the training store problem information may indicate that the suspected competitor is taking away customers, has shorter operating hours than the suspected competitor, has fewer employees than the suspected competitor, has less equipment than the suspected competitor, or other problems.

[0079] The store support system 1 of the modified example 1 includes an estimated competitor performance information acquisition unit 107. The estimated competitor performance information acquisition unit 107 acquires estimated competitor performance information, which is the performance of estimated competitors that are competitors of the estimated store. Estimated competitor performance information is information that indicates whether the estimated competitor performance is good or bad. Estimated competitor performance information can be any information that indicates some kind of estimated competitor performance and may include any items. Estimated competitor performance information is expressed by numbers, letters, other symbols other than numbers and letters that indicate estimated competitor performance, or a combination thereof.

[0080] For example, estimated competitive performance information may include estimated competitor sales, profits, customer traffic, average transaction value, sales volume, product or service lineup, product or service prices, discount rates, discount amounts, number of employees, competitor shifts, changes in these factors, or combinations thereof. Estimated competitive performance information may show estimated competitive performance for the entire past period, or for a specific period (e.g., daily, weekly, monthly, or yearly). Estimated competitive performance information may show estimated competitive performance at a specific point in time. Estimated competitive performance information may show future estimated competitive performance rather than past estimated competitive performance. That is, estimated competitive performance information may show estimated competitive performance that is projected to occur in the future.

[0081] In Modification 1, we take the example of a case where estimated competitor performance information is stored in the store database DB2. For example, the estimated competitor performance information of a competitor of a certain estimated store may be stored in the store database DB2 as the estimated store performance information of other estimated stores. The estimated store performance information acquisition unit 102 acquires the estimated competitor performance information of the competitor of the estimated store that is logged into the store's payment service from the store database DB2. The estimated competitor performance information may be stored in a database other than the competitor database, a computer other than the server 10 (for example, a store terminal 20), or an information storage medium. The estimated competitor performance information acquisition unit 107 may acquire the estimated competitor performance information from a database other than the competitor database, a computer other than the server 10, or an information storage medium. Note that the method of acquiring estimated competitor performance information is not limited to the above example. For example, estimated competitor performance information may be acquired from information disclosed on a website (for example, financial statements, etc.).

[0082] Information indicating which stores are competitors of individual estimated stores may be stored in the store database DB2. In this case, the estimated competitor performance information acquisition unit 107 identifies the competitors of the estimated stores by referring to this information and acquires the estimated competitor performance information of the identified competitors. The estimated competitor performance information acquisition unit 107 may also identify the competitors of the estimated stores by referring to at least one of the business type and location of the estimated stores and acquire the estimated competitor performance information of the identified competitors. The competitors of the estimated stores may be identified by AI based on at least one of the business type and location of the estimated stores. The competitors of the estimated stores may be designated by the person in charge of the estimated store or by the administrator of the store support system 1. Furthermore, it is not necessary to specify which stores are the competitors of individual estimated stores. For example, an anonymous store may be a competitor of the estimated store.

[0083] In the modified example 1, the input unit 103 inputs estimated competitor performance information to the AI. For example, as shown in Figure 10, the input unit 103 inputs not only estimated store performance information but also estimated competitor performance information to the AI. Similar to the embodiment, the input unit 103 may input a prompt to the AI ​​along with the estimated competitor performance information. The prompt may be a sentence indicating that estimated store factor information should be output based on the estimated competitor performance information. The prompt may also be a sentence indicating that estimated store problem information should be output based on the estimated competitor performance information.

[0084] In the modified example 1, the estimated store factor information acquisition unit 105 acquires estimated store factor information output from the AI ​​that has been input with estimated competitor performance. For example, when estimated competitor performance information is input to the AI, the AI ​​calculates an embedded representation of the estimated competitor performance information based on its own parameters. The AI ​​may divide the estimated competitor performance information into units called tokens and calculate an embedded representation for each token. Based on its own parameters, the AI ​​outputs estimated store factor information corresponding to the embedded representation. The estimated competitor problem information acquisition unit acquires the estimated store factor information output by the AI.

[0085] In the example shown in Figure 10, the fact that the estimated competitor discounted a specific product is indicated as estimated store factor information. The estimated store factor information may also indicate other estimated store factors, such as the estimated competitor adding a new product to its lineup. The estimated store factor information may indicate the estimated store's own estimated store factors, along with estimated store factors related to the estimated store's competitors. The estimated store factor information may also indicate estimated store factors corresponding to a particular problem. The processing of the support unit 106 after the estimated store factor information has been acquired may be the same as in the embodiment.

[0086] In addition, the store support system 1 in the modified example 1 may include an estimated store problem information acquisition unit 104. The estimated store problem information acquisition unit 104 may acquire estimated store problem information output from the AI ​​into which estimated competitor performance information has been input. In the example in Figure 10, the estimated store problem information indicates that customers are being taken away by estimated competitors. The processing of the support unit 106 after the estimated store problem information has been acquired may be the same as in the embodiment.

[0087] In the first modified example, the AI ​​has learned the relationship between training competitor performance information and training store factor information. Store support system 1 acquires estimated competitor performance information. Store support system 1 inputs the estimated competitor performance information into the AI. Store support system 1 acquires estimated store factor information output from the AI ​​that has been inputted with estimated competitor performance. As a result, the person in charge of the estimated store can learn the estimated store factors corresponding to their estimated competitor, so that store support system 1 can effectively support the operations of the estimated store. For example, by knowing the estimated store factors related to their estimated competitor, the person in charge of the estimated store can more easily consider effective measures to improve the estimated store performance of the estimated store.

[0088] [6-2. Variation 2] For example, in Modification Example 1, if there are multiple estimated competitors for a given estimated store, the AI ​​may estimate estimated store factor information based on the estimated competitor performance information of each of the multiple estimated competitors. In other words, the estimated competitors for which estimated competitor performance information is input to the AI ​​are not limited to just one. There may be multiple estimated competitors for which estimated competitor performance information is input to the AI.

[0089] In Modified Example 2, the AI ​​has learned the relationship between the training competition performance information of each of the multiple training competitors and the training store factor information. That is, if there are multiple training competitors for a given training store, the input portion of the training data includes the training competition performance information of each of the multiple different training competitors. The learning unit 101 of Modified Example 2 causes the AI ​​to learn the training competition performance information of each of the multiple training competitors. The learning unit 101 causes the AI ​​to learn such that when the training competition performance information of each of the multiple training competitors included in the input portion of the training data is input to the AI, the output portion corresponding to that input portion is output.

[0090] The estimated competitor performance information acquisition unit 107 in Modification Example 2 acquires estimated competitor performance information for each of the multiple estimated competitors. The method for identifying estimated competitors may be the same as in Modification Example 1. For example, when the estimated competitor performance information acquisition unit 107 identifies multiple estimated competitors for a given estimated store, it acquires estimated competitor performance information for each of the multiple estimated competitors. The estimated competitor performance information acquisition unit 107 can acquire estimated competitor performance information for each of the multiple estimated competitors from the store database DB2, other databases, other computers other than the server 10, or information storage media.

[0091] In the modified example 1, the input unit 103 inputs estimated competitor performance information for each of the multiple estimated competitors to the AI. The input unit 103 inputs not only estimated store performance information to the AI, but also estimated competitor performance information for each of the multiple estimated competitors. Similar to the embodiment, the input unit 103 may input a prompt to the AI ​​along with the estimated competitor performance information for each of the multiple estimated competitors. The prompt may be a sentence indicating that estimated store factor information should be output based on the estimated competitor performance information for each of the multiple estimated competitors. The prompt may also be a sentence indicating that estimated store problem information should be output based on the estimated competitor performance information for each of the multiple estimated competitors.

[0092] In the modified example 1, the estimated store factor information acquisition unit 105 acquires estimated store factor information output from the AI, which has been input with the estimated competitor performance information of each of the multiple estimated competitors. For example, when the estimated competitor performance information of each of the multiple estimated competitors is input to the AI, the AI ​​calculates an embedding representation of the estimated competitor performance information of each of the multiple estimated competitors based on its own parameters. The AI ​​may divide the estimated competitor performance information of each of the multiple estimated competitors into units called tokens and calculate an embedding representation for each token. Based on its own parameters, the AI ​​outputs estimated store factor information corresponding to the embedding representation. The estimated competitor problem information acquisition unit acquires the estimated store factor information output by the AI.

[0093] The store support system 1 in the modified example 2 may also include an estimated store problem information acquisition unit 104, similar to the modified example 1. The estimated store problem information acquisition unit 104 may acquire estimated store problem information output from an AI that has been input with estimated competitor performance information for each of the multiple estimated competitors. The estimated store problem information acquisition unit 104 may also acquire estimated store problem information output from an AI that has been input with estimated competitor performance information. The processing of the support unit 106 after the estimated store problem information has been acquired may be the same as in the embodiment.

[0094] In the modified example 2, the AI ​​has learned the relationship between the training competitor performance information and training store factor information for each of the multiple training competitors. Store support system 1 acquires the estimated competitor performance information for each of the multiple estimated competitors. Store support system 1 inputs the estimated competitor performance information for each of the multiple estimated competitors into the AI. Store support system 1 acquires the estimated store factor information output from the AI, which has been input with the estimated competitor performance information for each of the multiple estimated competitors. As a result, the person in charge of the estimated store can learn the estimated store factors corresponding to the multiple estimated competitors, so that store support system 1 can effectively support the operations of the estimated store. For example, by knowing the estimated store factors that comprehensively consider multiple estimated competitors, the person in charge of the estimated store can more easily consider effective measures to improve the estimated store performance of the estimated store.

[0095] [6-3. Modification 3] For example, the embodiment shows a case where the AI ​​outputs estimated store factor information and estimated store problem information, but the output of the AI ​​is not limited to estimated store factor information and estimated store problem information. The AI ​​only needs to output output information that is correlated with the input information it has received (for example, estimated store performance information). Modification example 3 shows a case where the AI ​​estimates solutions to estimated store performance and outputs them as output information.

[0096] Figure 11 shows an example of the input and output of the AI ​​in Modification 3. Figure 11 also shows the contents of the training data for Modification 3. As shown in Figure 11, in Modification 3, the output portion of the training data includes training store solution information. The AI ​​in Modification 3 has learned the relationship between training store performance information and training store solution information, which is a solution for the training store. A training store solution is a solution to improve the performance of the training store.

[0097] Training store solution information can be any information that indicates a training store solution and may include any items. Training store solution information is expressed by numbers, letters, other symbols other than numbers and letters, or a combination thereof that indicates a training store solution. For example, training store solution information may indicate the goods or services that the training store should handle, their prices, discount rates, discount amounts, number of employees, training store shifts, coupons, campaigns, information that the training store should provide to customers, advertisements, business hours, work content, or a combination thereof as a training store solution. Modification 3 gives an example where the training store solution information indicates a coupon for the training store.

[0098] In Modification 3, the learning unit 101 trains the AI ​​with training data that includes training store solution information in its output portion. In Modification 3, the example given is that training store factor information, training store problem information, and training store solution information are included in the output portion of the training data, but the output portion of the training data may include only training store solution information. The AI ​​in Modification 3 may learn training store solution information without learning training store factor information and training store problem information. That is, the AI ​​may output only the estimated store solution information described later, without outputting estimated store factor information and estimated store problem information. Such embodiments are also within the scope of this disclosure. This point is not limited to Modification 3, but also applies to other modifications (for example, modifications that include content related to coupons).

[0099] In the example shown in Figure 11, the input unit 103 is shown to input estimated store performance information to the AI, similar to the embodiment. When combining modifications 1 to 3, the input unit 103 may input estimated store performance information and estimated competitor performance information of at least one estimated competitor to the AI. The process in which the AI ​​calculates an embedded representation according to the input information it has received and outputs output information according to the calculated embedded representation may be the same as in the embodiment and modifications 1 to 3. In modification 3, estimated store solution information, which will be described later, is included in the output information.

[0100] The store support system 1 of Modification 3 includes an estimated store solution information acquisition unit 108. The estimated store solution information acquisition unit 108 acquires estimated store solution information relating to estimated store solutions, which are solutions to estimated store performance, and is output from an AI into which estimated store performance information has been input. An estimated store solution is a solution for improving the performance of an estimated store. The estimated store solution information can be any information that indicates some estimated store solution and may include any items. The estimated store solution information is expressed by numbers, letters, other symbols other than numbers and letters, or a combination thereof that indicates the estimated store solution.

[0101] For example, estimated store solution information may include the products or services that the estimated store should handle, their prices, discount rates, discount amounts, number of employees, the estimated store's shifts, coupons, campaigns, information that the estimated store should provide to customers, advertisements, business hours, work content, or a combination of these. Modification 3 takes the example of estimated store solution information showing coupons for the estimated store. In the example in Figure 11, the estimated store solution information shows coupons that the estimated store should distribute to customers. The AI ​​in Modification 3 outputs estimated store solution information corresponding to the embedded representation of the estimated store performance information.

[0102] Figure 12 shows an example of the estimation result screen SC2 of Modification Example 3. The support unit 106 of Modification Example 3 supports the operations of the estimated store based on the estimated store solution information. For example, as shown in Figure 12, the support unit 106 supports the operations of the estimated store by displaying the estimation result screen SC2, which includes the contents of the coupon indicated by the estimated store solution information, on the store terminal 20. In the example in Figure 12, the estimated store solution information shows recipes using the estimated store's products as information that the estimated store should provide to customers. In this way, the support unit 106 may support the operations of the estimated store by displaying the estimation result screen SC2, which includes multiple contents indicated by the estimated store solution information, on the store terminal 20.

[0103] Furthermore, the support unit 106 may support the operations of the estimated store by editing the coupons indicated in the estimated store solution information based on operations performed from the estimation result screen SC2. Operations for editing coupons may also be operations to change the content of the coupon (for example, the target product or service, the coupon's application period, the customers to whom it will be distributed, or the incentive). The support unit 106 may also support the operations of the estimated store by distributing coupons based on operations performed from the estimation result screen SC2. The recipients of the coupons may be specified by the person in charge at the estimated store, or they may be determined as shown in the modified examples 6 to 8 described below.

[0104] Furthermore, the estimated store solution information acquisition unit 108 may acquire estimated store solution information for all estimated store factor information, or it may acquire estimated store solution information for only some of the estimated store factor information. The estimated store solution information acquisition unit 108 may acquire estimated store solution information for estimated store factor information specified by the store manager. The number of estimated store solution information items output by the AI ​​may be specified by the store manager. The estimated store solution information acquisition unit 108 may acquire one estimated store solution information item for multiple estimated store factor information items.

[0105] In the modified example 3, the AI ​​has learned the relationship between training store performance information and training store solution information. Store support system 1 obtains estimated store solution information output from the AI, which has been input with estimated store performance information. Store support system 1 supports the operations of the estimated store based on the estimated store solution information. As a result, the person in charge of the estimated store can know the estimated store solutions they should implement, so store support system 1 can effectively support the operations of the estimated store.

[0106] [6-4. Modification 4] For example, in Modification 3, if multiple competing estimated stores use the store support system 1, the effectiveness of business improvement may be diminished if the training store solution information provided to each estimated store is similar to that of the others. For example, if a coupon for a product at one estimated store is proposed as training store solution information, and the same coupon for the same product is proposed as training store solution information at another estimated store, they may hinder each other, potentially reducing the effectiveness of the coupons. Therefore, Modification 4 describes a case where competing stores are presented with different estimated store solutions.

[0107] Figure 13 shows an example of the input and output of the AI ​​in Modification 4. Figure 13 also shows the contents of the training data for Modification 4. As shown in Figure 13, in Modification 4, the input portion of the training data includes information on training conflict solutions. The AI ​​in Modification 4 has learned the relationship between training conflict solution information, which concerns solutions proposed for training conflicts that are competitors to the training store, and training store solution information that is different from the training conflict solution information. A training conflict solution is a solution to improve the operations of the training conflict.

[0108] The training conflict solution information is the training store solution information of the training conflict. An example of the specific content shown by the training conflict solution information may be the same as the specific content of the training store solution information described in Modification 3. The learning unit 101 of Modification 4 trains the AI ​​with training data that includes the training conflict solution information as its input. The training store solution information included in the output portion of the training data shows different content from the training store solution information included in the input portion corresponding to that output portion. For example, if the training store solution information was a coupon for a specific product, the training store solution information will show coupons for other products. If the training store solution information was a coupon for a specific category, the training store solution information will show coupons for other categories.

[0109] The store support system 1 in Modification 4 includes an estimated competitor solution information acquisition unit 109. The estimated competitor solution information acquisition unit 109 acquires estimated competitor solution information, which is the solution output by the AI ​​for the estimated competitors that are competitors of the estimated store. The estimated competitor solution information is the estimated store solution information of the estimated competitor. The estimated competitor solution information is the estimated store solution information provided to the support unit 106 to support the operations of the estimated competitor. An example of the specific content shown in the estimated competitor solution information may be the same as the specific content of the estimated store solution information described in Modification 3.

[0110] In Modification 4, it is assumed that estimated competitor solution information is stored in the store database DB2. The support unit 106 stores the estimated store solution information provided to a certain estimated store in the store database DB2, associating it with the store identification information of that estimated store. If the estimated store corresponds to an estimated competitor, the estimated competitor solution information acquisition unit 109 acquires the estimated store solution information stored in the store database DB2 as estimated competitor solution information. For example, the estimated competitor solution information shows the content of the coupon presented to the estimated competitor.

[0111] In Modification 4, the input unit 103 inputs estimated competitor solution information to the AI. In the example in Figure 13, the input unit 103 inputs estimated store performance information and estimated competitor solution information to the AI. When combining Modifications 1 to 4, the input unit 103 may input estimated store performance information and estimated competitor performance information of at least one estimated competitor to the AI. The process in which the AI ​​calculates an embedded representation according to the input information it has received and outputs output information according to the calculated embedded representation may be the same as in the embodiment and Modifications 1 to 4. In Modification 4, when a prompt is given to the AI, the prompt may indicate that the AI ​​should output estimated store solution information different from the estimated competitor solution information.

[0112] The estimated store solution information acquisition unit 108 in Modification 4 acquires estimated store solution information that is different from the estimated competitor solution information, and which is output from the AI ​​that has the estimated competitor solution information as input. For example, when estimated competitor solution information is input to the AI, the AI ​​calculates an embedded representation of the estimated competitor solution information based on its own parameters. The AI ​​may divide the estimated competitor solution information into units called tokens and calculate an embedded representation for each token. The AI ​​outputs estimated store solution information corresponding to the embedded representation based on its own parameters. The estimated store solution information acquisition unit 108 acquires the estimated store solution information output by the AI. For example, the estimated store solution information indicates a coupon with different content from the coupon presented to the estimated competitor (for example, a coupon for a different product or category than the coupon presented to the estimated competitor).

[0113] Furthermore, the store support system 1 in Modification 4 may include at least one of the estimated store performance information acquisition unit 102 and the estimated store problem information acquisition unit 104, or it may not include at least one of the estimated store performance information acquisition unit 102 and the estimated store problem information acquisition unit 104. In other words, the store support system 1 in Modification 4 does not need to acquire at least one of the estimated store performance information and the estimated store problem information. The support unit 106 may support the operations of the estimated store based on estimated competitor solution information, without relying on at least one of the estimated store performance information and the estimated store problem information.

[0114] In the modified example 4, the AI ​​has learned the relationship between the trained competitor solution information and the trained store solution information. Store support system 1 acquires the estimated competitor solution information. Store support system 1 inputs the estimated competitor solution information into the AI. Store support system 1 acquires the estimated store solution information output from the AI ​​that has been inputted with the estimated competitor solution information. As a result, the person in charge of the estimated store can learn about estimated store solutions that are different from those presented to their estimated competitors, so that store support system 1 can effectively support the operations of the estimated store. For example, store support system 1 can prevent situations where competitors are presented with the same or similar coupons, causing them to undermine each other.

[0115] [6-5. Variation 5] For example, in Modifications 3 and 4, the estimated store may be offered an online service as its estimated store solution, rather than an estimated store solution in a physical store. The AI ​​in Modification 5 may have learned training store solution information that indicates online services that have a predetermined relationship (e.g., affiliates or partners) with the training store, which is a physical store, as training store solutions. The training store solution information in Modification 5 indicates an online service. For example, the training store solution information indicates which online service it is and what kind of service it is. For example, the training store solution information may indicate issuing coupons for an online service, suggesting products for an online service to a user, changing the price of products for an online service, changing the product lineup for an online service, or other content.

[0116] The online services indicated in the training store solution information may be any service. For example, the online services may be sales services that sell goods or services handled by the training store (e.g., online supermarket services), payment services, e-commerce services, travel booking services, beauty services, communication services, financial services, or other services. The online services should be services that have the potential to improve the performance of the training store. Although the content indicated in the training store solution information differs from that in Modifications 3 and 4, the method of training the AI ​​with the training data may be the same as in Modifications 3 and 4.

[0117] In Modification 5, the Estimated Store Solution Information Acquisition Unit 108 acquires Estimated Store Solution Information that indicates an online service as an Estimated Store Solution. The AI ​​in Modification 5 outputs Estimated Store Solution Information that indicates an appropriate online service, corresponding to the embedded representation of the Estimated Store Performance Information. The Estimated Store Solution Information Acquisition Unit 108 acquires the Estimated Store Solution Information output from the AI. For example, the Estimated Store Solution Information indicates content corresponding to the training store solution information learned by the AI. The Estimated Store Solution Information may indicate issuing coupons for online services, suggesting products for online services to users, changing the prices of products for online services, changing the product lineup for online services, or other content. The processing of the support unit 106 after the Estimated Store Problem Information has been acquired may be the same as in Modifications 3 and 4.

[0118] The AI ​​in Modification 5 has learned training store solution information that identifies online services that have a predetermined relationship with the physical training store as training store solutions. The store support system 1 acquires estimated store solution information that identifies online services as estimated store solutions. As a result, the person in charge at the estimated store can know which online services are the estimated store solutions they should implement, so the store support system 1 can effectively support the operations of the estimated store.

[0119] [6-6. Variation 6] For example, as explained to some extent in variations 3-5, the estimated store solution information may show coupons from the estimated store. In this case, if the estimated store distributes coupons to all customers, the estimated store may suffer a loss due to the large discounts. Even if the person in charge at the estimated store tries to distribute coupons to some customers, they may not be able to determine which customers are appropriate to receive the coupons. Therefore, variation 6 explains a case where AI is used to estimate the appropriate recipients of coupons.

[0120] Figure 14 shows an example of the input and output of the AI ​​in Modification 6. Figure 14 also shows the contents of the training data for Modification 6. As shown in Figure 14, in Modification 6, the input portion of the training data includes training conflict solution information. The AI ​​in Modification 6 has learned training store solution information that indicates a training store coupon, which is a coupon for a training store, and a training distribution destination, which is where the training store coupon is distributed, as a training store solution. For example, the training store solution information shows the specific contents of the training store coupon. The training store solution information may also show the products or services to which the training store coupon applies, the discount amount, the discount rate, the expiration date, the terms of use, or other information.

[0121] The training recipients are customers who are suitable recipients of training store coupons. For example, the training recipients may be customer attributes suitable recipients of training store coupons (e.g., gender, age group, residential area, or preferences), or customer identification information that can identify individual customers (e.g., user ID, login account, or email address). Similar to modifications 3-5, the training store solution information may be prepared by the administrator of the store support system 1. Although the content of the training store solution information differs from modifications 3-5, the method of training the AI ​​with the training data may be the same as in modifications 3-5.

[0122] In Modification 6, the Estimated Store Solution Information Acquisition Unit 108 acquires Estimated Store Solution Information, which indicates an Estimated Store Coupon, which is a coupon for an Estimated Store, and an Estimated Distribution Destination, which is the distribution destination of the Estimated Store Coupon, as an Estimated Store Solution. The AI ​​in Modification 6 outputs Estimated Store Solution Information, which corresponds to the embedded representation of Estimated Store Performance Information, and indicates an Estimated Store Coupon and an Estimated Distribution Destination. The Estimated Store Solution Information Acquisition Unit 108 acquires the Estimated Store Solution Information output from the AI. The processing of the support unit 106 after the Estimated Store Problem Information has been acquired may be the same as in Modifications 3 to 5. The support unit 106 may display an Estimated Result Screen SC2 on the store terminal 20, which shows the Estimated Store Coupon and Estimated Distribution Destination indicated by the Estimated Store Solution Information, or it may distribute the Estimated Store Coupon indicated by the Estimated Store Solution Information to the Estimated Distribution Destination indicated by the Estimated Store Solution Information.

[0123] The AI ​​in Modification 6 has learned training store solution information that indicates training store coupons and training distribution destinations as training store solutions. The store support system 1 acquires estimated store solution information that indicates estimated store coupons and estimated distribution destinations as estimated store solutions. As a result, the person in charge at the estimated store can know the appropriate training distribution destinations along with the training store coupons, so the store support system 1 can effectively support the operations of the estimated store. For example, the person in charge at the estimated store can enhance the promotional effect by distributing training store coupons to the appropriate training distribution destinations.

[0124] [6-7. Variation 7] For example, in Modification 6, if a customer receives coupons from multiple competing estimated stores, the customer may only use one of the estimated store coupons from those multiple stores. In this case, the effectiveness of the estimated store coupons may be diminished. Therefore, Modification 7 describes a case where AI estimation is performed so that a customer who has received an estimated store coupon from a particular estimated store does not receive another estimated store coupon from that store. That is, each customer receives an estimated store coupon from one of multiple competing estimated stores, but does not receive estimated store coupons from any other estimated stores among those multiple estimated stores.

[0125] In Modification Example 7, the AI ​​has learned training store solution information that identifies customers other than those to whom the training competitor coupon, which is a coupon from a competitor of the training store, is delivered, as training distribution destinations. In Modification Example 7, the training distribution destinations indicated by the training store solution information of a certain training store are different from the training distribution destinations indicated by the training store solution information of a training competitor that competes with that training store. For example, the customers to whom the training distribution destinations indicated by the training store solution information of a certain training store are different from the customers to whom the training distribution destinations indicated by the training store solution information of a training competitor that competes with that training store are indicated. The attributes of the customers to whom the training distribution destinations indicated by the training store solution information of a certain training store may also be different from the attributes of the customers to whom the training distribution destinations indicated by the training store solution information of a training competitor that competes with that training store are indicated.

[0126] In Modification 7, the Estimated Store Solution Information Acquisition Unit 108 acquires Estimated Store Solution Information that indicates, as an estimated distribution destination, customers other than those to whom the estimated competitor coupon, which is a coupon from a competitor of the estimated store, is distributed. For example, when Estimated Store Performance Information is input to the AI, the AI ​​calculates an embedded representation of the Estimated Store Performance Information based on its own parameters. The AI ​​may divide the Estimated Store Performance Information into units called tokens and calculate an embedded representation for each token. Based on its own parameters, the AI ​​outputs Estimated Store Solution Information that indicates an estimated distribution destination corresponding to the embedded representation. The Estimated Store Solution Information Acquisition Unit 108 acquires the Estimated Store Solution Information output by the AI. The processing of the support unit 106 after the Estimated Store Solution Information has been acquired may be the same as in Modification 6.

[0127] Furthermore, the training data input section may include information on the distribution destinations of the training competitor's coupons. In this case, the AI ​​will be trained so that a different distribution destination from the training competitor's coupon distribution destination input to the AI ​​becomes the training distribution destination. The estimated store solution information acquisition unit 108 may input information indicating the distribution destination of the estimated competitor's coupons to the AI. Based on the embedded representation of the information input to it, the AI ​​may output estimated store solution destination information so that a different distribution destination from the estimated competitor's coupon distribution destination becomes the estimated distribution destination.

[0128] In the modified example 7, the AI ​​has learned training store solution information that indicates other customers, different from the customers to whom the training competing coupon is delivered, as training delivery destinations. The store support system 1 acquires estimated store solution information that indicates other customers, different from the customers to whom the estimated competing coupon is delivered, as estimated delivery destinations. As a result, the store support system 1 can prevent estimated store coupons from multiple competing estimated stores from reaching the same customer, thereby increasing the promotional effectiveness of the estimated store coupons.

[0129] [6-8. Variation 8] For example, in Modification 7, even if there are multiple competing estimated stores, if the estimated store coupons have different content, the effectiveness of the estimated store coupons is unlikely to be diminished even if they are distributed to the same customer. Even if the same customer receives an estimated store coupon for one product from one estimated store and an estimated store coupon for another product from another estimated store, these products do not overlap, so the customer may use both estimated store coupons. Therefore, in Modification 8, estimated store coupons with similar content are not distributed to the same customer from multiple competing estimated stores.

[0130] The AI ​​in Modification 8 has learned training store solution information that identifies other customers as training recipients, different from the customers to whom training competitor coupons similar to training store coupons are distributed. Training competitor coupons similar to training store coupons are training competitor coupons for the same product or service as the training store coupon, training competitor coupons for the same genre (category) as the training store coupon, training competitor coupons whose usage period overlaps with that of the training store coupon, or training competitor coupons with a discount amount or a similar discount amount to that of the training store coupon. By preparing such training data, it is possible to make any of the following different: the coupon usage period, the target customers, the category of the product, etc., and the name of the product, etc.

[0131] For example, the training store solution information for a particular training store will not show customers who are training competitors competing with that training store and who receive training competitor coupons similar to those of that training store as training store coupons, as training distribution destinations. The training store solution information for a training store will show customers who are training competitors competing with that training store and who do not receive training competitor coupons similar to those of that training store (for example, customers who do not receive training competitor coupons at all, or customers who receive training competitor coupons with different content) as training distribution destinations.

[0132] In Modification 8, the Estimated Store Solution Information Acquisition Unit 108 acquires Estimated Store Solution Information that indicates, as an estimated distribution destination, customers other than those to whom an estimated competitor coupon similar to the Estimated Store Coupon is distributed. For example, when Estimated Store Performance Information is input to the AI, the AI ​​calculates an embedded representation of the Estimated Store Performance Information based on its own parameters. The AI ​​may divide the Estimated Store Performance Information into units called tokens and calculate an embedded representation for each token. Based on its own parameters, the AI ​​outputs Estimated Store Solution Information that indicates an estimated distribution destination corresponding to the embedded representation. The Estimated Store Solution Information Acquisition Unit 108 acquires the Estimated Store Solution Information output by the AI. The processing of the support unit 106 after the Estimated Store Solution Information has been acquired may be the same as in Modification 7.

[0133] In the modified example 8, the AI ​​has learned training store solution information that identifies different customers as training recipients, distinct from those who receive training competitor coupons similar to the training store coupons. The store support system 1 acquires estimated store solution information that identifies different customers as estimated recipients, distinct from those who receive estimated competitor coupons similar to the estimated store coupons. This allows the store support system 1 to prevent similar estimated store coupons from reaching the same customer at multiple competing estimated stores, thereby enhancing the promotional effectiveness of the estimated store coupons. Even if there are multiple competing estimated stores, the store support system 1 can still deliver different estimated store coupons to the same customer.

[0134] [6-9. Modification 9] For example, as explained to some extent in Modification 3, the AI ​​may output multiple estimated store solution information. The estimated store solution information output by the AI ​​is not limited to one, but may be multiple. The AI ​​in Modification 9 has learned the relationship between training store performance information and multiple training store solution information. That is, the training data includes one training store performance information as an input part and multiple training store solution information as an output part. The learning unit 101 in Modification 9 trains the AI ​​so that when training store performance information included in the input part of the training data is input to the AI, multiple training store solution information, which is the output part corresponding to that input part, is output.

[0135] The estimated store solution information acquisition unit 108 in Modification 9 acquires multiple estimated store solution information. For example, when estimated store performance information is input to the AI, the AI ​​calculates an embedded representation of the estimated store performance information and outputs multiple estimated store solution information corresponding to the calculated embedded representation. The estimated store solution information acquisition unit 108 acquires the multiple estimated store solution information output from the AI. The AI ​​may divide the estimated store performance information into units called tokens and output multiple estimated store solution information corresponding to the embedded representation of each token. The support unit 106 in Modification 9 supports the operations of the estimated store based on the multiple estimated store solution information. For example, the support unit 106 displays an estimation result screen SC2 containing the estimated store solutions indicated by each of the multiple estimated store solution information on the store terminal 20. The screen in this case may be the same as in Figure 11.

[0136] In the modified example 9, the AI ​​has learned the relationship between training store performance information and multiple training store solution information. Store support system 1 acquires multiple estimated store solution information. Store support system 1 supports the operations of the estimated store based on the multiple estimated store solution information. As a result, the person in charge of the estimated store can learn about the estimated store solutions indicated by each of the multiple estimated store solution information, and can refer to more estimated store solutions, so that store support system 1 can effectively support the operations of the estimated store.

[0137] [6-10. Variation 10] For example, in Modification 9, the multiple estimated store solution information output by the AI ​​may indicate different services. For example, among the multiple estimated store solution information, one estimated store solution information may indicate introducing an e-commerce service as an estimated store solution, while another estimated store solution information may indicate introducing a communication service as an estimated store solution. In Modification 10, the AI ​​is trained to output multiple estimated store solution information for different services.

[0138] In Modification 10, the AI ​​has learned the relationship between training store performance information and multiple training store solution information related to different services. In Modification 10, the output portion of the training data includes multiple training store solution information. For example, among the multiple training store solution information, the first training store solution information indicates that the training store performance is improved by the first service as the training store solution. The second training store solution information indicates that the training store performance is improved by the second service, which is different from the first service, as the training store solution. Similarly, if three or more training store solution information are shown in the output portion of the training data, the three or more training store solution information should indicate that the training store performance is improved by different services.

[0139] The estimated store solution information acquisition unit 108 in the modified example 10 acquires multiple estimated store solution information for services that are different from each other. When estimated store performance information is input to the AI, the AI ​​calculates an embedded representation of the estimated store performance information and outputs multiple estimated store solution information for services that are different from each other, corresponding to the calculated embedded representation. The estimated store solution information acquisition unit 108 acquires the multiple estimated store solution information output from the AI. The AI ​​may divide the estimated store performance information into units called tokens and output multiple estimated store solution information for services that are different from each other, corresponding to the embedded representation of each token.

[0140] In the modified example 10, the support unit 106 supports the operations of the estimated store based on multiple estimated store solution information relating to different services. For example, the support unit 106 causes the store terminal 20 to display an estimated result screen SC2 that includes the estimated store solutions indicated by each of the multiple estimated store solution information relating to different services. The support unit 106 may also cause the store terminal 20 to display an estimated result screen SC2 that includes first estimated store solution information indicating that the estimated store performance will be improved by a first service, and second estimated store solution information indicating that the estimated store performance will be improved by a second service different from the first service.

[0141] In the modified example 10, the AI ​​has learned the relationship between training store performance information and multiple training store solution information for different services. The store support system 1 acquires multiple estimated store solution information for different services. Based on the multiple estimated store solution information for different services, the store support system 1 supports the operations of the estimated store. As a result, the person in charge of the estimated store knows the estimated store solution indicated by each of the multiple estimated store solution information for different services, and can refer to estimated store solutions for more services, so the store support system 1 can effectively support the operations of the estimated store.

[0142] [6-11. Variation 11] For example, as briefly explained in the embodiment, store staff may be able to input questions to the AI ​​from the estimation results screen SC2. In the example in Figure 3, the store staff inputs a question to the AI ​​in the input form F20. The AI ​​generates an answer to the question from the store staff. The answer from the AI ​​is displayed on the estimation results screen SC2. Modification 11 describes a configuration in which store staff input questions to the AI.

[0143] The store support system 1 of the modified example 11 includes a question information acquisition unit 110, an answer information generation unit 111, and a support unit 106. The question information acquisition unit 110 acquires question information related to the question entered by the estimated store when support is provided by the support unit 106. Support is provided by the support unit 106 after support has been provided by the support unit 106. In the example of the embodiment, the display of the estimation result screen SC2 on the store terminal 20 corresponds to support being provided by the support unit 106.

[0144] For example, a store employee enters any characters (text), numbers, symbols, or combinations thereof as a question into input form F20. The store employee can also enter sentences written in natural language as questions. The store terminal 20 sends question information indicating the sentence entered by the store employee to the server 10. The question information may include other information besides sentences written in natural language (for example, attached files). The question information acquisition unit 110 acquires the question information from the store terminal 20.

[0145] The response information generation unit 111 causes a pre-trained large-scale language model to generate response information related to the answer to the question based on the question information. The pre-trained large-scale language model may be the same as the AI ​​that outputs estimated store factor information, etc., but in the modified example 11, a case where it is different from the AI ​​that outputs estimated store factor information, etc. is given as an example. Also, in the example, the pre-trained large-scale language model is stored on an external computer that cooperates with the server 10, but the pre-trained large-scale language model may also be stored in the data storage unit 100.

[0146] Large-scale language models are trained on a variety of documents. These large-scale language models may be well-known models, such as transformer-based models like GPT or BERT, or other models not classified as transformers. The large-scale language models include parameters (e.g., weights or biases) that have been adjusted through training. Based on its own parameters, the large-scale language model computes an embedding representation of the input information, predicts the next sentence as needed, and outputs the response information. The large-scale language model may also divide the input information into tokens and compute an embedding representation of each individual token.

[0147] For example, the answer information generation unit 111 inputs question information to a trained large-scale language model. In modified example 11, since the large-scale language model is stored on an external computer, the answer information generation unit 111 inputs the question information to the large-scale language model and generates answer information by sending the question information to the external computer. The external computer causes the large-scale language model to generate answer information. The external computer sends the answer information to the server 10. The answer information generation unit 111 retrieves the answer information from the external computer.

[0148] The response information generation unit 111 may also input a pre-prepared default prompt to the trained large-scale language model along with the question information. The default prompt indicates the content of the processing that the trained large-scale language model should perform. The default prompt may be a natural language sentence indicating that the AI ​​should answer a question from the store. The default prompt may be a sentence such as, "You are an AI that answers questions from the store. Please generate response information based on the question information you have entered." The default prompt is stored in the data storage unit 100.

[0149] Figure 15 shows an example of a screen displayed on the store terminal 20 of Modification 11. The support unit 106 of Modification 11 supports the operations of the estimated store based on the response information. The response information is natural language text that represents the response from a large-scale language model. For example, the response information may be letters, numbers, symbols, or a combination thereof. As shown in Figure 15, the support unit 106 supports the operations of the estimated store by displaying the response indicated by the response information on the estimation result screen SC2. The support unit 106 may also display the response indicated by the response information on a screen other than the estimation result screen SC2. The support unit 106 may also provide the response indicated by the response information to the store terminal 20 using other means such as email.

[0150] In the modified version 11, the store support system 1 acquires question information when support is provided by the support unit 106. Based on the question information, the store support system 1 causes a trained large-scale language model to generate answer information. Based on the answer information, the store support system 1 supports the operations of the estimated store. As a result, store staff can input questions into the large-scale language model and check the answers, so the store support system 1 can support the operations of the estimated store. For example, store staff can check information such as trained store factor information and then ask questions to the large-scale language model about things they are unsure of.

[0151] [6-12. Variation 12] For example, the questions entered by a store employee in Modification 11 may also be asked by employees at other stores. Therefore, the questions entered by store employees may be useful in supporting store operations. Thus, Modification 12 describes a case in which the questions entered by a store employee are provided to a payment service provider that provides the store with an estimated result screen SC2, etc.

[0152] The store support system 1 in Modification 12 includes a question information provision unit 112. The question information provision unit 112 provides question information to the service provider to which the store subscribes. In Modification 12, as in the embodiment, the case in which the service to which the store subscribes is a payment service is given as an example. Therefore, the parts describing payment services can be read as referring to services to which the store subscribes. As explained in the embodiment, the services to which the store subscribes are not limited to payment services.

[0153] For example, the provider is a business operator that provides payment services. The provider may also be an individual or a public institution, not a business operator. The question information provision unit 112 provides question information to the provider by transmitting the question information to the provider's computer. The question information provision unit 112 may also provide question information to the provider by recording the question information on a computer other than the data storage unit 100 or server 10, or on an information storage medium. The provider's computer retrieves the question information recorded therein at any time.

[0154] The store support system 1 of the modified version 12 provides question information to service providers that the store is affiliated with. This allows the providers to understand what the store is wondering about, and the store support system 1 can provide the providers with useful information to support the store's operations. For example, the providers can prepare training data corresponding to the question information to train the AI, or they can directly provide operational support to the store.

[0155] [6-13. Variation 13] For example, in Modifications 11 and 12, the content output by the AI ​​may be changed in response to questions entered by the store staff. In the example in Figure 3, the AI ​​outputs estimated store factor information and estimated store problem information, but some store staff may find this information difficult to understand and may enter questions. In the examples of Modifications 3 to 10, the store staff may want to change the estimated store solution information output by the AI. Therefore, Modification 13 provides an example in which the output by the AI ​​is changed based on questions entered by the store staff.

[0156] Figure 16 shows an example of AI input and output in Modification 13. The store support system 1 of Modification 13 further includes a modification unit 113. The modification unit 113 modifies the AI-generated results based on the question information. For example, the modification unit 113 inputs the question information and the AI-generated results into a large-scale language model. The generated results are the information described as being generated by the AI ​​in the embodiment and Modifications 1 to 12. For example, the generated results may be estimated store factor information, estimated store problem information, estimated store solution information, or a combination thereof.

[0157] In the example in Figure 16, estimated store solution information is shown as the generated result. Furthermore, a coupon is shown as an example of estimated store solution information. The modification unit 113 may input a default prompt to the large-scale language model indicating that the generated result should be modified based on the question information. The default prompt may be a sentence such as, "You are an AI that modifies estimated store solution information based on the question information. Please modify the estimated store solution information based on the question information and output it."

[0158] For example, a large-scale language model calculates an embedded representation of the question information input to it based on pre-tuned parameters, and outputs a modified generation result based on the embedded representation. In the example in Figure 16, the question information indicates that the changes to the coupon are described, so the modified generation result will be estimated store solution information that shows the modified coupon as determined by the large-scale language model. For example, suppose the AI ​​generates estimated store solution information that shows a vegetable coupon, but a store employee inputs a question stating that the delivery date of the coupon will be changed because the delivery of vegetables will be delayed. In this case, the large-scale language model generates estimated store solution information that shows the coupon with the changed delivery date. Similarly, if the modification unit 113 modifies any generation result other than the estimated store solution information, the modification unit 113 modifies the generation result.

[0159] In the modified version 13, the store support system 1 modifies the AI-generated results based on the question information. This allows store staff to modify the AI-generated results in response to their questions, enabling the store support system 1 to provide more flexible support. For example, store staff can modify the AI-generated results without having to perform any changes themselves, allowing the store support system 1 to effectively support store operations.

[0160] [6-14. Variation 14] For example, in this embodiment, estimated store performance information is obtained based on the usage status of the payment service or communication with the store terminal 20, but estimated store performance information may be obtained by other methods. Since the receipt issued by the store contains details such as the payment amount, estimated store performance information may be obtained by collecting a receipt image from the customer terminal 30. Estimated competitor performance information showing the performance of the estimated competitor may be obtained by analyzing a receipt image showing a receipt from an estimated competitor, which is a competitor of the estimated store.

[0161] In the modified version 14, the estimated store performance information acquisition unit 102 acquires estimated store performance information based on a receipt image relating to the estimated store's receipt. The receipt image shows a receipt. For example, a customer who has used the estimated store operates the customer terminal 30 to take a picture of the estimated store's receipt. The customer terminal 30 transmits the receipt image to the server 10. The estimated store performance information acquisition unit 102 analyzes the receipt image to acquire information that can identify the estimated store (e.g., store name, telephone number, or email address) and the payment amount.

[0162] The analysis of the receipt image may be performed using known image processing methods. For example, the estimated store performance information acquisition unit 102 may perform optical character recognition on the receipt image to obtain information that can identify the estimated store and the payment amount from the receipt image. The estimated store performance information acquisition unit 102 may also obtain information that can identify the product or service from the receipt image. Estimated store performance information may be obtained by storing this information obtained from the receipt image in a database and aggregating sales, etc., for each estimated store. The processing after the estimated store performance information has been obtained may be the same as in the embodiments and modifications 1 to 13.

[0163] The store support system 1 in modification 14 acquires estimated store performance information based on receipt images. This allows the store support system 1 to support store operations based on estimated store performance information. For example, the store support system 1 can acquire information such as estimated store sales that cannot be obtained solely from the usage status of payment services. Even in environments where sales information cannot be obtained from the store terminal 20, the store support system 1 can acquire information such as estimated store sales from receipt images.

[0164] [6-15. Variation 15] For example, not only estimated store performance but also trends in estimated store performance may be displayed on the sales summary screen SC1. In the example in Figure 2, "Trend" corresponds to the trend in estimated store performance. While the trend in estimated store performance may be obtained by analyzing the estimated store performance information with a program, Modification 15 explains the case where the trend in estimated store performance is estimated by AI.

[0165] Figure 17 shows an example of the input and output of the AI ​​in Modification 15. Figure 17 also shows the contents of the training data in Modification 1. For example, the AI ​​has learned the relationship between training store performance information and training store trend information, which is the trend of training store performance. In Modification 1, the training store trend information is included in the output portion of the training data. The training store trend information describes the trend of training store performance over a certain period (for example, the most recent day, week, month, or year) in natural language. The trend of training store performance can also be described as the change in training store performance.

[0166] The training data input, which is the training store performance information, is assumed to show changes in training store performance over a certain period of time. The learning unit 101 of Modification 15 trains the AI ​​on training data that includes training store performance information as its input and training store trend information as its output. The learning unit 101 trains the AI ​​so that when training store performance information is input, training store trend information is output. The output portion of the training data may include at least one of training store factor information, training store problem information, and training store solution information.

[0167] The store support system 1 of the modified example 15 includes an estimated store trend information acquisition unit 114. The estimated store trend information acquisition unit 114 acquires estimated store trend information, which is the trend of estimated store performance, output from the AI ​​that has been input with estimated store performance information. For example, the input unit 103 inputs estimated store performance information to the AI ​​in the same manner as in the embodiment. The AI ​​calculates an embedded representation of the estimated store performance information and outputs estimated store trend information corresponding to the embedded representation. The estimated store trend information acquisition unit 114 acquires the estimated store trend information output from the AI. The estimated store trend information describes the trend of estimated store performance over a certain period (for example, the most recent day, week, month, or year) in natural language. The trend of estimated store performance can also be described as a change in estimated store performance.

[0168] In the modified example 15, the support unit 106 supports the operations of the estimated store based on the estimated store trend information. For example, consider the case where the support unit 106 providing estimated store trend information to the store terminal 20 is equivalent to the support unit 106 supporting the operations of the estimated store. The support unit 106 supports the operations of the estimated store by generating display data for the sales summary screen SC1 based on the estimated store trend information and transmitting the display data to the store terminal 20. The support method by the support unit 106 may be other methods. For example, the support unit 106 may support the operations of the estimated store by sending an email to the estimated store based on the estimated store trend information. The support unit 106 may also support the operations of the estimated store by sending a message to the estimated store using a communication tool other than email based on the estimated store trend information.

[0169] The AI ​​in Experimental Example 15 has learned the relationship between training store performance information and training store trend information. Store support system 1 acquires estimated store trend information. Store support system 1 supports the operations of the estimated store based on the estimated store trend information. As a result, the person in charge of the estimated store can understand the estimated store trends, and store support system 1 can support the operations of the estimated store.

[0170] [6-16. Other variations] For example, the above variations may be combined.

[0171] For example, the functions described as being implemented on server 10 may be implemented on store terminal 20, customer terminal 30, or other computers. The functions described as being implemented on server 10 may be shared among multiple computers.

[0172] [7. Addendum] For example, a store support system can also be configured as follows: (1) The Estimated Store Performance Information Acquisition Unit acquires estimated store performance information, which is the performance of estimated stores, An AI (Artificial Intelligence) that has learned the relationship between training store performance information, which is the performance of the training store, and training store factor information, which is the factor in the training store performance, is input to input the estimated store performance information. Estimated store factor information relating to estimated store factors which are factors in the estimated store performance, and an estimated store factor information acquisition unit that acquires the estimated store factor information output from the AI ​​into which the estimated store performance information is input, Based on the estimated store factor information, a support unit is provided to support the operations of the estimated store. A store support system that includes this. (2) The AI ​​has learned the relationship between the training store performance information and the training store problem information, which is the problem of the training store. The store support system further includes an estimated store problem information acquisition unit that acquires estimated store problem information relating to the estimated store problem, which is a problem of the estimated store, and which acquires the estimated store problem information output from the AI ​​into which the estimated store performance information has been input. The support unit shall, based on the estimated store problem information, support the operations of the estimated store. (1) The store support system described above. (3) The AI ​​has learned the relationship between training competitor performance information, which is the performance of training competitors that are competitors of the training store, and training store factor information. The store support system further includes an estimated competitor performance information acquisition unit that acquires estimated competitor performance information relating to estimated competitor performance, which is the performance of estimated competitors that are competitors of the estimated store. The input unit inputs the estimated competitive performance information to the AI, The estimated store factor information acquisition unit acquires the estimated store factor information output from the AI ​​into which the estimated competitor performance information has been input. (1) or (2) The store support system described above. (4) The AI ​​has learned the relationship between the performance information of each of the multiple training competitors and the training store factor information. The estimated competitor performance information acquisition unit acquires the estimated competitor performance information for each of the multiple estimated competitors, The input unit inputs the estimated competitor performance information for each of the plurality of estimated competitors to the AI. The estimated store factor information acquisition unit acquires the estimated store factor information output from the AI ​​into which the estimated competitor performance information of each of the plurality of estimated competitors has been input. (3) The store support system described above. (5) The AI ​​has learned the relationship between the training store performance information and the training store solution information, which is the solution for the training store. The store support system further includes an estimated store solution information acquisition unit that acquires estimated store solution information output from the AI ​​into which the estimated store performance information is input, which is estimated store solution information relating to the estimated store solution that is a solution to the estimated store performance. The support unit supports the operations of the estimated store based on the estimated store solution information. A store support system as described in any of (1) to (4). (6) The AI ​​has learned the relationship between training competitor solution information, which concerns solutions proposed for training competitors that are competitors of the training store, and training store solution information that is different from said training competitor solution information. The store support system further includes an estimated competitor information acquisition unit that acquires estimated competitor information relating to estimated competitor solutions output by the AI ​​for estimated competitors, which are competitors of the estimated store. The input unit inputs the estimated conflict solution information to the AI, The estimated store solution information acquisition unit acquires estimated store solution information that is different from the estimated competitor solution information, and which is output from the AI ​​into which the estimated competitor solution information was input. (5) The store support system described above. (7) The AI ​​has learned training store solution information that indicates online services having a predetermined relationship with the training store, which is a physical store, as the training store solution. The estimated store solution information acquisition unit acquires the estimated store solution information that indicates the online service as the estimated store solution. (5) or (6) The store support system described above. (8) The AI ​​has learned the training store solution information, which indicates the training store coupon, which is a coupon for the training store, and the training distribution destination, which is the destination to which the training store coupon is distributed, as the training store solution. The estimated store solution information acquisition unit acquires estimated store solution information that indicates the estimated store coupon, which is a coupon for the estimated store, and the estimated distribution destination, which is the distribution destination of the estimated store coupon, as the estimated store solution. A store support system as described in any of (5) to (7). (9) The AI ​​has learned the training store solution information which indicates, as the training distribution destination, other customers different from the customers to whom the training competitor coupon, which is a coupon from a competitor of the training store, is distributed. The estimated store solution information acquisition unit acquires estimated store solution information that indicates, as the estimated distribution destination, other customers different from the customers to whom the estimated competitor coupon, which is a coupon from a competitor of the estimated store, is distributed. (8) The store support system described above. (10) The AI ​​has learned the training store solution information which indicates, as the training distribution destination, other customers different from the customers to whom the same training store coupon and the training competitor coupon are distributed. The estimated store solution information acquisition unit acquires estimated store solution information that indicates, as the estimated recipient, other customers who are different from the customers to whom the estimated competing coupon similar to the estimated store coupon is distributed. (9) The store support system described above. (11) The AI ​​has learned the relationship between the training store performance information and the multiple training store solution information. The estimated store solution information acquisition unit acquires a plurality of estimated store solution information, The support unit supports the operations of the estimated stores based on the multiple estimated store solution information. A store support system as described in any of (5) to (10). (12) The AI ​​has learned the relationship between the training store performance information and the information on multiple training store solutions related to different services. The estimated store solution information acquisition unit acquires the plurality of estimated store solution information relating to services that are different from each other. The support unit supports the operations of the estimated stores based on the multiple estimated store solution information relating to different services. (11) The store support system described above. (13) The aforementioned store support system is When support is provided by the aforementioned support unit, the question information acquisition unit acquires question information related to the question entered by the estimated store, A response information generation unit causes a trained large-scale language model to generate response information related to the answer to the question based on the aforementioned question information. It further includes, The support unit shall, based on the response information, support the operations of the estimated store. A store support system as described in any of (1) to (12). (14) The store support system further includes a question information provision unit that provides the question information to the service providers to which the store is affiliated. (13) The store support system described above. (15) The store support system further includes a modification unit that modifies the AI-generated results based on the question information. (13) or (14) The store support system described above. (16) The estimated store performance information acquisition unit acquires the estimated store performance information based on the receipt image relating to the estimated store's receipt. A store support system as described in any of (1) to (15). (17) The AI ​​has learned the relationship between the training store performance information and the training store trend information, which is the trend of the training store performance. The store support system further includes an estimated store trend information acquisition unit that acquires estimated store trend information, which is estimated store performance trend information, output from the AI ​​into which the estimated store performance information is input. The support unit provides support for the operations of the estimated stores based on the estimated store trend information. A store support system as described in any of (1) to (16). [Explanation of symbols]

[0173] 1 Store support system, N Network, 10 Server, 11,21,31 Control unit, 12,22,32 Storage unit, 13,23,33 Communication unit, 20 Store terminal, 30 Customer terminal, 24,34 Operation unit, 25,35 Display unit, 100 Data storage unit, 101 Learning unit, 102 Estimated store performance information acquisition unit, 103 Input unit, 104 Estimated store problem information acquisition unit, 105 Estimated store factor information acquisition unit, 106 Support unit, 107 Estimated competitor performance information acquisition unit, 108 Estimated store solution information acquisition unit, 109 Estimated competitor solution information acquisition unit, 110 Question information acquisition unit, 111 Answer information generation unit, 112 Question information provision unit, 113 Change unit, 114 Estimated store trend information acquisition unit, DB1 Training database, DB2 Store database, F20 Input form, SC1 Sales summary screen, SC2 Estimation result screen.

Claims

1. The Estimated Store Performance Information Acquisition Unit acquires estimated store performance information, which is the performance of estimated stores, An input unit inputs the estimated store performance information to an AI (Artificial Intelligence) that has learned the relationship between training store performance information, which is the performance of the training store, and training store factor information, which is the factor in the training store performance. Estimated store factor information relating to estimated store factors which are factors of the estimated store performance, and an estimated store factor information acquisition unit that acquires the estimated store factor information output from the AI ​​into which the estimated store performance information is input, Based on the estimated store factor information, a support unit is provided to support the operations of the estimated store. A store support system that includes this.

2. The AI ​​has learned the relationship between the training store performance information and the training store problem information, which is the problem of the training store. The store support system further includes an estimated store problem information acquisition unit that acquires estimated store problem information relating to the estimated store problem, which is a problem of the estimated store, and which acquires the estimated store problem information output from the AI ​​into which the estimated store performance information has been input. The support unit provides support for the operations of the estimated store based on the estimated store problem information. The store support system according to claim 1.

3. The AI ​​has learned the relationship between training competitor performance information, which is the performance of training competitors that are competitors of the training store, and training store factor information. The store support system further includes an estimated competitor performance information acquisition unit that acquires estimated competitor performance information relating to estimated competitor performance, which is the performance of estimated competitors that are competitors of the estimated store. The input unit inputs the estimated competitor performance information to the AI, The estimated store factor information acquisition unit acquires the estimated store factor information output from the AI ​​into which the estimated competitor performance information has been input. The store support system according to claim 1 or 2.

4. The AI ​​has learned the relationship between the performance information of each of the multiple training competitors and the training store factor information. The estimated competitor performance information acquisition unit acquires the estimated competitor performance information for each of the multiple estimated competitors, The input unit inputs the estimated competitor performance information for each of the plurality of estimated competitors to the AI. The estimated store factor information acquisition unit acquires the estimated store factor information output from the AI ​​into which the estimated competitor performance information of each of the plurality of estimated competitors has been input. The store support system according to claim 3.

5. The AI ​​has learned the relationship between the training store performance information and the training store solution information, which is the solution for the training store. The store support system further includes an estimated store solution information acquisition unit that acquires estimated store solution information, which is a solution to the estimated store performance, and which acquires the estimated store solution information output from the AI ​​into which the estimated store performance information is input. The support unit supports the operations of the estimated store based on the estimated store solution information. The store support system according to claim 1 or 2.

6. The AI ​​has learned the relationship between training competitor solution information, which concerns solutions proposed to training competitors that are competitors of the training store, and training store solution information that is different from said training competitor solution information. The store support system further includes an estimated competitor information acquisition unit that acquires estimated competitor information relating to estimated competitor solutions, which are solutions output by the AI ​​for estimated competitors, which are competitors of the estimated store. The input unit inputs the estimated conflict solution information to the AI, The estimated store solution information acquisition unit acquires estimated store solution information that is different from the estimated competitor solution information, and which is output from the AI ​​into which the estimated competitor solution information was input. The store support system according to claim 5.

7. The AI ​​has learned the training store solution information, which indicates online services that have a predetermined relationship with the training store, which is a physical store, as the training store solution. The estimated store solution information acquisition unit acquires the estimated store solution information that indicates the online service as the estimated store solution. The store support system according to claim 5.

8. The AI ​​has learned the training store solution information, which indicates the training store coupon, which is a coupon for the training store, and the training distribution destination, which is the recipient of the training store coupon, as the training store solution. The estimated store solution information acquisition unit acquires estimated store solution information that indicates the estimated store coupon, which is a coupon for the estimated store, and the estimated distribution destination, which is the distribution destination of the estimated store coupon, as the estimated store solution. The store support system according to claim 5.

9. The AI ​​has learned the training store solution information which indicates, as the training distribution destination, other customers different from the customers to whom the training competitor coupon, which is a coupon from a competitor of the training store, is distributed. The estimated store solution information acquisition unit acquires estimated store solution information that indicates, as the estimated distribution destination, other customers different from the customers to whom the estimated competitor coupon, which is a coupon from a competitor of the estimated store, is distributed. The store support system according to claim 8.

10. The AI ​​has learned the training store solution information which indicates, as the training distribution destination, other customers different from the customers to whom the same training store coupon and the training competitor coupon are distributed. The estimated store solution information acquisition unit acquires estimated store solution information that indicates, as the estimated recipient, other customers who are different from the customers to whom the estimated competing coupon similar to the estimated store coupon is distributed. The store support system according to claim 9.

11. The AI ​​has learned the relationship between the training store performance information and the multiple training store solution information. The estimated store solution information acquisition unit acquires a plurality of the estimated store solution information, The support unit supports the operations of the estimated stores based on the multiple estimated store solution information. The store support system according to claim 5.

12. The AI ​​has learned the relationship between the training store performance information and the information on multiple training store solutions related to different services. The estimated store solution information acquisition unit acquires the plurality of estimated store solution information relating to services that are different from each other, The support unit supports the operations of the estimated stores based on the multiple estimated store solution information relating to different services. The store support system according to claim 11.

13. The aforementioned store support system is When support is provided by the aforementioned support unit, the question information acquisition unit acquires question information related to the question entered by the estimated store, A response information generation unit causes a trained large-scale language model to generate response information related to the answer to the question based on the aforementioned question information. It further includes, The support unit shall, based on the response information, support the operations of the estimated store. The store support system according to claim 1 or 2.

14. The store support system further includes a question information provision unit that provides the question information to the service provider to which the estimated store belongs. The store support system according to claim 13.

15. The store support system further includes a modification unit that modifies the AI-generated results based on the question information. The store support system according to claim 13.

16. The estimated store performance information acquisition unit acquires the estimated store performance information based on the receipt image relating to the estimated store's receipt. The store support system according to claim 1 or 2.

17. The AI ​​has learned the relationship between the training store performance information and the training store trend information, which is the trend of the training store performance. The store support system further includes an estimated store trend information acquisition unit that acquires estimated store trend information, which is estimated store trend information, that is output from the AI ​​into which the estimated store performance information is input. The support unit provides support for the operations of the estimated stores based on the estimated store trend information. The store support system according to claim 1 or 2.

18. A computer, We obtain estimated store performance information, which is the performance of estimated stores. The estimated store performance information is input to an AI (Artificial Intelligence) that has learned the relationship between training store performance information, which is the performance of the training store, and training store factor information, which is the factor in the training store performance. Estimated store factor information relating to estimated store factors which are factors of the estimated store performance, wherein the estimated store factor information is obtained from the AI ​​into which the estimated store performance information is input. Based on the estimated store factor information, support the operations of the estimated store. Store support methods.

19. Estimated Store Performance Information Acquisition Unit, which acquires estimated store performance information, which is the performance of estimated stores. An AI (Artificial Intelligence) that has learned the relationship between training store performance information, which is the performance of the training store, and training store factor information, which is the factor in the training store performance, inputs the estimated store performance information into an input unit. Estimated store factor information relating to estimated store factors which are factors of the estimated store performance, and an estimated store factor information acquisition unit that acquires the estimated store factor information output from the AI ​​into which the estimated store performance information has been input. Based on the estimated store factor information, a support department that assists the operations of the estimated store, A program that makes a computer function.

Images