Generation apparatus, generation method, and generation program

Abstract

This enables the appropriate guidance of customer purchasing behavior. [Solution] The generation device 100 receives a prompt that includes a command to generate information on guiding product purchasing behavior, which is used to guide customer purchasing behavior at the target store, to a large-scale language model that has been provided with information on the business operations of the target store, and generates information on guiding product purchasing behavior. The generation device 100 outputs the generated information on guiding product purchasing behavior in a predetermined format.

Description

Technical Field

[0006] , , ,

[0001] The present invention relates to a generation device, a generation method, and a generation program.

Background Art

[0002] In a store, various measures may be implemented to promote the sale of products and guide the purchasing behavior of customers who visit the store. For example, "shelf allocation" can be cited as a method for guiding customers' purchasing behavior to promote sales.

[0003] Shelf allocation refers to planning the products to be placed on the shelves (hereinafter, may be referred to as "display shelves") for sale, the placement positions, the quantities to be placed, etc. in order to guide customers who take purchasing actions in the store to the product shelves or the placement positions of the products and promote the sale of the products (hereinafter, may be referred to as "shelf allocation plan").

[0004] Shelf allocation is implemented based on a shelf allocation plan for the display shelves created in advance. Therefore, as a conventional technique for appropriately implementing shelf allocation, there is known a technique of performing a shelf allocation simulation based on product information, shelf information, etc. and creating a shelf allocation reproduction image in which products are virtually placed on the product shelves (for example, see Patent Document 1).

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, conventional technologies sometimes have difficulty appropriately guiding customer purchasing behavior. For example, conventional technologies perform shelf layout simulations based on product information and shelf information, so they have difficulty creating appropriate shelf layout plans that take into account the store's situation and guide customers who visit the store to the target display shelves. [Means for solving the problem]

[0007] Therefore, in order to solve the above-mentioned problems and achieve the objective, the present invention is characterized by comprising: a generation unit that receives a prompt including a command to generate information on guiding product purchasing behavior used to guide customer purchasing behavior at a target store, to a large-scale language model provided with information on the business operations of the target store, and generates said information on guiding product purchasing behavior; and an output unit that outputs the said information on guiding product purchasing behavior generated by the generation unit in a predetermined format. [Effects of the Invention]

[0008] The present invention has the effect of enabling appropriate guidance of customer purchasing behavior. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 is a diagram illustrating the overall process of generating behavioral guidance information related to purchasing according to the embodiment. [Figure 2] Figure 2 shows the configuration of the generating apparatus according to this embodiment. [Figure 3] Figure 3 is a table diagram showing an example of store business information according to the embodiment. [Figure 4] Figure 4 is a table diagram showing an example of shelf layout information according to the embodiment. [Figure 5] Figure 5 is a table diagram showing an example of posted information according to the embodiment. [Figure 6] Figure 6 shows an example of a two-dimensional code display according to the embodiment. [Figure 7]Figure 7 shows an example of shelf layout information generation according to the embodiment. [Figure 8] Figure 8 shows an example of route information generation according to this embodiment. [Figure 9] Figure 9 shows an example of the generation of posted information according to this embodiment. [Figure 10] Figure 10 is a flowchart showing the processing performed by the generation apparatus according to the embodiment. [Figure 11] Figure 11 shows an example of a computer that realizes the generation apparatus according to the embodiment. [Modes for carrying out the invention]

[0010] Hereinafter, embodiments for carrying out the present invention (hereinafter referred to as "embodiments") will be described with reference to the drawings. However, each embodiment is not limited to those described below.

[0011] <Overview> (background) To promote product sales in stores, various measures are taken, including store and exterior layout design, shelf allocation, in-store POP (Point of Purchase advertising), and promotions targeting customers who visit the target store (hereinafter sometimes simply referred to as "customers").

[0012] For example, in the "shelf arrangement" described above, products are arranged on the shelves in a way that makes them easily noticeable to customers browsing the store and considering purchasing them, thereby guiding those customers towards making a purchase.

[0013] Regarding the shelf division of the display shelves, based on the intentions of the manufacturers of the products, the business plan of the store, the inventory status, the season, the attributes of the customers visiting the store, the experience of the person in charge of shelf division, etc., a "shelf division plan" for the products to be arranged, the positions to be arranged, the quantities to be arranged, etc. for the target display shelves is created in advance. Therefore, in order to efficiently create the shelf division plan, a reference technique for virtually arranging products on the display shelves by executing a shelf division simulation based on information such as products and display shelves is known.

[0014] However, since the above reference technique is a technique for performing a shelf division simulation using information on products and display shelves, it is difficult to create an appropriate shelf division plan based on other external factors such as other store conditions, customer conditions, seasons, and weather conditions.

[0015] Therefore, for example, it is difficult for the reference technique to realize an appropriate shelf division according to the store conditions and customer conditions and induce the customer's purchasing behavior so that the customer can appropriately stroll around the store. In addition, in other respects, it is difficult for the reference technique to guide customers to the display shelves where products with high customer purchasing desire are arranged or to the display shelves where products such as unsold items are arranged based on the results of the shelf division simulation.

[0016] (Processing by the generation device 100) Therefore, the generation device 100 according to the present embodiment outputs information for creating an appropriate shelf division plan and guiding customers to the target store or the target display shelves, using information related to the business of the target store including information on the sales status of products and information on customers visiting the store.

[0017] Note that the above "information related to the business of the target store including information on the sales status of products and information on customers visiting the store" may be referred to as "store business information" in the following items. In addition, the information for guiding customers to the target store or the target display shelves (information related to guiding the purchasing behavior of products used to guide the behavior related to the purchase of customers in the target store) may be referred to as "behavior guiding information".

[0018] Here, we will explain the overall process performed by the generation device 100. Figure 1 is a diagram illustrating the overall process of generating behavioral guidance information related to purchasing according to this embodiment. The generation device 100 shown in Figure 1 is an example of a computer that provides the technology to realize the information processing described below.

[0019] First, the generation device 100 provides a predetermined large-scale language model with store operation information (Figure 1 (1-1)) as prior knowledge, such as the sales status of products at the target store, past shelf layout information, and information about customers who visit the store (Figure 1 (1-2)).

[0020] The "provision of prior knowledge" mentioned above refers to inputting specific information into a large-scale language model in advance, and includes training and tuning the large-scale language model, inputting information into the large-scale language model, and inserting or substituting information related to the prior knowledge into prompts input to the large-scale language model.

[0021] Next, the generation device 100 inputs a command to generate behavioral guidance information (Figure 1(2-1)) for providing predetermined guidance to customers' purchasing behavior, such as shelf layout information, route information, and posting information, into a large-scale language model using a prompt expressed in natural language text (Figure 1(2-2)), causing the large-scale language model to generate behavioral guidance information (Figure 1(2-3)).

[0022] The "shelf allocation information" described above refers to the information used to allocate product shelves in the target store, and includes information related to the shelf allocation plan, such as the products to be placed on the target shelves, the placement location of those products, and the quantity to be placed. The "route information" is information generated using the above shelf allocation information, and includes information that shows customers visiting the target store the route within the store to the shelves where the products they desire or are presumed to desire are located. The "posted information" includes images of the product and text information such as descriptions, impressions, and opinions about the product, which are generated for the purpose of generating a predetermined number of impressions when the target product is posted on the internet.

[0023] The generation device 100 then outputs the above-mentioned shelf layout information, route information, posting information, and other behavioral guidance information to store staff, customers, etc., in a predetermined format (Figure 1 (3-1)).

[0024] For example, the generating device 100 outputs a list or layout diagram of the display shelves (shelf allocation information) that shows which products to place on the target display shelves, where on the shelves, and in what quantities (Figure 1 (3-2)). Also, for example, the generating device 100 outputs information showing the route within the store to the display shelves where the target products are placed, superimposed on a map of the target store (hereinafter sometimes referred to as the "in-store map") to customers who visit the store (Figure 1 (3-3)). Also, for example, if a customer who visits the store meets predetermined conditions, the generating device 100 outputs posted information about the target products to that customer (Figure 1 (3-4)).

[0025] In this way, the generation device 100 according to this embodiment does not simply perform shelf layout simulations based on information about products and display shelves, but can generate and output information to guide customers to take desired purchasing actions according to the store's operating conditions and the customers' circumstances. As a result, the generation device 100 has the effect of enabling appropriate guidance of customer purchasing behavior.

[0026] <Description of Generator 100> The configuration of the generation device 100 according to this embodiment will now be described. Figure 2 is a diagram showing the configuration of the generation device 100 according to this embodiment. As shown in Figure 2, the generation device 100 has a communication unit 110, a storage unit 120, and a control unit 130.

[0027] Although not shown in Figure 2, the generation device 100 may be equipped with an input unit such as a keyboard or mouse to receive input such as operations from an administrator. Furthermore, the generation device 100 may be equipped with a display or the like to show the acquired store business information, generated behavior guidance information, etc., to an administrator.

[0028] (Communications Department 110) The communication unit 110 performs data communication related to the input of store business information and commands for generating behavior guidance information, which are input by the person in charge or acquired from an external information processing device. The communication unit 110 also performs data communication related to the output of generated behavior guidance information, etc.

[0029] The communication unit 110 is implemented using a NIC (Network Interface Card) or the like, and controls communication via telecommunication lines such as a LAN (Local Area Network) or the Internet. The communication unit 110 can be connected to the network via wired or wireless connection as needed, and can send and receive information bidirectionally with terminal devices, etc.

[0030] (Storage unit 120) The storage unit 120 stores data and programs used for various processes by the control unit 130, as well as various data acquired through the operation of the control unit 130. The storage unit 120 is implemented using semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or storage devices such as hard disks and optical discs. As shown in Figure 2, the storage unit 120 includes a store business information DB 121, a shelf allocation information DB 122, a posting information DB 123, and a generation model DB 124.

[0031] (Store Business Information DB121) The Store Operations Information DB121 is a database that stores information about the operations of a target store (store operations information), including product sales status and information about customers visiting the store. Specifically, the Store Operations Information DB121 stores information such as the inventory status and sales status of products in the store (product sales information), information about past shelf layout plans (past shelf layout information), information about customers visiting the store (dwell time information / customer information), weather information (weather information), and a map of the target store (in-store map).

[0032] Here, an example of store business information stored in the store business information DB121 will be explained using Figure 3. Figure 3 is a table diagram showing an example of store business information according to the embodiment.

[0033] The store business information DB121 associates information related to each item in the store business information, such as "No." (information that identifies individual data included in the store business information), with "store identification information," "product sales information," "past shelf layout information," "customer dwell information," "customer information," "weather information," and "in-store map," and stores this information in a table format, for example, as shown in Figure 3. The letters "A to G" written in each item in the table diagram shown in Figure 3 are legends for the information contained in each item.

[0034] The "store identification information" mentioned above is information used to identify the target store, and includes, for example, the store name, store number, and other information combining text, numbers, symbols, etc., to identify the store. "Product sales information" includes information on the sales history of products at the store, such as inventory information, sales amount, and sales quantity. "Past shelf layout information" is historical information on shelf layouts implemented at the target store in the past, and includes, for example, historical information on the target display shelves, products, product placement, and quantity when shelf layouts were implemented in the past. "Customer dwell time information" is information on customer dwell time at the target store, and includes, for example, information shown in the form of a list or map, such as how many customers are in the store and where customers are concentrated within the store. "Customer information" is information on customers who visit the target store, and includes, for example, customer identification information, gender and age, hobbies and preferences, etc., in a form that excludes information that could identify an individual customer. "Weather information" is weather information related to the target store, and includes, for example, information indicating the temperature, humidity, weather, and season in the vicinity of the target store. Furthermore, the "store map" contains layout information for the target store.

[0035] Here, the "dwell time information," "customer information," and "weather information" mentioned above may be information acquired in real time. For example, "dwell time information" may be updated in real time based on information acquired by cameras, motion sensors, etc., installed in the store. Similarly, "customer information" may be updated in real time based on information acquired by cameras installed at the store entrance or by devices that perform predetermined wireless communication with customer terminals installed in the store. Furthermore, "weather information" may be updated based on information acquired in real time or at predetermined intervals from external information processing equipment, etc.

[0036] (Shelf allocation information DB122) The shelf allocation information DB122 is a database that stores information (shelf allocation information) regarding the products to be placed on the target display shelves in the target store, as well as the placement location and quantity of those products. Specifically, the shelf allocation information DB122 stores information such as information about the products to be placed on the target display shelves (product information) and information about the location where those products will be placed (placement location).

[0037] Here, an example of shelf layout information stored in the shelf layout information DB122 will be explained using Figure 4. Figure 4 is a table diagram showing an example of shelf layout information according to the embodiment. The shelf layout information DB122 associates information related to each item, such as "No," which is information that identifies individual data included in the shelf layout information, with "product information" and "placement location," and stores it in a table format, for example, as shown in Figure 4. The letters "H" and "I" written in each item of the table diagram shown in Figure 4 are legends for the information included in each item.

[0038] The "product information" mentioned above includes, for example, the product name, product type, quantity, sales price, and sales trends of the product collected from the internet, etc., to be placed on the target display shelf. The "placement location" includes, for example, the location of the display shelf where the product will be placed, and the placement of the product within that display shelf.

[0039] (Post Information DB123) The Post Information DB123 is a database that stores text information (post information) such as product images, descriptions, impressions, and opinions about a product, which is generated with the aim of generating a predetermined number of impressions when a product is posted on the internet. Specifically, the Post Information DB123 stores information such as the name of the product related to the post information (product name), product images and text related to the product (post content), and information indicating the classification of the post information (classification).

[0040] Here, an example of post information stored in the post information DB123 will be explained using Figure 5. Figure 5 is a table diagram showing an example of post information according to the embodiment. The post information DB123 associates information related to each item, such as "No," which is information that identifies individual data included in the post information, with "product name," "post content," and "classification," and stores it in a table format, for example, as shown in Figure 5. The letters "J to L" written in each item of the table diagram shown in Figure 5 are legends for the information included in each item.

[0041] The "product name" mentioned above is information that identifies the product to be posted on the internet, and includes, for example, information in which the product name is expressed by a combination of text, numbers, symbols, etc. The "posted content" includes information that combines an image of the product to be posted on the internet with text that includes a description of the product, opinions, impressions, and expressions to attract attention. The "classification" includes information that identifies whether the posted information generated more than a specified number of impressions (hereinafter sometimes referred to as "becoming a topic of discussion") when it was posted on the internet, etc., and whether or not critical opinions were received in the past (hereinafter referred to as "going viral").

[0042] (Generative model DB124) The generative model DB124 is a database that stores predetermined models used by the generation unit 134 (described later) to generate behavioral guidance information. For example, the generative model DB124 can store large-scale language models as generative models.

[0043] Specifically, the generation device 100 according to this embodiment can use at least one of the following as a large-scale language model: "ChatGPT®", a large-scale language model possessing general-purpose knowledge, and "tsuzumi®", a predetermined large-scale language model on which adapter tuning is performed (see, for example, References 1 and 2).

[0044] (Reference 1):ChatGPT(OpenAI),<URL:https: / / openai.com / chatgpt> ,<Searched on October 7, 2021> (Reference 2): NTT version of large-scale language model "tsuzumi",<URL:https: / / www.rd.ntt / research / LLM_tsuzumi.html> ,<Searched on October 7, 2021>

[0045] Furthermore, the generation device 100 according to this embodiment can perform processing using a predetermined learning model that falls within the category of machine learning models, in addition to the aforementioned "ChatGPT®" and "tsuzumi®".

[0046] (Control unit 130) Now, let's return to Figure 2 and continue the explanation. The control unit 130 has an internal memory for temporarily storing programs and processing data that define various processing procedures of the generation device 100, and is realized by electronic circuits such as a CPU (Central Processing Unit) and an MPU (Micro Processing Unit), and integrated circuits such as an ASIC (Application Specific Integrated Circuit) and an FPGA (Field Programmable Gate Array). As shown in Figure 2, the control unit 130 has an acquisition unit 131, a supply unit 132, a reception unit 133, a generation unit 134, and an output unit 135.

[0047] (Acquisition part 131) The acquisition unit 131 stores store business information acquired from external information processing devices, etc., via the communication unit 110, etc., in the store business information DB 121. Specifically, the acquisition unit 131 acquires product sales information including inventory information and sales history of products for the store, past shelf layout information, dwell time information, customer information, etc., from an information processing device that manages information related to the store, and stores it in the store business information DB 121 as store business information.

[0048] Furthermore, the acquisition unit 131 acquires weather information related to the store from an external information processing device or the like that stores weather information in the target area, and stores it in the store business information DB 121 as store business information.

[0049] (Provider 132) The provisioning unit 132 provides the generation model with store operation information as prior knowledge, which includes at least one of the following: product sales information, past shelf layout information, dwell time information, customer information, and weather information.

[0050] For example, the provision unit 132 can input information such as the inventory of products at the target store, real-time product sales status, weather conditions, customer dwell time in the store, and past shelf arrangement performance performed by the person in charge of shelf arrangement (hereinafter sometimes referred to as "store staff") into a large-scale language model using prompts, and pre-train the large-scale language model to generate behavior guidance information taking the above information into consideration.

[0051] (Reception desk 133) The reception unit 133 receives predetermined information and accesses that trigger the generation unit 134, described later, to begin generating behavioral guidance information. The reception unit 133 then uses the received predetermined information and accesses as a trigger to transmit a command to generate behavioral guidance information to the generation unit 134.

[0052] For example, the reception unit 133 receives information including a processing start command transmitted from a terminal device operated by a store employee. The reception unit 133 then uses the received processing start command as a trigger to send a shelf layout information generation command to the generation unit 134.

[0053] For example, when a customer enters the store, the reception unit 133 receives information including a processing start command transmitted from a camera or motion sensor that detected the customer's arrival. The reception unit 133 then uses the received processing start command as a trigger to send a command to the generation unit 134 to generate route information using shelf layout information.

[0054] Furthermore, the reception unit 133 accepts access from a terminal device operated by a customer based on the reading of an identifier such as a two-dimensional code displayed at the target store by the terminal device. The reception unit 133 then uses the access from the terminal device as a processing start command and sends a command to the generation unit 134, described later, to generate the posted information.

[0055] Here, an example of displaying an identifier read by customers will be explained using Figure 6. Figure 6 is a diagram showing an example of a two-dimensional code display according to the embodiment. As shown in Figure 6, the identifier to be read is displayed, for example, near the display shelf 10. For example, the identifier for reading is displayed together with text such as "Let's make a post about '○○ chocolate'!" as shown in Figure 6(1).

[0056] The customer reads the identifier using their own terminal device. The customer's terminal device then accesses the generation device 100 based on information such as the URL (Uniform Resource Locator) contained in the identifier. The reception unit 133 then accepts the access from the terminal device.

[0057] (Generation unit 134) The generation unit 134 receives prompts containing commands to generate behavioral guidance information from a large-scale language model provided with store business information, and generates behavioral guidance information. For example, the generation unit 134 generates shelf layout information, route information using the shelf layout information, posting information, etc., as behavioral guidance information. A specific example of the behavioral guidance information generation process by the generation unit 134 will be explained in the following sections.

[0058] (Output section 135) The output unit 135 outputs the behavior guidance information generated by the generation unit 134 in a predetermined format. For example, the output unit 135 outputs the behavior guidance information, such as route information and posting information, generated by the generation unit 134 using the shelf layout information, in a predetermined format, such as displaying it overlaid on a list of shelf layout information and route information or on an in-store map, or displaying the posting information on the customer's terminal.

[0059] When outputting in the predetermined format described above, the output unit 135 can output behavioral guidance information to, for example, terminal devices operated by store staff, terminal devices such as smartphones operated by customers, and terminal devices such as digital signage installed in the store. Specific examples of the output processing of behavioral guidance information by the output unit 135 based on the predetermined format will be described in subsequent sections.

[0060] (An example of processing) From here, using Figures 7 to 9, we will explain an example of the generation and output processing of behavior guidance information by the generation device 100. The first example shown below is an example in which the generation device 100 generates shelf layout information. The second example is an example in which the generation device 100 generates route information based on the shelf layout information. The third example is an example in which the generation device 100 generates information to be posted on the internet, etc., for sales promotion at the target store.

[0061] The large-scale language models used in the following examples 1 through 3 are, for example, models that have been pre-trained using information provided by the provider, such as product inventory, product sales status, weather conditions, customer dwell time, and past shelf allocation performance.

[0062] (Example 1) First, as a first example, we will explain with reference to Figure 7 an example in which the generation device 100 generates shelf layout information using product sales information, past shelf layout information, dwell time information, weather information, etc., included in store business information, and outputs it to store staff, etc. Figure 7 is a diagram showing an example of shelf layout information generation according to the embodiment.

[0063] In the first example, the generation device 100 (generation unit) receives a command from a large-scale language model via prompts to generate shelf layout information including product names, product locations, and product sales prices, and generates the shelf layout information as behavioral guidance information.

[0064] For example, as shown in (1) of Figure 7, the generation device 100 (generation unit) uses prompts including "<role>", "<constraints>", "<commands>", etc., to cause a large-scale language model to generate shelf layout information.

[0065] The "<Role>" above is a text that defines the role of a person who embodies a specific role in the large-scale language model and generates information for the large-scale language model. As described above, the generation device 100 (generation unit) can accurately generate desired behavioral guidance information for the large-scale language model by using prompts that include the "definition of the role".

[0066] For example, the generation device 100 (generation unit) can assign the role of "responsible for creating a shelf layout plan for arranging products on shelves in a store" to the large-scale language model by using a prompt that includes the "<Role>" shown in (1-1) of Figure 7. As a result, the generation device 100 (generation unit) can generate a more accurate shelf layout plan (shelf layout information) by having the large-scale language model perform the processing as an expert in creating shelf layout plans.

[0067] Furthermore, the "<Constraints>" mentioned above is text that describes predetermined constraints on the execution of processing when the large-scale language model generates information. As described above, the generation device 100 (generation unit) can accurately generate desired behavioral guidance information from the large-scale language model by using prompts that include "processing instructions based on processing constraints".

[0068] For example, the generation device 100 (generation unit) can use prompts containing the "<Constraints>" shown in (1-2) of Figure 7 to impose constraints on the large-scale language model, such as "generate shelf layout information for products sold at the target store," "use the display shelves used at the target store," "consider past product sales history, nearby weather information, and past shelf layout information," and "follow the store layout of the target store when generating route information." As a result, the generation device 100 (generation unit) can prevent the generation of shelf layout information unrelated to the store when the large-scale language model generates shelf layout information, and can generate a more accurate shelf layout plan (shelf layout information).

[0069] Furthermore, the "<command>" mentioned above is a text containing instructions for causing the large-scale language model to perform a desired process. The generation device 100 (generation unit) can cause the large-scale language model to perform a desired process by inputting a prompt containing the "<command>" which describes the desired process into the large-scale language model.

[0070] For example, the generation device 100 (generation unit) can use a prompt containing "<command>" as shown in (1-3) of Figure 7 to cause the large-scale language model to generate shelf allocation information that includes "the target product, the location of the display shelf, the placement position on the display shelf, the sales amount for each product, and route information indicating the location of the display shelf." As a result, when the generation device 100 (generation unit) causes the large-scale language model to generate shelf allocation information, it can generate a more accurate shelf allocation plan (shelf allocation information) that includes the desired information.

[0071] The generation device 100 (output unit) then outputs the generated shelf layout information to the store staff (Figure 7 (2)). Specifically, based on the generated shelf layout information, the generation device 100 (output unit) outputs information to the store staff as behavioral guidance information, which shows the product name, the location of the product, and the sales price of the product in a list format.

[0072] For example, the generating device 100 (output unit) can display a list including the product name, the location of the product, and the sales price of the product, as shown in (2-1) of Figure 7. Furthermore, the generating device 100 (output unit) can display this list superimposed on a store map.

[0073] Furthermore, the generating device 100 (output unit) outputs information to the store staff as behavioral guidance information, which displays the placement of products on the target display shelf superimposed on the layout of the target display shelf.

[0074] For example, the generation device 100 (output unit) can display image information, such as (2-2) in Figure 7, indicating where and in what quantity the target product should be placed on the target display shelf. Furthermore, the generation device 100 (output unit) can overlay this image information onto a store map.

[0075] As shown in the first example, the generation device 100 supports the efficient and appropriate implementation of shelf allocation by displaying shelf allocation information to store staff by overlaying it on a list or in-store map.

[0076] (Second example) Next, as a second example, we will explain, using Figure 8, an example in which the generation device 100 uses the shelf layout information, dwell time information, and customer information generated in the first example to generate route information for guiding customers who have come to the store to the target display shelves, and outputs it to the customers. Figure 8 is a diagram showing an example of route information generation according to the embodiment.

[0077] In the second example, the generation device 100 (generation unit) generates route information for guiding customers to designated display shelves using a large-scale language model, based on the shelf layout information generated in the first example, and using the prompt shown in Figure 8 (1).

[0078] For example, as shown in (1) of Figure 8, the generation device 100 (generation unit) uses prompts including "<role>", "<constraints>", "<commands>", etc., to cause a large-scale language model to generate route information.

[0079] Specifically, the generation device 100 (generation unit) uses prompts with the "<role>" shown in (1-1) of Figure 8 to assign the large-scale language model the roles of "responsible for creating shelf layout plans for placing products on shelves in stores" and "guiding customers who visit the store to the sales location of the target products."

[0080] Furthermore, the generation device 100 (generation unit) imposes constraints on the large-scale language model, such as "use the generated shelf allocation information" and "generate route information based on the in-store layout," by using prompts that contain the "<Constraints>" shown in (1-2) of Figure 8.

[0081] Furthermore, the generation device 100 (generation unit) uses prompts containing "<command>" as shown in (1-3) of Figure 8 to cause the large-scale language model to perform tasks such as "estimating the products that customers who visit the store wish to purchase," "generating route information to the display shelves where the estimated products are located," and "integrating the route information superimposed on the in-store map with the shelf layout information."

[0082] The generation device 100 (output unit) then outputs the generated route information to customers who visit the target store (Figure 8 (2)).

[0083] Furthermore, as shown in (2-1) of Figure 8, the generating device 100 (output unit) outputs to the customer as behavioral guidance information, which includes the product name, the sales price of the product, and route information plotted on a map related to the target store showing the route to the product's location.

[0084] For example, as shown in (2-2) of Figure 8, the generating device 100 (output unit) outputs information to the customer as behavioral guidance information, displaying the product name, the location of the product, and the sales price of the product side by side.

[0085] As shown in the second example, the generating device 100 appropriately guides the customer's purchasing behavior by overlaying route information to the location of products that the customer is expected to want to purchase, based on shelf layout information, onto a store map and displaying it to the customer.

[0086] (Third example) Next, as a third example, we will explain with reference to Figure 9 an example in which the generation device 100 uses product sales information and past posting information of the target store to generate posting information that customers who visit the store can use to post about the products sold on websites and the like, and outputs it to the customer. Figure 9 is a diagram showing an example of posting information generation according to the embodiment.

[0087] In the third example, the generation device 100 (providing unit) first provides the large-scale language model with inventory information related to the inventory products and the content of posts related to the inventory products that have generated a predetermined number of impressions on the internet as prior knowledge, and performs pre-training. The generation device 100 (providing unit) also provides the large-scale language model with the content of posts related to the inventory products that have generated a predetermined number of impressions on the internet as prior knowledge, and performs pre-training.

[0088] The generating device 100 (generating unit) then reads the identifier displayed at the target store and, using prompts, inputs a command to generate posting information that includes the product name of the in-stock product, a description of the in-stock product, the sales location of the in-stock product, and the content of the post related to the in-stock product, thereby generating the posting information as action guidance information.

[0089] Furthermore, the generation device 100 (generation unit) generates posting information using prompts that include constraints, at least one of the following: prohibition of generating posting content that has been generated in the past, and prohibition of generating posting content related to products that are identical or similar to posting content that has caused controversy on the internet in the past.

[0090] For example, as shown in (1) of Figure 9, the generation device 100 (generation unit) uses prompts including "<role>", "<constraints>", "<commands>", etc., to cause a large-scale language model to generate route information.

[0091] Specifically, the generation device 100 (generation unit) uses prompts with the "<role>" shown in (1-1) of Figure 9 to assign the roles of "person in charge of developing promotional strategies for sales promotion in stores" and "specialist in developing posts that will become popular when posted on the internet" to the large-scale language model.

[0092] Furthermore, the generation device 100 (generation unit) imposes constraints on the large-scale language model, such as "generate posting information for products that satisfy predetermined conditions," by using prompts containing the "<constraint conditions>" shown in (1-2) of Figure 9.

[0093] Specifically, the generation device 100 (generation unit) imposes constraints on the large-scale language model, such as "products that are sold at the target store and have a certain number of units in stock," "products that have generated more than XX impressions on the internet in the past," and "products whose sales as of Month XX, Day XX are less than XX yen." As a result, the generation device 100 (generation unit) can generate posting information to make products that were popular in the past but are currently unsold as inventory items popular again.

[0094] The generation device 100 (generation unit) imposes a constraint on the large-scale language model that "it does not generate content identical to previously posted content." This allows the generation device 100 (generation unit) to prevent duplicate posts from being posted on the internet.

[0095] Furthermore, the generation device 100 (generation unit) can generate posting information while imposing the following constraints. For example, the generation device 100 (generation unit) can impose the constraint that "the content must not be identical or similar to content that has received a certain number of criticisms on the internet" on a large-scale language model. This allows the generation device 100 (generation unit) to prevent posting information on the internet from becoming a source of online outrage.

[0096] For example, the generation device 100 (generation unit) can impose a constraint on the large-scale language model that "posted information must always have the word "PR" added to it to clearly indicate that it is an advertisement." This allows the generation device 100 (generation unit) to avoid criticism, such as being accused of engaging in stealth marketing, and prevent posted information on the internet from becoming a source of online outrage.

[0097] The generation device 100 (generation unit) uses prompts containing the "<command>" shown in (1-3) of Figure 9 to cause the large-scale language model to generate "posting information used by customers who visit the store to post about the products sold on websites on the internet, etc."

[0098] The generation device 100 (output unit) then outputs the generated post information to the customer who has read the identifier displayed at the target store (Figure 9 (2)).

[0099] As shown in the third example, the generating device 100 supports efficient and appropriate product sales promotion by displaying posting information that is expected to become a topic of conversation for customers, thereby guiding customers to post such information on the internet or the like.

[0100] (Processing procedure by the generating device 100) Next, the processing procedure implemented by the generation device 100 according to this embodiment will be explained using Figure 10. Figure 10 is a flowchart showing the processing performed by the generation device 100 according to this embodiment.

[0101] The supply unit 132 provides store business information to the generation model (S101). If no behavior guidance information is to be generated (No. in S102), the generation device 100 waits for processing.

[0102] On the other hand, if behavioral guidance information is to be generated (Yes in S102), the generation device 100 performs the following processing. For example, the generation device 100 may start the process of generating behavioral guidance information when it receives a generation command from a store employee, when a customer enters the store, when a customer reads an identifier posted inside the store, etc.

[0103] The generation unit 134 inputs commands to generate behavior guidance information into a large-scale language model using prompts, and generates behavior guidance information (S103). Next, the output unit 135 outputs the generated behavior guidance information in a predetermined format (S104). Then, the generation device 100 terminates processing.

[0104] (effect) Next, we will explain the effects of the generation device 100 according to this embodiment. Conventionally, when performing shelf layout simulations using product and display shelf information, there have been challenges in creating appropriate shelf layout plans based on store conditions, customer conditions, and other external factors such as seasonal and weather conditions, making it difficult to appropriately guide customer purchasing behavior within the store.

[0105] Therefore, the generation unit 134 of the generation device 100 according to this embodiment receives a prompt that includes a command to generate behavior guidance information from a large-scale language model provided with store business information, and generates behavior guidance information. The output unit 135 of the generation device 100 outputs the behavior guidance information generated by the generation unit 134 in a predetermined format.

[0106] Therefore, the generation device 100 according to this embodiment has the effect of enabling appropriate guidance of customer purchasing behavior. For example, the generation device 100 enables store staff to create and execute shelf layout plans more efficiently and effectively than before, based on the generated shelf layout information. Furthermore, the generation device 100 enables appropriate guidance of customers based on the generated shelf layout information, including products tailored to the customer visiting the target store, the placement location of those products, and route information including guidance routes to those placement locations.

[0107] Furthermore, the generation apparatus 100 according to this embodiment achieves predetermined effects by performing the processes described below.

[0108] The generation unit 134 receives a command via prompts to generate shelf layout information, which includes product names, product placement locations, and product sales prices, from a large-scale language model that is provided with store operation information, which includes at least one of the following as prior knowledge: product sales information, past shelf layout information, customer dwell time information, customer information, and weather information. The generation unit 134 generates shelf layout information as behavioral guidance information.

[0109] Through the process described above, the generation device 100 can generate appropriate shelf layout information in real time, tailored to the store's conditions, customer behavior, and other external factors. As a result, the generation device 100 enables the implementation of appropriate shelf layouts and customer guidance to facilitate optimal purchasing behavior for customers visiting the target store.

[0110] Based on the shelf layout information generated by the generation unit 134, the output unit 135 outputs to the store staff, as behavioral guidance information, at least one of the following: information showing the product name, the location of the product, and the sales price of the product in a list format; and information showing the product name and the location of the product on the target display shelf superimposed on the layout of the target display shelf.

[0111] Through the process described above, the generating device 100 can output appropriate shelf layout information to store staff, generated according to the store's situation, customer situation, and other external factors. As a result, the generating device 100 has the effect of enabling store staff to appropriately implement shelf layouts that guide customers visiting the target store to the optimal purchasing behavior.

[0112] Based on the shelf layout information generated by the generation unit 134, the output unit 135 outputs route information to the customer as guidance information, which includes at least one of the following: information displaying the product name, the product's location, and the product's selling price side by side, and information plotting the product name, the product's selling price, and the route to the product's location on a map related to the target store.

[0113] Through the process described above, the generating device 100 outputs route information based on appropriate shelf layout information generated according to the store's conditions, customer conditions, and other external factors to customers visiting the target store. As a result, the generating device 100 has the effect of guiding customers visiting the target store to make optimal purchasing decisions.

[0114] The generation unit 134 receives a command via prompts to generate posting information, which includes the product name of the product in stock, a description of the product, the sales location of the product, and the content of posts related to the product, based on the reading of identifiers displayed at the target store. This command is then used to generate posting information as behavioral guidance information.

[0115] Through the process described above, the generation device 100 generates and outputs to customers post information about products that were popular in the past but have since become unsold. In other words, the generation device 100 promotes sales of the product by encouraging customers to post the generated information on the internet based on the benefits of "becoming popular," etc.

[0116] The generation unit 134 generates posting information using a prompt that includes a constraint condition that includes at least one of the following: prohibition of generating posting content that has been generated in the past, and prohibition of generating posting content related to products that are identical or similar to posting content that has received a predetermined criticism on the internet in the past.

[0117] Through the process described above, the generator 100 can prevent customers from being suspected of "plagiarizing information" or criticized for "inappropriate posting" when they post information on the internet. As a result, customers can confidently post the information generated by the generator 100 on the internet. Therefore, the generator 100 has the effect of enabling customers to post the information it has generated.

[0118] The provisioning unit 132 provides the large-scale language model with prior knowledge of posts related to inventory products generated by the generation unit 134, specifically those posts that have generated a predetermined number of impressions on the internet.

[0119] The above-described process enables the generation device 100 to generate posting information that contains content more likely to become a topic of discussion. As a result, customers can actively post the posting information generated by the generation device 100 onto the internet. Therefore, the generation device 100 has the effect of enabling customers to post the generated posting information.

[0120] <Variation> The following describes modifications that can be implemented by the generation apparatus 100 according to this embodiment.

[0121] (Data, etc.) The behavior guidance information, shelf allocation information, route information, posting information, "roles, constraints, and commands" described in the prompts, the names of the functional parts of the generation device 100, steps, processes, and names of steps or processes used in the description of the above embodiment are merely examples and can be changed at will.

[0122] For example, while it was explained that the store business information DB121 stores information related to each of the following items in a table format, such as "No," which identifies individual data included in the store business information, "store identification information," "past shelf layout information," "dwell time information," "customer information," "weather information," and "in-store map," the items and information within each item that are stored are not limited to those shown in Figure 3. Similarly, while it was explained that the shelf layout information DB122 stores information related to each of the following items in a table format, such as "No," which identifies individual data included in the shelf layout information, "product information," and "placement," the items and information within each item that are stored are not limited to those shown in Figure 4. Furthermore, while it was explained that the posting information DB123 stores information related to each of the following items in a table format, such as "No," which identifies individual data included in the posting information, "product name," "post content," and "classification," the items and information within each item that are stored are not limited to those shown in Figure 5.

[0123] (Regarding the use of generative models) In this embodiment, the model (large-scale language model) used by the generation device 100 is described as being stored in the generation model DB 124 of the storage unit 120, but this is not limited to this. For example, the generation device 100 can access an external information processing device (server, etc.) and use a predetermined model.

[0124] (Flowcharts, etc.) In flowcharts, each step may be rearranged as long as it does not create inconsistencies, and some steps may be omitted. Furthermore, conjunctions such as "next," "continue," "in addition," "at this time," and "on this occasion" in flowchart descriptions do not limit the order or timing of the processes in the flowchart.

[0125] <Hardware Configuration> Each component of the illustrated device is a functional concept and does not necessarily have to be physically configured as shown. In other words, the specific forms of distribution and integration of each device are not limited to those shown, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads and usage conditions. Furthermore, each processing function performed by each device can be implemented, all or any part of it, by a CPU and the program that is analyzed and executed by that CPU, or by hardware using wired logic.

[0126] Furthermore, among the processes described in this embodiment, all or part of those described as being performed automatically can be performed manually using known methods. In addition, the processing procedures, control procedures, specific names, and information including various data and parameters shown in the drawings can be arbitrarily changed unless otherwise specified.

[0127] <Program> In one embodiment, the various devices constituting the generation device 100 can be implemented by installing the generation program as packaged software or online software on a desired computer. For example, by having the above generation program executed on an information processing device, the various devices constituting the generation device 100 can be made to function. The information processing device referred to here includes desktop or notebook personal computers. In addition, the information processing device also includes mobile communication terminals such as smartphones and mobile phones, and slate terminals such as PDAs (Personal Digital Assistants).

[0128] Figure 11 shows an example of a computer that implements the generation device 100 according to the embodiment. The computer 1000 has, for example, memory 1010 and CPU 1020. The computer 1000 also has a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These components are connected by a bus 1080.

[0129] Memory 1010 includes ROM (Read Only Memory) 1011 and RAM 1012. ROM 1011 stores, for example, a boot program such as BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to the hard disk drive 1090. The disk drive interface 1040 is connected to the disk drive 1100. For example, a removable storage medium such as a magnetic disk or optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to, for example, a mouse 1110 and a keyboard 1120. The video adapter 1060 is connected to, for example, a display 1130.

[0130] The hard disk drive 1090 stores, for example, an OS (Operating System) 1091, an application program 1092, a program module 1093, and program data 1094. That is, the programs that define the various processes of the various devices constituting the generation device 100 are implemented as program modules 1093 in which executable code for a computer is written. The program modules 1093 are stored, for example, in the hard disk drive 1090. For example, a program module 1093 for performing processes similar to the functional configuration of the various devices constituting the generation device 100 is stored in the hard disk drive 1090. Note that the hard disk drive 1090 may be replaced by an SSD (Solid State Drive).

[0131] Furthermore, the configuration data used in the processing of the embodiment described above is stored as program data 1094 in, for example, memory 1010 or hard disk drive 1090. The CPU 1020 then reads the program module 1093 and program data 1094 stored in memory 1010 or hard disk drive 1090 into RAM 1012 as needed and executes the processing of the embodiment described above.

[0132] Furthermore, the program module 1093 and program data 1094 are not limited to being stored in the hard disk drive 1090; for example, they may be stored in a removable storage medium and read by the CPU 1020 via a disk drive 1100 or the like. Alternatively, the program module 1093 and program data 1094 may be stored in another computer connected via a network (LAN, WAN (Wide Area Network), etc.). The program module 1093 and program data 1094 may then be read from the other computer by the CPU 1020 via a network interface 1070.

[0133] <Other> Although this embodiment has been described above, this embodiment is not limited by the description and drawings that constitute part of the disclosure. That is, all other embodiments, examples, and operational techniques made by those skilled in the art based on this embodiment are included in the scope of this embodiment. [Explanation of Symbols]

[0134] 100 generator 110 Communications Department 120 Storage section 121 Store Business Information Database 122 Shelf allocation information DB 123 Post Information Database 124 Generative Model DB 130 Control Unit 131 Acquisition Department 132 Provision Department 133 Reception Department 134 Generation part 135 Output section

Claims

1. A generation unit receives a prompt containing a command to generate information on guiding product purchasing behavior, which is used to guide customer purchasing behavior at a target store, from a large-scale language model that has been provided with information on the business operations of the target store, and generates the information on guiding product purchasing behavior. An output unit outputs the information related to guiding the purchasing behavior of the product generated by the generation unit in a predetermined format, A generating apparatus characterized by having the following features.

2. The generating unit is For a large-scale language model that is provided with prior knowledge of the store's operations, including at least one of the following: product inventory information, product sales history, past shelf layout information, customer dwell time at the store, customer visits to the store, and weather information related to the store, The system generates shelf allocation information for use in arranging product displays, including the product name, the product's location, and its selling price, by inputting a command using a prompt expressed in natural language text, thereby generating the shelf allocation information as information for inducing purchasing behavior for the product. The generating apparatus according to feature 1.

3. The output unit is, Based on the shelf layout information generated by the generation unit, Information showing the aforementioned product name, the location where the aforementioned product is placed, and the sales price of the aforementioned product in a list format, And, Information showing the placement location of the product on the target display shelf, superimposed on the layout of the target display shelf. At least one of the above is output to the person in charge of shelf arrangement as information that guides the purchasing behavior of the said product. The generating apparatus according to feature 2.

4. The output unit is, Based on the shelf layout information generated by the generation unit, Information displayed side by side: the product name, the location where the product is placed, and the selling price of the product. And, The product name, the sales price of the product, and the route to the location where the product is placed are plotted on a map related to the target store. Route information including at least one of the above is output to customers visiting the target store as information for inducing purchasing behavior of the said product. The generating apparatus according to feature 2.

5. The generating unit is For a large-scale language model provided with prior knowledge consisting of inventory information related to inventory products and posts related to said inventory products that have generated a predetermined number of impressions on the internet, Based on reading the identifier displayed at the target store, a command is input using the prompt expressed in natural language text to generate posting information including the product name of the stocked product, a description of the stocked product, the sales location of the stocked product, and the content of the post related to the stocked product. The posted information is used to generate information that guides purchasing behavior for the product. The generating apparatus according to feature 1.

6. The generating unit is Prohibition of generating previously generated post content. And, Prohibition of creating postings related to the aforementioned product that are identical or similar to postings that have previously received specific criticism on the internet. Using the prompt which includes constraints that include at least one of the above, the posting information is generated. The generating apparatus according to feature 5.

7. The system further includes a provisioning unit that provides the large-scale language model with, as prior knowledge, the content of posts related to the inventory products generated by the generation unit that have generated a predetermined number of impressions on the internet. The generating apparatus according to feature 5 or 6.

8. The generating unit is As the aforementioned large-scale language model, at least one of the following is used: a large-scale language model possessing general knowledge, and a predetermined large-scale language model on which adapter tuning is performed. The generating apparatus according to any one of claims 1 to 6.

9. A generation method to be executed by a generation device, A generation process involves inputting a prompt containing a command to generate information on guiding product purchasing behavior, which is used to guide customer purchasing behavior at the target store, to a large-scale language model that has been provided with information on the business operations of the target store, and generating the information on guiding product purchasing behavior. An output step that outputs the information related to guiding purchasing behavior of the product generated in the generation step in a predetermined format, A method for generating a product, characterized by including the following:

10. A generation step involves inputting a prompt containing a command to generate information on product purchasing behavior used to guide customer purchasing behavior at a target store, to a large-scale language model provided with information on the store's operations, and generating the information on product purchasing behavior. An output step that outputs the information related to guiding purchasing behavior of the product generated in the generation step in a predetermined format, A generation program that causes a computer to execute something.

Images