System

Abstract

To provide a system.SOLUTION: A system includes: means for acquiring track record data of promotion measures; means for analyzing the track record data using generative AI and for proposing measures based on the analysis result; and means for proposing an improvement action based on the measures.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] There is a need to speed up the PDCA cycle of promotional measures and reduce the amount of work required. However, with conventional methods, it is difficult to shorten the lead time between analyzing advertising creatives and market trends, proposing hypotheses and countermeasures, and taking action to improve. [Means for solving the problem]

[0005] We provide a means to accelerate the PDCA cycle of promotional measures using generative AI. Specifically, generative AI reads acquisition results and analyzes them in comparison with advertising creatives and market trends. Based on the analysis results, we propose hypotheses and measures, shortening the lead time until improvement actions are taken. This will result in a reduction of labor hours of 10 hours / month x 12 months x 4 measures = 480 hours / year. [Brief explanation of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment Example 2. [Figure 15]FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, the processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0011] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0012] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0018] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0022] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0023] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0024] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] The present invention is a system that uses generative AI to accelerate the PDCA cycle of promotional initiatives. Specifically, the generative AI reads in acquisition results and analyzes them against advertising creatives and market trends. Based on the analysis results, it proposes hypotheses and measures, shortening the lead time until improvement actions are taken. This achieves a reduction in labor costs of 10 hours / month x 12 months x 4 initiatives = 480 hours / year.

[0029] "Example 2"

[0030] As a concrete example, generative AI reads performance data from past ad creatives and market trend data. Based on this data, it proposes the direction of the next ad creative and measures to address market trends. The proposed measures are implemented as part of the PDCA cycle, and the results are fed back to the generative AI. This speeds up the PDCA cycle and reduces labor costs.

[0031] "Example 3"

[0032] Furthermore, generative AI predicts the effectiveness of proposed measures and prioritizes improvement actions based on these predictions, significantly shortening lead times and enabling faster PDCA cycles.

[0033] The processing flow of each embodiment will be described below.

[0034] "Example 1"

[0035] Step 1: Load the generative AI with past achievements, such as acquisitions. This allows the AI ​​to learn based on past data.

[0036] Step 2: Analyze against ad creative and market trends. This analysis will help you understand the current market situation and how your ads are performing.

[0037] Step 3: Based on the analysis results, hypotheses and countermeasures are proposed, which then lead to the creation of the next action plan.

[0038] Step 4: Reduce the lead time to improvement actions, which speeds up the PDCA cycle and reduces man-hours.

[0039] "Example 2"

[0040] Step 1: The generative AI reads past performance data of advertising creatives and market trend data, allowing the AI ​​to learn from past data.

[0041] Step 2: Based on this data, the generative AI proposes the next direction for advertising creative and measures to address market trends. This allows the next action plan to be formulated. Step 3: The proposed measures are implemented as part of the PDCA cycle, and the results are fed back to the generative AI. This speeds up the PDCA cycle and reduces labor costs.

[0042] "Example 3"

[0043] Step 1: Generative AI predicts the effectiveness of the proposed measures, which in turn predicts the effectiveness of the next action plan.

[0044] Step 2: The generative AI prioritizes improvement actions based on the prediction results, enabling the most effective actions to be implemented first.

[0045] Step 3: This significantly reduces lead times and enables a faster PDCA cycle.

[0046] Example 1

[0047] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0048] The traditional plan-do-check-act (PDCA) cycle for promotional measures required a great deal of time and effort for each step, such as collecting data, analyzing it, formulating hypotheses, and implementing measures, making it difficult to respond quickly. Furthermore, data cleansing and analysis required specialized knowledge, making efficient operation difficult. This delayed prompt improvement actions to maximize the effectiveness of promotional measures, potentially resulting in a decline in market competitiveness.

[0049] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0050] In this invention, the server includes means for accelerating the planning, execution, evaluation, and improvement cycle of promotional measures using generative artificial intelligence, means for loading performance data, means for analyzing the data against advertising materials and market trends, means for cleansing the performance data, filling in missing values, and removing outliers, means for analyzing the performance data using generative artificial intelligence and identifying effective patterns and trends, means for proposing specific measures based on hypotheses, and means for presenting specific improvement actions based on the proposed measures. This enables the PDCA cycle of promotional measures to be executed quickly and efficiently, reducing labor hours and improving market competitiveness.

[0051] "Generative AI" refers to AI that has the ability to generate new information and patterns based on data.

[0052] "Promotional measures" refer to plans and activities to increase awareness of and promote sales of products and services.

[0053] The "Plan-Do-Check-Act cycle (PDCA cycle)" is a method for achieving continuous improvement in business processes and projects by repeating the four steps of Plan, Do, Check, and Act.

[0054] "Performance data" refers to data obtained as a result of past activities and measures, including the number of clicks, conversion rates, and sales.

[0055] "Advertising materials" refers to the visuals, text, and other creative elements used in an advertising campaign.

[0056] "Market trends" refers to current market conditions and trends, such as consumer behavior and competitor actions in a particular market.

[0057] "Cleansing" refers to the process of preparing data for analysis, such as filling in missing values ​​and removing outliers.

[0058] "Patterns and trends" refer to certain regularities or characteristics found in data.

[0059] A "hypothesis" is a prediction or assumption made based on data analysis.

[0060] "Countermeasures" refer to specific actions or measures implemented based on a hypothesis.

[0061] "Improvement actions" refer to specific improvement measures implemented based on analysis results and hypotheses.

[0062] This invention is a system that uses generative artificial intelligence to accelerate the plan-do-check-act (PDCA) cycle of promotional measures. Specific embodiments of this system are described below.

[0063] System configuration

[0064] Hardware and Software Configuration

[0065] The server provides computational resources for running a generative artificial intelligence model (e.g., OpenAI's GPT-4®). The server includes a database, a data processing unit, and a generative artificial intelligence model.

[0066] The terminal provides an interface for users to access the system, upload data, check analysis results, etc. Terminals include PCs, tablets, smartphones, etc.

[0067] Program processing

[0068] Data collection and preprocessing

[0069] Users upload promotional campaign performance data into the system in CSV file format, including the number of clicks, conversion rate, sales, and other data.

[0070] The server receives the uploaded data and stores it in a database, where it is cleansed, missing values ​​are filled, and outliers are removed.

[0071] Analyzing the data

[0072] The server uses a generative artificial intelligence model to analyze performance data, specifically matching it with advertising materials and market trends to identify effective patterns and trends.

[0073] The server identifies that particular advertising materials have demonstrated high conversion rates in particular market segments.

[0074] Hypothesis and proposed measures

[0075] The server generates a hypothesis based on the analysis results, for example, "Advertising material B shows a high conversion rate among young people."

[0076] The server then proposes specific measures based on the hypothesis, such as "Use advertising material B, which targets younger generations, in the next campaign."

[0077] Suggestion of improvement actions

[0078] The server then presents specific improvement actions to the user based on the proposed measures, such as "Use advertising material B in the next campaign and target younger generations."

[0079] The user checks the proposed improvement actions and reflects them in the next promotional measures. Specifically, the user plans and executes a new campaign using advertising material B.

[0080] Examples of concrete examples and prompts

[0081] Specific examples

[0082] User-Provided Data:

[0083] Ad Material A: 5,000 clicks, 2.5% conversion rate

[0084] Ad Material B: 3,000 clicks, conversion rate 3.0%

[0085] Market Trend Data: Online Shopping Increases Among Young People

[0086] Prompt Sentence Examples

[0087] "Based on the performance data of advertising campaigns over the past six months, please analyze the most effective advertising materials and market trends, and propose hypotheses and measures for the next campaign."

[0088] In this way, the server can propose quick and specific improvement actions to the user, accelerating the PDCA cycle.

[0089] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0090] Step 1:

[0091] Collection of performance data

[0092] Users upload promotional campaign performance data into the system in CSV file format.

[0093] Input: A CSV file containing performance data such as clicks, conversion rates, and sales.

[0094] The server receives the uploaded CSV file and stores it in a database.

[0095] Output: Actual data stored in a database.

[0096] Step 2:

[0097] Data Preprocessing

[0098] The server cleanses the stored performance data, specifically by filling in missing values ​​and removing outliers.

[0099] Input: Actual data stored in the database.

[0100] The server will fill in missing values ​​with the average value, remove outliers, and standardize the data format. For example, it will standardize the date format to "YYYY-MM-DD."

[0101] Output: Cleansed actual data.

[0102] Step 3:

[0103] Analyzing the data

[0104] The server analyzes the performance data using a generative artificial intelligence model (e.g., GPT-4).

[0105] Input: Cleansed actual data.

[0106] The server compares advertising materials and market trends with performance data to identify effective patterns and trends, such as identifying whether certain advertising materials result in higher conversion rates among certain market segments.

[0107] Output: Effective patterns and trends as a result of the analysis.

[0108] Step 4:

[0109] Hypothesis and proposed measures

[0110] The server generates a hypothesis based on the analysis results, for example, "Advertising material B shows a high conversion rate among young people."

[0111] Input: Effective patterns and trends as a result of analysis.

[0112] The server then proposes specific measures based on the hypothesis, such as "Use advertising material B, which targets younger generations, in the next campaign."

[0113] Output: Hypothesis and specific measures.

[0114] Step 5:

[0115] Suggestion of improvement actions

[0116] The server then presents specific improvement actions to the user based on the proposed measures, such as "Use advertising material B in the next campaign and target younger generations."

[0117] Input: Hypothesis and specific measures.

[0118] The server generates and presents remedial actions to the user.

[0119] Output: Specific improvement actions presented to the user.

[0120] Step 6:

[0121] Taking improvement actions

[0122] The user checks the proposed improvement actions and reflects them in the next promotional measures. Specifically, the user plans and executes a new campaign using advertising material B.

[0123] Input: The specific improvement actions suggested to the user.

[0124] The user runs a new campaign based on the improvement actions.

[0125] Output: The new campaign that was executed.

[0126] (Application example 1)

[0127] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0128] Traditional promotional measures require a great deal of time and effort in each phase of the PDCA cycle, and require rapid response, particularly in analyzing advertising creatives and market trends, formulating hypotheses, and implementing measures. However, there is a lack of systems to efficiently carry out these processes, which results in long lead times for improvement actions, making it difficult to implement effective promotional measures. In addition, there is no way to monitor advertising campaign performance in real time and quickly implement improvement actions, which makes it difficult to maximize advertising effectiveness.

[0129] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0130] In this invention, the server includes a means for accelerating the PDCA cycle of promotional measures using generative AI, a means for inputting results such as acquisitions into the generative AI, and a means for analyzing the results in comparison with advertising creatives and market trends. This makes it possible to propose hypotheses and measures based on the analysis results, shortening the lead time until improvement actions are taken. In addition, by including a means for monitoring the performance of advertising campaigns in real time and quickly taking improvement actions as an application installed on a smartphone, it is possible to maximize advertising effectiveness.

[0131] "Generative AI" refers to artificial intelligence that generates new information and content based on data.

[0132] "Promotional measures" are plans and activities to increase awareness of and promote sales of a product or service.

[0133] The PDCA cycle is a method for continuously improving business processes and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0134] "Results" refers to the data and results obtained as a result of past activities and measures.

[0135] "Advertising creative" refers to the elements that make up an advertisement, such as its content, design, and message.

[0136] "Market trends" refer to changes and trends in consumer preferences and behavior, competitor trends, etc. in a particular market.

[0137] A hypothesis is a prediction or premise made to find an explanation or solution to a phenomenon or problem.

[0138] "Countermeasures" refer to specific actions or measures taken to address a specific problem or issue.

[0139] "Lead time" is the time it takes from the start to the end of a process.

[0140] A "smartphone" is a highly functional mobile phone that can connect to the Internet and use applications.

[0141] An "application" is software designed to provide a specific function or service.

[0142] "Performance" refers to the results or effects of a particular activity or measure.

[0143] "Real-time" refers to the ability to process and display ongoing events and data immediately.

[0144] An "improvement action" is a specific improvement measure implemented to solve a specific problem or issue.

[0145] The system for carrying out this invention uses generative AI to accelerate the PDCA cycle of promotional measures. A specific embodiment of this system will be described below.

[0146] System Configuration

[0147] The system includes a server, a smartphone, a generative AI model, a database, and an internet connection. The server runs the generative AI model and accesses the database to retrieve advertising performance data. The smartphone runs an application that allows users to monitor advertising campaign performance in real time and take action to improve.

[0148] Program processing

[0149] The server processes the data in the following steps:

[0150] 1. Data loading:

[0151] The server reads advertising performance data from the database, including click-through rates (CTR), conversion rates (CVR), advertising costs, etc.

[0152] 2. Data Analysis:

[0153] The server uses a generative AI model to analyze advertising performance data, specifically by comparing it with advertising creative and market trends, and then propose hypotheses and countermeasures.

[0154] 3. Improvement suggestions:

[0155] The server uses a generative AI model to suggest improvement actions based on the analysis results, which are then communicated to the user via a smartphone application.

[0156] Hardware and software used

[0157] Hardware: Servers, smartphones

[0158] Software: Python, OpenAI API, Pandas

[0159] Data processing and calculation

[0160] The server uses Python to read advertising performance data from the database and processes the data using Pandas. It then analyzes the data using a generative AI model (OpenAI API) to generate hypotheses and countermeasures. The generated hypotheses and countermeasures are then notified to the user via a smartphone application.

[0161] Specific examples

[0162] For example, assume that the following advertising performance data exists:

[0163] Click-through rate: 2.5%

[0164] Conversion rate: 1.2%

[0165] Advertising cost: 500,000 yen

[0166] Based on this data, the server inputs the following prompt sentence into the generative AI model:

[0167] Example prompt sentence:

[0168] Based on the advertising performance data below, please analyze the advertising creative and market trends and suggest areas for improvement.

[0169] Advertising performance data:

[0170] Click-through rate: 2.5%

[0171] Conversion rate: 1.2%

[0172] Advertising cost: 500,000 yen

[0173] Based on this prompt, the generative AI model analyzes advertising creatives and market trends and suggests improvements. The proposed improvements are notified to the user via a smartphone application, allowing them to quickly take action.

[0174] In this way, the system can accelerate the PDCA cycle of promotional initiatives and maximize the performance of advertising campaigns.

[0175] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0176] Step 1:

[0177] The server reads ad performance data from the database. As input, it requires database connection information and a query. The server executes the query to obtain data such as click-through rate (CTR), conversion rate (CVR), and ad cost, and stores it in a Pandas data frame. The output is a data frame containing ad performance data.

[0178] Step 2:

[0179] The server generates a prompt to input the acquired advertising performance data into the generative AI model. Advertising performance data is required as input. The server converts the data frame into text format and generates a prompt like the following: "Based on the following advertising performance data, please analyze the advertising creative and market trends and suggest improvements. Advertising performance data: Click-through rate: 2.5%, Conversion rate: 1.2%, Advertising cost: 500,000 yen." The output is the generated prompt.

[0180] Step 3:

[0181] The server uses a generative AI model (OpenAI API) to input a prompt text and analyze ad creatives and market trends. The prompt text is required as input. The server sends the prompt text to the OpenAI API and receives a response from the generative AI model. The output is the analysis results from the generative AI model.

[0182] Step 4:

[0183] The server generates hypotheses and proposed countermeasures based on the analysis results from the generative AI model. The analysis results from the generative AI model are required as input. The server analyzes the analysis results and proposes specific improvement actions. The output is proposed hypotheses and countermeasures.

[0184] Step 5:

[0185] The server notifies the smartphone application of the proposed hypotheses and countermeasures. The proposed hypotheses and countermeasures are required as input. The server sends a notification to the smartphone application, prompting the user to take an improvement action. The output is the improvement action notified to the user.

[0186] Step 6:

[0187] The user checks the notified improvement actions through a smartphone application and executes them as necessary. The input requires a notification from the server. The user checks the notification and executes the improvement action according to the application's instructions. The output is the executed improvement action.

[0188] In this way, the system can accelerate the PDCA cycle of promotional initiatives and maximize the performance of advertising campaigns.

[0189] Example 2

[0190] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0191] Conventional promotional measures have the problem that each step of the PDCA cycle takes a lot of time and effort, making it difficult to make quick improvements. In addition, it was difficult to effectively utilize performance data on advertising creatives and market trend data, making it difficult to propose optimal measures. This resulted in the promotional effect not being maximized and resources being wasted.

[0192] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0193] In this invention, the server includes: a means for a user to upload past advertising creative performance data and market trend data to the server; a means for the server to save the uploaded data in a database and perform preprocessing; a means for the server to input the preprocessed data into a generative AI model and send a prompt sentence to the generative AI model; a means for the server to generate proposals and return them to the server; a means for the server to display the proposals to the user; a means for the user to implement the proposed measures and feed back the results to the server; and a means for the server to re-input the feedback data into the generative AI model and reflect it in the next proposal. This speeds up the PDCA cycle, reduces labor hours, and enables optimal measures to be proposed that effectively utilize advertising creative performance data and market trend data.

[0194] "Generative AI" is artificial intelligence that makes new suggestions and predictions based on past data and market trends.

[0195] The PDCA cycle is a method for continuously improving business operations and processes by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0196] "Promotional measures" are specific activities and strategies for increasing awareness of products and services and promoting sales.

[0197] "Performance data" refers to numerical data that indicates the effectiveness of advertising creatives and promotional measures, and includes the number of clicks, conversion rates, costs, etc.

[0198] "Market trend data" is data that shows current market trends and consumer behavior patterns.

[0199] A "server" is a computer system that stores, processes, and manages data.

[0200] A "database" is a system for efficiently storing, searching, and managing data.

[0201] "Preprocessing" refers to the process of converting data into a format suitable for analysis or model input.

[0202] A "generative AI model" is an artificial intelligence model that generates new suggestions and predictions based on input data.

[0203] A "prompt" is text that gives specific instructions or questions to a generative AI model.

[0204] The "proposal content" refers to the new advertising creative direction and measures for market trends generated by the generative AI model.

[0205] "Feedback" refers to data that is re-entered into the system to reflect the results of the measures implemented in the next proposal.

[0206] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures and reduce man-hours. This system consists of multiple components, including a user, a server, and a generative AI model.

[0207] First, the user uploads past advertising creative performance data and market trend data to the server. This data is generally provided in CSV or Excel file format. The server then stores the uploaded data in a database. For the database, a relational database management system (RDBMS) such as MySQL (registered trademark) or PostgreSQL is used.

[0208] The server then preprocesses the stored data. This includes imputing missing values ​​and normalizing the data. These processes can be performed using the Python Pandas library. The preprocessed data is then converted into a format that is easy for the generative AI model to understand. Specifically, it is often converted into JSON format.

[0209] The server inputs the preprocessed data into a generative AI model. For example, natural language processing (NLP) models such as GPT-4 or BERT are used as generative AI models. The server inputs a prompt to the generative AI model. An example of a prompt is, "Based on the performance data of advertising creatives over the past six months and current market trend data, please suggest a direction for the next advertising campaign."

[0210] Based on the input data and prompt text, the generative AI model proposes the next direction for advertising creative and measures to address market trends. The generated proposals are returned to the server in text format. The server receives the proposals returned by the generative AI model and displays them to the user. For display, data visualization tools such as Tableau or Power BI are used.

[0211] The user implements the proposed measures and feeds the results back to the server. The feedback data is then saved in the database and reflected in the next proposal. The server then inputs the feedback data back into the generative AI model, accelerating the PDCA cycle.

[0212] This system will speed up the PDCA cycle, reduce labor costs, and enable the proposal of optimal measures that effectively utilize advertising creative performance data and market trend data.

[0213] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0214] Step 1:

[0215] Users upload historical advertising creative performance data and market trend data to the server.

[0216] Input: Past ad creative performance data (e.g., CSV file), market trend data (e.g., Excel file)

[0217] Output: Data stored on the server

[0218] Specific operation: A user selects a CSV or Excel file through the web interface and clicks the upload button. The server receives the uploaded file and stores it in the database.

[0219] Step 2:

[0220] The server stores the uploaded data in a database and performs preprocessing.

[0221] Input: Uploaded data (CSV file, Excel file)

[0222] Output: Preprocessed data (JSON format)

[0223] How it works: The server reads the data using Python's Pandas library, imputes missing values, normalizes the data, and converts the preprocessed data into a JSON format that is easy for the generative AI model to understand.

[0224] Step 3:

[0225] The server inputs the preprocessed data into the generative AI model and sends the prompt sentence to the generative AI model.

[0226] Input: Preprocessed data (JSON format), prompt statement

[0227] Output: Proposals from the generative AI model (text format)

[0228] Specific operation: The server calls the generative AI model's API and sends the preprocessed data and a prompt. An example of a prompt is, "Based on the performance data of the ad creative over the past six months and current market trend data, please suggest the direction of our next ad campaign."

[0229] Step 4:

[0230] Based on the input data and prompt text, the generative AI model suggests the next direction for advertising creative and measures to address market trends.

[0231] Input: Preprocessed data (JSON format), prompt statement

[0232] Output: Proposal content (text format)

[0233] How it works: The generative AI model uses its internal algorithms to analyze data and generate optimal ad creative direction and countermeasures for market trends. The generated proposals are returned to the server in text format.

[0234] Step 5:

[0235] The server receives the suggestions returned by the generative AI model and displays them to the user.

[0236] Input: Proposal content from the generative AI model (text format)

[0237] Output: The suggestions displayed to the user

[0238] Specific operation: The server receives the proposals, creates a dashboard using a data visualization tool (e.g., Tableau, Power BI), and displays it to the user.

[0239] Step 6:

[0240] The user implements the proposed measures and feeds back the results to the server.

[0241] Input: Implementation result data (e.g. Excel file)

[0242] Output: Feedback data stored on the server

[0243] Specific operation: The user compiles the results of the implemented measures in an Excel file and uploads it to the server. The server receives the uploaded file and stores it in a database.

[0244] Step 7:

[0245] The server then inputs the feedback data back into the generative AI model and reflects it in the next proposal.

[0246] Input: Feedback data (JSON format)

[0247] Output: Data reflected in the next proposal

[0248] Specific operation: The server preprocesses the feedback data and re-inputs it into the generative AI model, which then takes the feedback data into account to generate the next proposal.

[0249] (Application example 2)

[0250] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0251] Conventional promotional measures had the problem of taking a lot of time and effort to analyze advertising creative performance data and market trend data and propose hypotheses and countermeasures. There was also a lack of means to speed up the PDCA cycle and efficiently implement improvement actions. Furthermore, there was no established process for efficient data input using smart devices, or for implementing and providing feedback on proposals using generative AI models.

[0252] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for accelerating the PDCA cycle of promotional measures using a generative AI; means for loading the generative AI with actual results such as acquisitions; means for analyzing the results in comparison with advertising creatives and market trends; means for proposing hypotheses and measures based on the analysis results; means for shortening the lead time to improvement actions; means for inputting performance data and market trend data of past advertising creatives using an application installed on a smart device and inputting prompt statements into the generative AI model; and means for implementing measures in response to the advertising creative direction and market trends proposed by the generative AI model and providing feedback on the results. This makes it possible to speed up the PDCA cycle of promotional measures and significantly reduce labor costs.

[0253] "Generative AI" is artificial intelligence that generates new advertising creatives and strategies based on past data and market trends.

[0254] "Promotional measures" are plans and activities to increase awareness of products and services and promote sales.

[0255] The PDCA cycle is a method for continuously improving business operations by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0256] "Performance data" refers to data used to measure the effectiveness of advertising creatives, including click-through rates and conversion rates.

[0257] "Market trend data" is data that indicates current market trends and consumer interests.

[0258] A "smart device" is a portable electronic device that can connect to the Internet, including smartphones and smart glasses.

[0259] A "prompt" is an instruction given to generative AI, and serves as the basis for generating the direction and measures for advertising creatives.

[0260] "Feedback" is the process of inputting the results of implemented measures back into the generative AI and reflecting them in the next proposal.

[0261] The system for carrying out this invention uses generative AI to accelerate the PDCA cycle of promotional measures. A specific embodiment of this system will be described below.

[0262] Hardware and software used

[0263] Hardware: Smartphones, smart glasses, servers

[0264] Software: Generative AI models (e.g., OpenAI GPT-4), database management systems (e.g., MySQL, PostgreSQL), data analysis tools (e.g., Tableau, Google Analytics)

[0265] Data processing and calculation

[0266] 1. Data Collection:

[0267] The server retrieves historical performance data of advertising creatives from a database, and also retrieves market trend data from an external API (e.g., Google Trends API).

[0268] 2. Data Preprocessing:

[0269] The server cleans the acquired data, imputes missing values, normalizes the data, and converts it into a format suitable for analysis.

[0270] 3. Data Analysis:

[0271] The server uses data analysis tools to evaluate the performance of past advertising creatives and analyzes market trend data to identify current trends.

[0272] 4. Utilizing generative AI models:

[0273] The server inputs the following prompt sentences into the generative AI model to generate the next direction for advertising creative and measures for market trends.

[0274] Prompt Sentence Examples

[0275] Based on past ad creative performance data and current market trend data, please propose the next ad creative direction and measures to address market trends.

[0276] 5. Implementation and feedback of proposals:

[0277] Using an application installed on a smart device, users can implement the advertising creative direction and market trend measures suggested by the generative AI model. The results are then fed back into the generative AI model, accelerating the PDCA cycle.

[0278] Specific examples

[0279] Example 1: Smartphone application

[0280] Users open the smartphone app and input past ad creative performance data and market trend data. The app uses a generative AI model to suggest the direction of the next ad creative. Users can then review and implement the suggestions, completing the PDCA cycle.

[0281] Example 2: Smart Glasses Application

[0282] The marketer creating the ad creative wears the smart glasses and uses voice commands to access past performance data and market trend data. The smart glasses use a generative AI model to suggest the direction of the next ad creative in real time. The marketer then reviews the suggestions and implements them, completing the PDCA cycle.

[0283] In this way, it is possible to speed up the PDCA cycle of promotional measures and significantly reduce labor costs.

[0284] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0285] Step 1:

[0286] The server retrieves past performance data of ad creatives from a database. The inputs are database connection information and a query, and the output is performance data of ad creatives. Specifically, the server connects to the database and executes the specified query to retrieve the data.

[0287] Step 2:

[0288] The server retrieves market trend data from an external API. The input requires the API endpoint and authentication information, and the output is market trend data. Specifically, the server sends a request to the API and receives market trend data as a response.

[0289] Step 3:

[0290] The server cleans the acquired data and fills in missing values. It takes advertising creative performance data and market trend data as input, and obtains cleaned data as output. Specifically, the server identifies missing values ​​in the data and fills them in appropriately.

[0291] Step 4:

[0292] The server normalizes the data and converts it into a format suitable for analysis. It takes cleaned data as input and normalized data as output. Specifically, the server scales and encodes the data.

[0293] Step 5:

[0294] The server uses a data analysis tool to evaluate the performance of past advertising creatives. Normalized data is required as input, and performance evaluation results are obtained as output. Specifically, the server inputs data into the data analysis tool and obtains the evaluation results.

[0295] Step 6:

[0296] The server analyzes the market trend data to identify the current trend. The input required is normalized market trend data, and the output is trend analysis results. Specifically, the server uses data analysis tools to identify the trend.

[0297] Step 7:

[0298] The server inputs prompt statements into the generative AI model and generates the direction of the next ad creative and measures for market trends. The required inputs are performance evaluation results, trend analysis results, and prompt statements, and the output is the direction of the generated ad creative and measures. Specifically, the server inputs prompt statements into the generative AI model and obtains the generated results.

[0299] Step 8:

[0300] Using an application installed on a smart device, users implement the advertising creative direction and market trend countermeasures suggested by the generative AI model. The inputs are the generated advertising creative direction and countermeasures, and the output is the implementation results. Specifically, users operate the application to implement the suggested countermeasures.

[0301] Step 9:

[0302] The user then feeds back the results of the measures they have implemented to the generative AI model. The input is the implementation results, and the output is the fed-back data. Specifically, the user uses an application to input the implementation results and feeds them back to the generative AI model.

[0303] Example 3

[0304] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0305] The PDCA cycle for conventional promotional measures had long lead times and made it difficult to take prompt action to improve. Furthermore, there was a lack of concrete methods for effectively utilizing the analysis results using generative AI, making it difficult to reduce man-hours and prioritize effective measures.

[0306] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes a means for the user to input a prompt sentence, a means for the terminal to send the prompt sentence to the server, a means for the server to preprocess the prompt sentence, a means for the server to input the prompt sentence to the generative AI model, a means for the generative AI model to predict the effectiveness of a countermeasure, a means for the server to receive the prediction result, a means for the server to analyze the prediction result and determine the priority of improvement actions, a means for the server to return the priority to the user, and a means for the user to execute the improvement action based on the priority. This enables shortening of lead time and realization of a rapid PDCA cycle.

[0307] "Generative AI" is an artificial intelligence technology that generates data based on user input and performs analysis and predictions.

[0308] "Promotional measures" are marketing activities aimed at increasing awareness of products and services and promoting sales.

[0309] The PDCA cycle is a method for achieving continuous improvement in business operations and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0310] "Lead time" is the time required from the start to the end of a process.

[0311] A "prompt" is text that contains instructions or questions that a user inputs to a generative AI.

[0312] A "server" is a computer system that provides services to other computers on a network.

[0313] A "terminal" is a device that a user uses to communicate with a server.

[0314] "Preprocessing" refers to the initial process of converting data into a format suitable for analysis or prediction.

[0315] A "generative AI model" is a specific implementation of generative AI, an algorithm that generates data and makes predictions based on user input.

[0316] "Prediction results" are the results of analysis and predictions that a generative AI model outputs based on user input.

[0317] "Improvement actions" are specific measures or policies implemented based on the prediction results.

[0318] "Priority" refers to the order of importance or urgency of execution among multiple improvement actions.

[0319] This invention relates to a system that uses generative AI to accelerate the PDCA cycle of promotional measures and shorten lead times. Specific embodiments of this system are described below.

[0320] First, the user uses a terminal to input a prompt. A prompt is text used to give instructions or ask questions to the generative AI. For example, a prompt such as "Please predict the effect of increasing sales if a new marketing strategy is implemented" can be input.

[0321] The terminal sends the prompt text entered by the user to the server. The HTTPS protocol is used for communication to ensure data security. The server then preprocesses the received prompt text. Specifically, it normalizes the text and removes unnecessary spaces.

[0322] Next, the server inputs the preprocessed prompt sentence into a generative AI model. The generative AI model runs on a high-performance server equipped with an NVIDIA GPU. For example, GPT-4 is used as the generative AI model. This generative AI model predicts the effectiveness of countermeasures based on the prompt sentence. For example, it predicts the effect of increasing sales based on past data and similar cases.

[0323] The server receives prediction results from the generative AI model. The prediction results are returned as numerical or text data. The server analyzes the prediction results and prioritizes improvement actions. For example, it ranks marketing strategies in order of their effectiveness in increasing sales.

[0324] The server returns the determined priorities to the user, who then displays them to the terminal. The user then takes action to improve the situation based on the priorities received from the server. For example, the user may first implement the marketing strategies predicted to be most effective.

[0325] This system significantly shortens lead times and enables a faster PDCA cycle. In addition, by effectively utilizing the analysis results using generative AI, it is possible to reduce man-hours and prioritize effective countermeasures.

[0326] As a concrete example, consider the case where a user inputs a prompt such as, "Please predict the sales increase effect of introducing a new marketing strategy." The device sends this prompt to the server, which then predicts the sales increase effect using a generative AI model. Based on the prediction results, the server determines the priority of the marketing strategies and returns it to the user. The user then implements the marketing strategy predicted to be most effective based on this priority.

[0327] In this way, the PDCA cycle of promotional measures utilizing generative AI can be accelerated and the lead time can be shortened. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0328] Step 1:

[0329] The user inputs a prompt sentence, such as "Please predict the sales increase effect of introducing a new marketing strategy" through the terminal interface. The input prompt sentence is temporarily stored in the terminal's memory.

[0330] Step 2:

[0331] The terminal sends the prompt text to the server. The terminal sends the prompt text entered by the user to the server using the HTTPS protocol. The input is the prompt text, and the output is a notification to the server that transmission has been completed.

[0332] Step 3:

[0333] The server preprocesses the prompt text. The server normalizes the received prompt text and removes unnecessary spaces and special characters. The input is the prompt text and the output is the preprocessed prompt text. Specific operations include text cleaning and tokenization.

[0334] Step 4:

[0335] The server inputs a prompt sentence to the generative AI model. The server then inputs a preprocessed prompt sentence to the generative AI model. The input is the preprocessed prompt sentence, and the output is a notification to the generative AI model that the input has been completed. Specific operations include API calls and sending data to the model.

[0336] Step 5:

[0337] The generative AI model predicts the effectiveness of countermeasures. The generative AI model predicts the effectiveness of countermeasures based on the prompt sentence. The input is the preprocessed prompt sentence, and the output is the predicted result. Specific operations include the model's inference process.

[0338] Step 6:

[0339] The server receives the prediction results. The server receives the prediction results from the generative AI model. The input is the prediction results, and the output is saving the prediction results to the server. Specific operations include saving the results to a database and analyzing the results.

[0340] Step 7:

[0341] The server analyzes the prediction results and determines the priorities of improvement actions. The server analyzes the prediction results and determines the priorities of improvement actions. The input is the prediction results and the output is a priority list. Specific operations include analysis using an algorithm and generation of rankings.

[0342] Step 8:

[0343] The server returns the priority to the user. The server returns the determined priority to the user. The input is a priority list, and the output is a notification to the user that transmission is complete. Specific operations include data format conversion and the transmission process.

[0344] Step 9:

[0345] The user executes improvement actions based on the priorities. The user executes improvement actions based on the priorities received from the server. The input is a priority list, and the output is the executed improvement actions. Specific operations include the introduction of marketing strategies and the execution of measures.

[0346] (Application example 3)

[0347] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0348] In conventional promotional measures and factory robot operations, there is a demand for accelerating the PDCA cycle and shortening lead times, but there is a lack of efficient ways to achieve this. Furthermore, there is no system in place that can collect factory robot operation data in real time, predict the effects of improvement measures using generative AI, and determine priorities, making it difficult to maximize production efficiency. A new system is needed to solve these issues.

[0349] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0350] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time to improvement actions, means for collecting operation data of robots in the factory in real time and predicting the effects of the proposed improvement measures using generative AI, and means for determining the priorities of improvement actions based on the prediction results and notifying the factory manager. This makes it possible to improve the efficiency of promotional measures and maximize the production efficiency of factory robots.

[0351] "Generative AI" refers to artificial intelligence that generates new information and predictions based on data.

[0352] The PDCA cycle is a method for improving business operations by repeating the four steps of Plan, Do, Check, and Act.

[0353] "Lead time" is the time it takes for a product or service to be delivered after it is ordered.

[0354] A "factory robot" is a mechanical device that performs work automatically in a factory.

[0355] "Operational data" refers to data that records the operating status of a machine or system.

[0356] "Improvement measures" refer to specific means or methods for solving current problems.

[0357] "Priority" refers to determining the order in which multiple items or tasks should be done first.

[0358] A "factory manager" is a position that involves the operation and management of a factory.

[0359] "Promotional measures" are specific activities or plans aimed at promoting products or services.

[0360] "Advertising creative" refers to the content, design, and expression of the advertisement.

[0361] A "market trend" is a change in consumer preferences and behavior in a particular market.

[0362] A "hypothesis" is a tentative explanation or prediction about a phenomenon or problem.

[0363] A "proposal" is a presentation of a solution or new idea to a problem.

[0364] The following system configuration will be described as an embodiment of the present invention.

[0365] The server uses generative AI to provide a means to accelerate the PDCA cycle of promotional measures. Specifically, the server loads performance data such as acquisition data into the generative AI, and analyzes it in comparison with advertising creatives and market trends. Based on the analysis results, the server proposes hypotheses and countermeasures, shortening the lead time until improvement actions are taken.

[0366] Furthermore, the server collects operational data from robots in the factory in real time and uses generative AI to predict the effectiveness of proposed improvement measures. Based on the prediction results, it determines the priority of improvement actions and provides a means to notify factory managers. This makes it possible to maximize factory production efficiency and shorten lead times.

[0367] This system uses the following hardware and software:

[0368] Hardware: Smartphone (iOS or ANDROID), factory robot

[0369] Software: Python, Pandas, Scikit-learn, Matplotlib

[0370] The server reads the factory robot's operation data from a CSV file and splits the data into features and targets. It splits the data into training data and test data, and uses a random forest regression model to learn the training data. It makes predictions on the test data and evaluates the predicted effects. It calculates the importance of features and determines the priority of improvement actions. It displays the results in a graph and notifies the factory manager.

[0371] As a specific example, assume that the operation data of a factory robot is in the following format:

[0372] time, speed, load, temperature, target

[0373] 0, 1.2, 50, 30, 0.8

[0374] 1, 1.5, 55, 32, 0.9

[0375] 2, 1.1, 52, 31, 0.85

[0376] ...

[0377] An example of a prompt sentence to input to the generative AI model is as follows:

[0378] "Based on the operation data of a factory robot, predict the effect of improvement measures and determine priorities. The data is in the following format: time, speed, load, temperature, target. Output the prediction results and priorities."

[0379] Using this prompt, the generative AI model can make appropriate predictions and prioritization decisions.

[0380] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0381] Step 1:

[0382] The server reads the factory robot's operation data from a CSV file. It requires a CSV file as input and obtains data in a data frame format as output. This data contains columns for time, speed, load, temperature, and target. Specifically, it reads the CSV file using Python's Pandas library.

[0383] Step 2:

[0384] The server splits the loaded data into features (X) and targets (y). Data in a data frame format is required as input, and a dataset of features and targets is obtained as output. Specifically, it separates the target column from the data frame and treats the remaining columns as features.

[0385] Step 3:

[0386] The server splits the features and targets into training data and test data. As input, it requires a dataset of features and targets, and as output, it obtains training data and test data. Specifically, it splits the data using the train_test_split function of Scikit-learn.

[0387] Step 4:

[0388] The server uses a random forest regression model to learn the training data. The training data is required as input, and the trained model is obtained as output. Specifically, the model is created using Scikit-learn's RandomForestRegressor class, and training is performed using the fit method.

[0389] Step 5:

[0390] The server uses the trained model to make predictions on test data. Test data is required as input, and prediction results are obtained as output. Specific operations include making predictions using the predict method.

[0391] Step 6:

[0392] The server evaluates the effect prediction by comparing the predicted results with the actual values. The predicted results and actual values ​​are required as input, and an evaluation index (e.g., mean squared error) is obtained as output. Specifically, the evaluation is performed using the mean_squared_error function of Scikit-learn.

[0393] Step 7:

[0394] The server calculates the importance of features and determines the priority of improvement actions. A trained model is required as input, and the importance and ranking of features are obtained as output. Specifically, the server obtains the importance using the feature_importances_ attribute of the model and determines the ranking based on the importance.

[0395] Step 8:

[0396] The server displays the results in a graph and notifies the factory manager. The input required is the feature importance and its ranking, and the output is a visual graph and a notification message. Specifically, the server creates the graph using the Matplotlib library and sends the results to the factory manager through a notification system.

[0397] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0398] "Example 1"

[0399] One embodiment of the present invention is a system that combines generative AI and an emotion engine. This system recognizes a user's emotions and proposes promotional measures based on those emotions. Specifically, if the system recognizes that the user is feeling joy or excitement, it proposes promotional measures such as new product release information or special discount information to the user. If the system recognizes that the user is feeling dissatisfied or anger, it proposes promotional measures such as support information for resolving the user's problem or an apology message to the user.

[0400] "Example 2"

[0401] Another embodiment of the present invention is a system that suggests improvement actions according to a user's emotions based on the emotion recognition results of an emotion engine. Specifically, if it is recognized that a user is feeling confused or perplexed, it suggests improvement actions such as providing the user with information on how to use a product or details of a service. Also, if it is recognized that a user is feeling relieved or satisfied, it suggests improvement actions such as suggesting the user to purchase an additional product or upgrade a service.

[0402] "Example 3"

[0403] Furthermore, another embodiment of the present invention is a system that generates advertising creatives according to a user's emotions based on the emotion recognition results of an emotion engine. Specifically, if it is recognized that the user is feeling joy or excitement, it generates advertising creatives using bright colors and lively music that enhance the user's emotions. On the other hand, if it is recognized that the user is feeling relief or satisfaction, it generates advertising creatives using calm colors and tranquil music that deepen the user's emotions.

[0404] The processing flow of each embodiment will be described below.

[0405] "Example 1"

[0406] Step 1: Generative AI collects user behavior data.

[0407] Step 2: Based on the collected data, the generative AI predicts the user's emotions.

[0408] Step 3: The emotion engine receives the predictions from the generative AI and recognizes the user's emotions.

[0409] Step 4: Based on the emotions recognized by the emotion engine, appropriate promotional measures are proposed.

[0410] "Example 2"

[0411] Step 1: Generative AI collects user behavior data.

[0412] Step 2: Based on the collected data, the generative AI predicts the user's emotions.

[0413] Step 3: The emotion engine receives the predictions from the generative AI and recognizes the user's emotions.

[0414] Step 4: The emotion engine suggests appropriate improvement actions based on the emotions it recognizes. "Example 3"

[0415] Step 1: Generative AI collects user behavior data.

[0416] Step 2: Based on the collected data, the generative AI predicts the user's emotions.

[0417] Step 3: The emotion engine receives the predictions from the generative AI and recognizes the user's emotions.

[0418] Step 4: The emotion engine generates appropriate ad creatives based on the emotions it recognizes.

[0419] Example 1

[0420] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0421] Conventional promotional measures require a lot of time and effort for data analysis, hypothesis formulation, and countermeasure proposals at each phase of the PDCA cycle, making it difficult to take prompt action to improve. Furthermore, promotional measures do not take into account user emotions, which means that user satisfaction cannot be increased.

[0422] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0423] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for inputting results such as acquisitions into the generative AI, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, and means for recognizing user emotions and proposing promotional measures according to those emotions. This enables the implementation of prompt and effective promotional measures and appropriate responses according to user emotions.

[0424] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[0425] The PDCA cycle is a business improvement method consisting of four phases: Plan, Do, Check, and Act.

[0426] "Results" are data and results obtained as a result of past activities and measures.

[0427] "Advertising creative" is a general term that refers to elements such as the content, design, and message of an advertisement.

[0428] "Market trends" refer to changes in consumer preferences and behavior, trends in fashion, etc. in a particular market.

[0429] A "hypothesis" is an explanation or proposal based on a prediction or conjecture about a particular phenomenon or problem.

[0430] "Countermeasures" are specific actions or measures taken to address a specific problem or issue.

[0431] "Lead time" is the time it takes from the start to the end of a process.

[0432] "Emotion recognition" is a technology that analyzes and identifies emotions from a user's facial expressions, voice, etc.

[0433] "Promotional measures" are specific activities and strategies aimed at promoting products and services.

[0434] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures. This system works in cooperation with three parties: a server, a terminal, and a user.

[0435] Hardware and software used

[0436] Hardware: Servers, terminals

[0437] Software: Generative AI models (e.g., GPT-4), emotion engines (e.g., Affectiva)

[0438] Data collection and preprocessing

[0439] The server retrieves user performance data (e.g., ad clicks, purchase history) from a database. It also retrieves data on ad creatives and market trends from external APIs. The retrieved data is cleansed and missing values ​​and outliers are processed. The data is standardized and converted into a format suitable for generative AI models.

[0440] Data analysis

[0441] The server then inputs the preprocessed data into a generative AI model, which compares the performance data with ad creatives and market trends to extract patterns and trends. For example, it might discover that a particular ad creative matches a particular market trend, resulting in a higher click-through rate.

[0442] Hypothesis and proposed measures

[0443] The server makes hypotheses based on the analysis results of the generative AI model. For example, it may make a hypothesis that "advertising creatives that match specific market trends are effective." The server then proposes countermeasures based on the hypothesis. For example, it may propose a countermeasure such as "use creatives that match current market trends in the next advertising campaign."

[0444] Emotion recognition and promotional measures

[0445] While the user is browsing a website, the device uses a camera and microphone to collect the user's facial expressions and voice. The emotion engine analyzes the collected data and recognizes the user's emotions. For example, if the user is smiling, it is recognized as feeling happy. The server proposes promotional measures appropriate for the user based on the emotion engine's results. For example, if the user is feeling happy or excited, it may display information about new product releases or special discounts. If the user is feeling frustrated or angry, it may display support information for resolving the problem or an apology message.

[0446] Specific examples

[0447] Measuring the effectiveness of advertising campaigns

[0448] The server retrieves click counts and purchase history from advertising campaigns over the past year from a database. The retrieved data is then cleansed, missing values ​​are filled, and outliers are corrected. The preprocessed data is then input into a generative AI model, which analyzes which advertising creative was most effective. Based on the analysis results, a hypothesis is made for the next advertising campaign, and a countermeasure is proposed, such as "creative content tailored to specific market trends will be effective."

[0449] Promotions based on user emotions

[0450] While the user is browsing a page introducing a new product, the device uses a camera and microphone to collect the user's facial expressions and voice. The emotion engine analyzes the collected data and recognizes that the user is excited. The server displays special discount information based on the emotion engine's results. The user checks the displayed discount information and considers purchasing.

[0451] Prompt Sentence Examples

[0452] "Analyze the most effective ad creatives based on data from advertising campaigns over the past year."

[0453] "Analyze user emotions and propose appropriate promotional measures to users who are happy."

[0454] The above is an embodiment of the present invention.

[0455] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0456] Step 1: Data collection

[0457] The server retrieves user performance data (e.g., ad clicks, purchase history) from a database, as well as ad creative and market trend data from external APIs. These data serve as input for subsequent analysis.

[0458] Step 2: Data Preprocessing

[0459] The server cleanses the acquired data and processes missing and outliers. Specifically, it imputes missing values ​​with the mean and replaces outliers with the median. It then standardizes the data and converts it into a format suitable for generative AI models. The preprocessed data becomes the input data for analysis.

[0460] Step 3: Data analysis

[0461] The server inputs the preprocessed data into a generative AI model. The generative AI model compares the performance data with ad creatives and market trends to extract patterns and trends. For example, it may discover that click-through rates are higher when specific ad creatives match specific market trends. The analysis results become output data for hypothesis generation.

[0462] Step 4: Propose hypotheses and countermeasures

[0463] The server formulates hypotheses based on the analysis results of the generative AI model. For example, it may formulate a hypothesis that "advertising creatives tailored to specific market trends are effective." The server then proposes countermeasures based on the hypothesis. For example, it may propose a countermeasure such as "In the next advertising campaign, use creatives that are tailored to current market trends." The proposed countermeasures become output data for the execution phase.

[0464] Step 5: Emotion Recognition

[0465] The device uses a camera and microphone to collect the user's facial expressions and voice while they are browsing a website. The emotion engine analyzes the collected data and recognizes the user's emotions. For example, if the user is smiling, it will recognize that they are feeling happy. The recognized emotion data becomes input data for proposing promotional measures.

[0466] Step 6: Propose promotional measures

[0467] The server proposes promotional measures appropriate for the user based on the results of the emotion engine. For example, if the user feels happy or excited, it displays information about new product releases or special discounts. If the user feels dissatisfied or angry, it displays support information for resolving the problem or an apology message. The proposed promotional measures become output data for the user.

[0468] (Application example 1)

[0469] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0470] The traditional PDCA cycle for promotional measures required a lot of time and effort to collect data, analyze it, propose hypotheses, and implement improvement actions. It was also difficult to propose promotional measures that took user emotions into account, making it difficult to achieve effective marketing. This made it difficult to implement promotional measures quickly and effectively, reducing the efficiency of corporate marketing activities.

[0471] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0472] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for inputting results such as acquisitions into the generative AI, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, means for recognizing user emotions, and means for proposing promotional measures according to the user emotions, thereby enabling the implementation of promotional measures quickly and effectively.

[0473] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[0474] "Promotional measures" are specific marketing activities aimed at increasing awareness of products and services and promoting sales.

[0475] The PDCA cycle is a method for continuously improving business processes and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0476] "Results" are data and results obtained as a result of past activities and measures.

[0477] "Advertising creative" refers to the elements that make up an advertisement, such as its content, design, and message.

[0478] "Market trends" are changes or trends in consumer preferences and behavior, competitor trends, etc. in a particular market.

[0479] A "hypothesis" is a prediction or conjecture about a particular phenomenon or problem, and is a premise whose validity is confirmed through verification.

[0480] A "measure" is a specific action or measure to address a specific problem or issue.

[0481] "Lead time" is the time required from the start to the end of a process.

[0482] "Emotion recognition" is a technology that determines a user's emotions from facial expressions, voice, text, etc.

[0483] A "promotional measure proposal" is a proposal for specific marketing activities that should be carried out to achieve a specific objective.

[0484] A system for carrying out the present invention uses generative AI to accelerate the PDCA cycle of promotional measures. Specific embodiments of the system will be described below.

[0485] System Configuration

[0486] The system includes a server, a user device, and a generative AI model. The server is responsible for collecting data, analyzing it, proposing hypotheses, and implementing improvement actions. The user device is used to recognize user emotions and receive promotional proposals.

[0487] Hardware and software used

[0488] Server: A high-performance computer equipped with a database, generative AI models, and an emotion recognition engine.

[0489] User devices: smartphones, tablets, computers, etc.

[0490] Generative AI model: Uses OpenAI API.

[0491] Emotion Recognition Engine: Uses the EmotionRecognizer library.

[0492] Data collection and analysis

[0493] The server collects performance data from ad campaigns, including clicks, conversion rates, and user feedback. This data is then fed into a generative AI model and analyzed against ad creative and market trends.

[0494] Hypothesis and proposed measures

[0495] The generative AI model proposes hypotheses and countermeasures based on collected data and market trends, thereby shortening the lead time for improving promotional measures.

[0496] Emotion recognition and promotional measures

[0497] The user device recognizes emotions using user input data (e.g., feedback and comments). The emotion recognition engine identifies the user's emotions in real time and proposes promotional measures based on those emotions. For example, if the user is dissatisfied, it can provide special discount information or support information for resolving the problem.

[0498] Specific examples

[0499] User input data: "I don't find recent ads very appealing."

[0500] Perceived Emotion: Frustration

[0501] Proposed promotion: "Test new ad creative designs. Offer special discount information."

[0502] Prompt Sentence Examples

[0503] Advertising campaign performance data: {ad_data}

[0504] User emotion: {emotion}

[0505] Based on this, please propose the best promotional measures.

[0506] In this way, by implementing a system that combines generative AI and an emotion recognition engine, it becomes possible to implement promotional measures quickly and effectively.

[0507] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0508] Step 1:

[0509] The server collects performance data of the advertising campaign, specifically, data such as the number of clicks, conversion rate, user feedback, etc. from a database. The input is the performance data of the advertising campaign, and the output is a set of collected data.

[0510] Step 2:

[0511] The server loads the collected performance data into the generative AI model. Specifically, the data is input using OpenAI's API and analyzed against advertising creatives and market trends. The input is the collected performance data, and the output is the analysis results from the generative AI model.

[0512] Step 3:

[0513] The server proposes hypotheses and countermeasures based on the analysis results of the generative AI model. Specifically, it generates optimal promotional measures based on the analysis results obtained from the generative AI model. The input is the analysis results of the generative AI model, and the output is proposed hypotheses and countermeasures.

[0514] Step 4:

[0515] The server transmits the hypotheses and proposed measures to the user terminal. Specifically, it notifies the user terminal of the proposed promotion measures. The input is the hypotheses and proposed measures, and the output is a notification to the user terminal.

[0516] Step 5:

[0517] The user terminal collects user input data. Specifically, it acquires user feedback and comments. The input is the user input data, and the output is the collected user data.

[0518] Step 6:

[0519] The user device inputs the collected user data into an emotion recognition engine to recognize the user's emotions. Specifically, it uses the EmotionRecognizer library to determine the user's emotions. The input is the user's input data, and the output is the recognized emotion.

[0520] Step 7:

[0521] The server proposes promotional measures based on the recognized user emotions. Specifically, it uses a generative AI model to generate optimal promotional measures according to the user emotions. The input is the recognized emotions, and the output is the proposed promotional measures.

[0522] Step 8:

[0523] The server transmits the proposed promotional measures to the user terminal. Specifically, it notifies the user terminal of the promotional measures according to the user's emotions. The input is the proposed promotional measures, and the output is the notification to the user terminal.

[0524] Example 2

[0525] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0526] Conventional promotional measures have a long cycle of planning, execution, evaluation, and improvement, which requires a lot of man-hours. It is also difficult to propose appropriate improvement actions based on user emotions, which means that the user experience is not fully improved. This results in the effectiveness of advertising creative not being maximized, and marketing efficiency is declining.

[0527] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0528] In this invention, the server includes means for accelerating the planning, execution, evaluation, and improvement cycle of promotional measures using generative artificial intelligence, means for loading past performance data, means for analyzing the data in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time to improvement actions, means for collecting user emotion data and analyzing it with an emotion engine, and means for proposing improvement actions according to the user's emotions based on the analysis results.This speeds up the PDCA cycle of promotional measures, reduces labor costs, and makes it possible to propose appropriate improvement actions according to the user's emotions.

[0529] "Generative AI" is AI that has the ability to generate new advertising creatives and strategies based on past data and market trends.

[0530] "Promotional measures" are plans and activities to increase awareness of and promote sales of a product or service.

[0531] The "plan-do-evaluate-improve cycle" is a series of processes for planning promotional measures, implementing them, evaluating the results, and improving them.

[0532] "Performance data" refers to data on the results and performance of promotional measures taken in the past.

[0533] "Advertising creative" refers to the creative elements of an advertisement, such as its content, design, and message.

[0534] "Market trends" refers to information such as current market conditions and trends, and consumer behavior patterns.

[0535] A "hypothesis" is a premise for making predictions or explanations about a particular problem or phenomenon.

[0536] A "measure" is a specific action or plan to address a particular problem or issue.

[0537] "Lead time" is the time it takes from planning to executing a promotional measure.

[0538] "Emotion data" refers to data relating to emotions obtained from the user's facial expressions, voice, etc.

[0539] An "emotion engine" is a technology that analyzes a user's emotional data and recognizes those emotions.

[0540] "Improvement actions" are specific improvement measures implemented based on user sentiment and the results of promotional measures.

[0541] This invention is a system that uses generative artificial intelligence to accelerate the planning, execution, evaluation, and improvement cycle of promotional measures and proposes improvement actions based on user emotions. Specific embodiments of this system are described below.

[0542] System configuration

[0543] Hardware and Software

[0544] 1. Server

[0545] The server collects past performance data from a database and analyzes the data using a generative artificial intelligence model. Specifically, it uses a cloud-based database service (e.g., Amazon RDS or Google Cloud SQL).

[0546] The server runs a generative artificial intelligence model (e.g., OpenAI's GPT-4) and proposes hypotheses and countermeasures based on advertising creative and market trends.

[0547] 2. Terminal

[0548] The device collects the user's facial expression and voice data and sends it to the emotion engine. Specifically, the device collects data using a camera and microphone.

[0549] The terminal uses an emotion engine (for example, Microsoft® Azure® Emotion API) to analyze the collected data and recognize the user's emotions.

[0550] 3. Users

[0551] The user checks the direction and measures of the advertising creative proposed by the server and implements them, specifically by setting up the advertising campaign and producing the creative.

[0552] The user receives suggestions from the server and takes necessary improvement actions, for example, watching a video on how to use the product.

[0553] Data processing and calculation

[0554] 1. Data collection and preprocessing

[0555] The server collects historical ad creative performance data and market trend data from the database, and the collected data is preprocessed using Python's Pandas library, which performs missing value imputation and data normalization.

[0556] 2. Analysis using generative AI models

[0557] The server inputs the pre-processed data into a generative AI model for analysis, which then generates the next direction for advertising creative and measures to address market trends.

[0558] 3. Emotion Data Analysis

[0559] The device sends the user's facial expressions and voice data to the emotion engine, which analyzes their emotions. The analysis results are sent to the server, which then suggests improvement actions based on the user's emotions.

[0560] Specific examples

[0561] Ad creative suggestion system using generative AI

[0562] Example prompt: "Based on ad creative performance data from the past six months and current market trend data, please suggest a direction for our next advertising campaign."

[0563] Improvement action suggestion system using emotion engine

[0564] Example prompt: "If you identify a user as confused, what remedial action would you suggest for that user?"

[0565] This enables the system to speed up the PDCA cycle of promotional measures, reduce labor costs, and suggest appropriate improvement actions based on user emotions.

[0566] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0567] Step 1:

[0568] Data collection

[0569] The server collects historical ad creative performance data and market trend data from a database. Specifically, it retrieves the data using a cloud-based database service (e.g., Amazon RDS or Google Cloud SQL). The input is the database query results, and the output is the collected dataset.

[0570] Step 2:

[0571] Data Preprocessing

[0572] The server preprocesses the collected data. Specifically, it uses the Python Pandas library to impute missing values ​​and normalize the data. The input is the collected dataset, and the output is the preprocessed dataset.

[0573] Step 3:

[0574] Analysis using generative AI models

[0575] The server inputs the preprocessed data into a generative AI model (e.g., OpenAI's GPT-4) for analysis. The input is the preprocessed dataset, and the output is the analysis results, which are the direction of advertising creative and measures against market trends. Specifically, the server inputs a prompt statement into the generative AI model and obtains the analysis results.

[0576] Step 4:

[0577] Proposal Generation

[0578] The server generates specific ad creative proposals based on the analysis results obtained from the generative AI model. The input is the analysis results of the generative AI model, and the output is specific ad creative proposals. Specific actions include generating new ad copy and visual ideas.

[0579] Step 5:

[0580] Implementing the proposal

[0581] The user confirms and implements the direction and measures of the advertising creative proposed by the server. The input is the proposal from the server, and the output is the implemented advertising campaign. Specific operations include setting up the advertising campaign and producing the creative.

[0582] Step 6:

[0583] Feedback of results

[0584] The server again collects performance data of the executed advertising creative and feeds it back to the generative AI model. The input is the performance data of the executed advertising campaign, and the output is an updated generative AI model. Specifically, the collected data is preprocessed again and input into the generative AI model.

[0585] Step 7:

[0586] Collecting user emotion data

[0587] The device collects the user's facial expression and voice data and sends it to the emotion engine. The input is the user's emotion data collected through the camera and microphone, and the output is the data sent to the emotion engine. Specifically, the device collects data using the camera and microphone.

[0588] Step 8:

[0589] Emotional Data Analysis

[0590] The server uses an emotion engine (for example, Microsoft Azure's Emotion API) to analyze the collected data and recognize the user's emotions. The input is the emotion data sent from the device, and the output is the user's emotional state as an analysis result. Specifically, the data is input into the emotion engine and the analysis results are obtained.

[0591] Step 9:

[0592] Proposal of improvement actions

[0593] The server proposes improvement actions based on the user's emotions based on the recognition results from the emotion engine. The input is the analysis result of the emotion engine, and the output is a proposal for a specific improvement action. As a specific operation, if the user is confused, it will suggest providing a video on how to use the product.

[0594] Step 10:

[0595] Implementing the proposal

[0596] The user receives the suggestion from the server and performs the necessary action. The input is the suggestion for improvement action from the server, and the output is the implemented improvement action. The specific action is to watch a video on how to use the product.

[0597] Step 11:

[0598] Feedback of results

[0599] The server collects the results of the user's implementation of the suggestions and feeds them back to the emotion engine. The input is the result of the implemented improvement action, and the output is an updated emotion engine. Specifically, the collected data is preprocessed again and input into the emotion engine.

[0600] (Application example 2)

[0601] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0602] In conventional promotional measures, the process of analyzing performance data of advertising creatives and market trend data, and proposing the direction of the next advertising creative and measures to address market trends, takes time and effort, so there is a need to speed up the PDCA cycle. Also, there is the issue of how difficult it is to properly recognize user emotions and propose corresponding improvement actions.

[0603] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0604] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, means for recognizing user emotions using an emotion recognition engine, and means for proposing improvement actions based on the recognized emotions. This makes it possible to speed up the PDCA cycle of promotional measures and quickly propose appropriate improvement actions according to the user's emotions.

[0605] "Generative AI" is artificial intelligence that generates new advertising creatives and strategies based on past data and market trends.

[0606] The PDCA cycle is a method for improving business operations and processes by repeating four steps: Plan, Do, Check, and Act.

[0607] "Ad creative" refers to the visual and content elements of an advertisement, such as its design, content, and message.

[0608] "Market trends" are data and trends that indicate the current market situation, such as consumer preferences and behavior in a particular market, and the actions of competitors.

[0609] An "emotion recognition engine" is a technology for recognizing and analyzing emotions from a user's facial expressions, voice, etc.

[0610] "Improvement actions" are specific improvements or measures proposed based on the analysis results and user sentiment.

[0611] "Lead time" refers to the time it takes from the start to the end of a process.

[0612] A "hypothesis" is a premise or way of thinking that allows one to make predictions or explanations about a particular problem or phenomenon.

[0613] "Countermeasures" are specific actions or measures taken to address a specific problem or issue.

[0614] The system for implementing this invention uses generative AI to accelerate the PDCA cycle of promotional measures, recognizes user emotions, and proposes appropriate improvement actions. Specific embodiments of this system are described below.

[0615] The server uses generative AI to read past ad creative performance data and market trend data to accelerate the PDCA cycle of promotional initiatives. This makes it possible to propose the direction of the next ad creative and measures to address market trends. Specifically, it uses OpenAI's API to analyze the data and generate hypotheses and measures.

[0616] The server also uses an emotion recognition engine to recognize the user's emotions. It receives the user's image and voice data as input and analyzes their emotions using EmotionRecognizer. Based on the recognized emotions, the server suggests appropriate improvement actions. For example, if the user is confused, it provides instructions on how to use the product, and if the user is satisfied, it suggests additional purchases.

[0617] The specific hardware configuration of this system includes a server, a user device (such as a smartphone or PC), and a camera and microphone to run the emotion recognition engine. The software uses OpenAI's API, EmotionRecognizer, and a database management system.

[0618] As a concrete example, let's say the performance data for an ad creative shows that "click-through rates are declining." Based on this data, the prompt to suggest the direction of the next ad creative is as follows:

[0619] Ad creative performance data: Click-through rates are declining. Please suggest the next direction for ad creative and measures to address market trends.

[0620] By inputting this prompt into OpenAI's API, the next direction for advertising creative and measures to address market trends are suggested, accelerating the PDCA cycle for promotional measures and enabling the rapid proposal of appropriate improvement actions based on user emotions.

[0621] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0622] Step 1:

[0623] The server collects historical ad creative performance data and market trend data, including retrieving data from a database and / or external APIs. The input is the ad creative performance data and market trend data, and the output is a consolidated dataset of these data.

[0624] Step 2:

[0625] The server inputs the collected data into generative AI to generate the next direction for advertising creative and countermeasures for market trends. Specifically, it uses OpenAI's API to analyze the data and generate hypotheses and countermeasures. The input is the integrated dataset, and the output is the generated hypotheses and countermeasures.

[0626] Step 3:

[0627] Based on the generated hypotheses and measures, the server creates a specific action plan to accelerate the PDCA cycle of promotional measures. The input is the generated hypotheses and measures, and the output is a specific action plan.

[0628] Step 4:

[0629] The user device recognizes the user's emotions using an emotion recognition engine. Specifically, it acquires the user's image and voice data through a camera and microphone, and analyzes the emotions using EmotionRecognizer. The input is the user's image and voice data, and the output is the recognized emotion.

[0630] Step 5:

[0631] The server then suggests appropriate improvement actions based on the recognized emotions. For example, if the user is confused, it provides product usage instructions, and if the user is satisfied, it suggests additional purchases. The input is the recognized emotions, and the output is the suggested improvement actions.

[0632] Step 6:

[0633] The server sends the proposed improvement action to the user terminal and notifies the user. The input is the proposed improvement action, and the output is the notification to the user.

[0634] Step 7:

[0635] The user checks the notified improvement actions and executes them as necessary. The input is the notified improvement action, and the output is the result of the executed improvement action.

[0636] This will speed up the PDCA cycle for promotional measures and make it possible to quickly propose appropriate improvement actions based on user emotions.

[0637] Example 3

[0638] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0639] Conventional promotional measures have a long lead time for the PDCA cycle, making it difficult to take quick action to improve. In addition, the creation of advertising creatives based on user emotions is done manually, which is inefficient and makes it difficult to provide advertisements that match user emotions.

[0640] The specification process by the specification processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes a means for accelerating the PDCA cycle of promotional measures using a generative AI, a means for inputting the results of acquisitions and the like into the generative AI, a means for analyzing the results in comparison with advertising creatives and market trends, a means for proposing hypotheses and measures based on the analysis results, a means for shortening the lead time until improvement actions, a means for recognizing user emotions using an emotion engine, and a means for generating advertising creatives according to the user emotions. This enables the PDCA cycle to be executed quickly and advertising creatives that are in line with the user's emotions to be provided.

[0641] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[0642] The PDCA cycle is a method for continuously improving business operations and processes by repeating the four steps of Plan, Do, Check, and Act.

[0643] "Lead time" is the time required from the start to the end of a process.

[0644] An "emotion engine" is a technology that analyzes a user's text and voice data and recognizes their emotions.

[0645] "Advertising creative" refers to elements that appeal to users visually or audibly, such as the content, design, and music of the advertisement.

[0646] "Market trends" refer to changes in consumer preferences and behaviors in a particular market, or trends therein.

[0647] A "hypothesis" refers to a premise or prediction that is used to propose an explanation or solution to a phenomenon or problem.

[0648] "Countermeasures" refer to specific actions or measures taken to address a specific problem or issue.

[0649] "Achievements" refers to results or accomplishments achieved in the past.

[0650] "User" refers to an individual or organization that uses the system or service.

[0651] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures and generate advertising creatives that respond to user emotions. A specific embodiment of this system is described below.

[0652] The server uses a generative AI model to accelerate the PDCA cycle of promotional measures. Specifically, the server receives data sent by users and analyzes it using a generative AI model (e.g., a general-purpose natural language processing model). Based on the analysis results, it proposes hypotheses and countermeasures and determines the priority of improvement actions. At this time, it also references past data stored in a database (e.g., a relational database management system).

[0653] Furthermore, the server uses an emotion engine to recognize the user's emotions. It analyzes text and voice data entered by the user through a device (e.g., a smartphone or PC) to obtain emotion recognition results. For example, an emotion analysis API can be used as the emotion engine.

[0654] Based on the emotion recognition results, the server uses a generative AI model to generate ad creatives. For example, if the user is feeling happy or excited, the server generates an ad using bright colors and lively music. On the other hand, if the user is feeling relieved or satisfied, the server generates an ad using calm colors and tranquil music.

[0655] As a concrete example, consider the case where a user posts a review of a new product. The user uses a device to input a review such as, "This product is very easy to use and has a great design." The server receives the review and uses a generative AI model to generate a prediction such as, "Improving the design of this product will increase user satisfaction by 20%." Next, the server uses an emotion engine to recognize the user's emotions and determine that the user is excited. Finally, the server uses the generative AI model to input a prompt such as, "Generate an ad creative that will enhance this user's excitement," and generates an ad featuring bright colors and lively music.

[0656] Example prompt sentence:

[0657] "Analyze user reviews, recognize emotions, and generate ad creatives that reflect those emotions."

[0658] "Predict the effectiveness of proposed measures and prioritize remedial actions"

[0659] In this way, the server, terminal, and user work together to execute each processing step of the system, enabling the rapid execution of the PDCA cycle and the provision of advertising creatives that are in line with the user's emotions. The flow of the specific processing in Example 3 will be described with reference to Figure 21.

[0660] Step 1:

[0661] The user inputs data through a device. The user uses a smartphone or PC to input reviews, feedback, and other text data. For example, a user might write a review of a new product, saying, "This product is very easy to use and has a great design." The input data is sent to the server via an HTTP request.

[0662] Step 2:

[0663] The server receives input data and predicts the effectiveness of countermeasures using a generative AI model. The server receives data sent by users and analyzes the data using a generative AI model (for example, a natural language processing model). Specifically, it analyzes the content of user reviews and generates a predicted result, such as "Improving the design of this product will increase user satisfaction by 20%." The input is the user review, and the output is the predicted result.

[0664] Step 3:

[0665] The server determines the priority of improvement actions based on the prediction results. The server determines the priority of improvement actions based on the prediction results. In doing so, it also refers to past data stored in a database (e.g., a relational database management system). For example, if it predicts that design improvements will be most effective, the server sets a priority such as "implement design improvements as the top priority." The inputs are the prediction results and past data, and the output is the priority.

[0666] Step 4:

[0667] The server uses an emotion engine to recognize the user's emotions. The server uses an emotion engine (for example, an emotion analysis API) to recognize the user's emotions. It analyzes the text data entered by the user and obtains emotion recognition results. For example, from a review that says, "This product is very easy to use and has a great design," it is recognized that the user is feeling happy or excited. The input is the user's review, and the output is the emotion recognition result.

[0668] Step 5:

[0669] The server generates ad creatives based on the emotion recognition results. The server generates ad creatives using a generative AI model based on the emotion recognition results. For example, if a user is feeling happy or excited, it generates an ad using bright colors and lively music. Specifically, a prompt such as "Generate ad creatives that will enhance this user's excitement" is input into the generative AI model. The inputs are the emotion recognition results and the prompt, and the output is the ad creatives.

[0670] Step 6:

[0671] The server sends the generated ad creative to the terminal. The server sends the generated ad creative to the user's terminal. For example, if the generated ad is in video format, the video data is sent to the terminal. The input is the ad creative, and the output is transmission to the terminal.

[0672] Step 7:

[0673] The user reviews the ad creative on their device. The user reviews the ad creative received on their device. For example, they may watch a video ad on their smartphone and provide feedback on the content. The input is the ad creative, and the output is the user's feedback.

[0674] (Application example 3)

[0675] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0676] Conventional promotional measures have the problem that the PDCA cycle takes time to execute, making it difficult to take quick action to improve. In addition, it is not possible to generate advertising creatives that correspond to the user's emotions, making it difficult to deliver effective advertisements.

[0677] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means. In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, means for recognizing user emotions using an emotion engine, and means for generating advertising creatives in accordance with the user emotions. This enables rapid execution of the PDCA cycle and effective ad delivery in accordance with user emotions.

[0678] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[0679] The PDCA cycle is a method for improving business processes and projects by repeating the four steps of Plan, Do, Check, and Act.

[0680] "Lead time" is the time it takes from the start to the end of a process.

[0681] An "emotion engine" is a technology that recognizes emotions from a user's facial expressions, voice, etc.

[0682] "Advertising creative" is a term that refers to elements such as the content, design, and message of an advertisement.

[0683] A "market trend" is a tendency that indicates changes in consumer preferences and behavior in a particular market.

[0684] A "hypothesis" is a premise or assumption used to explain or predict a phenomenon or problem.

[0685] A "measure" is a specific action or measure to address a specific problem or issue.

[0686] "User emotion" refers to a psychological state such as joy, excitement, relief, or satisfaction that a user feels in response to a particular situation or stimulus.

[0687] The system for implementing this invention uses generative AI to accelerate the PDCA cycle of promotional measures and generate advertising creatives that respond to user emotions. A specific embodiment of this system is described below.

[0688] System Program

[0689] This system is implemented using the following hardware and software.

[0690] Hardware: Smartphone camera and microphone

[0691] Software: OpenCV, Keras, playsound, PIL (Python Imaging Library)

[0692] Processing flow

[0693] The server captures video in real time using the smartphone's camera. It uses OpenCV to detect faces and converts the detected facial area to grayscale. Next, it predicts the user's emotion from the facial area using an emotion recognition model pre-trained with Keras. Depending on the predicted emotion, it generates an appropriate advertising creative (image) using PIL and plays the corresponding music using playsound. Finally, the advertising creative is displayed on the smartphone screen.

[0694] Specific examples

[0695] For example, if a user is using a smartphone and the camera captures the user's face, and the emotion recognition model predicts that the user is "happy," an advertisement with bright images and lively music will be displayed.

[0696] Prompt Sentence Examples

[0697] An example of a prompt sentence to input to the generative AI model is as follows:

[0698] Develop a smartphone application that recognizes user emotions in real time and generates and displays advertising creatives according to those emotions. If the user is happy, display an advertisement using bright colors and lively music, and if the user is relieved, display an advertisement using calm colors and gentle music.

[0699] In this way, the server can quickly execute the PDCA cycle and effectively deliver advertisements according to the user's emotions.

[0700] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0701] Step 1:

[0702] The server captures video in real time using the smartphone camera, the input is video data from the smartphone camera, and the output is real-time video frames, which are used in subsequent processing steps.

[0703] Step 2:

[0704] The server uses OpenCV to detect faces from video frames. The input is the video frame acquired in step 1, and the output is the coordinate information of the detected face area. Specifically, the video frame is converted to grayscale and the face area is identified by applying a face detection algorithm.

[0705] Step 3:

[0706] The server converts the detected face area to grayscale and performs preprocessing to input it into the emotion recognition model. The input is the coordinate information of the face area obtained in step 2 and the video frame, and the output is preprocessed face image data. Specifically, the face area is cut out, resized to 48x48 pixels, and normalized.

[0707] Step 4:

[0708] The server uses an emotion recognition model pre-trained with Keras to predict the user's emotion from the pre-processed facial image data. The input is the facial image data obtained in step 3, and the output is the emotion prediction result. Specifically, the facial image data is input into the emotion recognition model and the probability distribution of emotions is obtained.

[0709] Step 5:

[0710] The server generates advertising creatives according to the predicted emotions. The input is the emotion prediction result obtained in step 4, and the output is the generated advertising creatives (images) and music files. Specifically, if the emotion is "joy," it selects brightly colored images and lively music, and if the emotion is "relief," it selects calmly colored images and tranquil music.

[0711] Step 6:

[0712] The server displays the generated ad creative on the smartphone screen and plays the corresponding music. The input is the ad creative and music file obtained in step 5, and the output is the ad displayed to the user and the music played. Specifically, PIL is used to display the image and plays the music using playsound.

[0713] In this way, the server can generate and display advertising creatives in real time according to the user's emotions, achieving effective advertising delivery.

[0714] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0715] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0716] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are listed.

[0717] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0718] [Second embodiment]

[0719] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0720] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0721] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0722] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0723] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0724] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0725] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0726] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0727] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0728] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0729] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0730] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0731] "Example 1"

[0732] The present invention is a system that uses generative AI to accelerate the PDCA cycle of promotional initiatives. Specifically, the generative AI reads in acquisition results and analyzes them against advertising creatives and market trends. Based on the analysis results, it proposes hypotheses and measures, shortening the lead time until improvement actions are taken. This achieves a reduction in labor costs of 10 hours / month x 12 months x 4 initiatives = 480 hours / year.

[0733] "Example 2"

[0734] As a concrete example, generative AI reads performance data from past ad creatives and market trend data. Based on this data, it proposes the direction of the next ad creative and measures to address market trends. The proposed measures are implemented as part of the PDCA cycle, and the results are fed back to the generative AI. This speeds up the PDCA cycle and reduces labor costs.

[0735] "Example 3"

[0736] Furthermore, generative AI predicts the effectiveness of proposed measures and prioritizes improvement actions based on these predictions, significantly shortening lead times and enabling faster PDCA cycles.

[0737] The processing flow of each embodiment will be described below.

[0738] "Example 1"

[0739] Step 1: Load the generative AI with past achievements, such as acquisitions. This allows the AI ​​to learn based on past data.

[0740] Step 2: Analyze against ad creative and market trends. This analysis will help you understand the current market situation and how your ads are performing.

[0741] Step 3: Based on the analysis results, hypotheses and countermeasures are proposed, which then lead to the creation of the next action plan.

[0742] Step 4: Reduce the lead time to improvement actions, which speeds up the PDCA cycle and reduces man-hours.

[0743] "Example 2"

[0744] Step 1: Generative AI uses historical ad creative performance data and market trends

[0745] This allows the AI ​​to learn from past data.

[0746] Step 2: Based on this data, the generative AI proposes the next direction for advertising creative and measures to address market trends. This allows the next action plan to be formulated. Step 3: The proposed measures are implemented as part of the PDCA cycle, and the results are fed back to the generative AI. This speeds up the PDCA cycle and reduces labor costs.

[0747] "Example 3"

[0748] Step 1: Generative AI predicts the effectiveness of the proposed measures, which in turn predicts the effectiveness of the next action plan.

[0749] Step 2: The generative AI prioritizes improvement actions based on the prediction results, enabling the most effective actions to be implemented first.

[0750] Step 3: This significantly reduces lead times and enables a faster PDCA cycle.

[0751] Example 1

[0752] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0753] The traditional plan-do-check-act (PDCA) cycle for promotional measures required a great deal of time and effort for each step, such as collecting data, analyzing it, formulating hypotheses, and implementing measures, making it difficult to respond quickly. Furthermore, data cleansing and analysis required specialized knowledge, making efficient operation difficult. This delayed prompt improvement actions to maximize the effectiveness of promotional measures, potentially resulting in a decline in market competitiveness.

[0754] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0755] In this invention, the server includes means for accelerating the planning, execution, evaluation, and improvement cycle of promotional measures using generative artificial intelligence, means for loading performance data, means for analyzing the data against advertising materials and market trends, means for cleansing the performance data, filling in missing values, and removing outliers, means for analyzing the performance data using generative artificial intelligence and identifying effective patterns and trends, means for proposing specific measures based on hypotheses, and means for presenting specific improvement actions based on the proposed measures. This enables the PDCA cycle of promotional measures to be executed quickly and efficiently, reducing labor hours and improving market competitiveness.

[0756] "Generative AI" refers to AI that has the ability to generate new information and patterns based on data.

[0757] "Promotional measures" refer to plans and activities to increase awareness of and promote sales of products and services.

[0758] The "Plan-Do-Check-Act cycle (PDCA cycle)" is a method for achieving continuous improvement in business processes and projects by repeating the four steps of Plan, Do, Check, and Act.

[0759] "Performance data" refers to data obtained as a result of past activities and measures, including the number of clicks, conversion rates, and sales.

[0760] "Advertising materials" refers to the visuals, text, and other creative elements used in an advertising campaign.

[0761] "Market trends" refers to current market conditions and trends, such as consumer behavior and competitor actions in a particular market.

[0762] "Cleansing" refers to the process of preparing data for analysis, such as filling in missing values ​​and removing outliers.

[0763] "Patterns and trends" refer to certain regularities or characteristics found in data.

[0764] A "hypothesis" is a prediction or assumption made based on data analysis.

[0765] "Countermeasures" refer to specific actions or measures implemented based on a hypothesis.

[0766] "Improvement actions" refer to specific improvement measures implemented based on analysis results and hypotheses.

[0767] This invention is a system that uses generative artificial intelligence to accelerate the plan-do-check-act (PDCA) cycle of promotional measures. Specific embodiments of this system are described below.

[0768] System configuration

[0769] Hardware and Software Configuration

[0770] The server provides computational resources for running a generative artificial intelligence model (e.g., OpenAI's GPT-4). The server includes a database, a data processing unit, and a generative artificial intelligence model.

[0771] The terminal provides an interface for users to access the system, upload data, check analysis results, etc. Terminals include PCs, tablets, smartphones, etc.

[0772] Program processing

[0773] Data collection and preprocessing

[0774] Users upload promotional campaign performance data into the system in CSV file format, including the number of clicks, conversion rate, sales, and other data.

[0775] The server receives the uploaded data and stores it in a database, where it is cleansed, missing values ​​are filled, and outliers are removed.

[0776] Analyzing the data

[0777] The server uses a generative artificial intelligence model to analyze performance data, specifically matching it with advertising materials and market trends to identify effective patterns and trends.

[0778] The server identifies that particular advertising materials have demonstrated high conversion rates in particular market segments.

[0779] Hypothesis and proposed measures

[0780] The server generates a hypothesis based on the analysis results, for example, "Advertising material B shows a high conversion rate among young people."

[0781] The server then proposes specific measures based on the hypothesis, such as "Use advertising material B, which targets younger generations, in the next campaign."

[0782] Suggestion of improvement actions

[0783] The server then presents specific improvement actions to the user based on the proposed measures, such as "Use advertising material B in the next campaign and target younger generations."

[0784] The user checks the proposed improvement actions and reflects them in the next promotional measures. Specifically, the user plans and executes a new campaign using advertising material B.

[0785] Examples of concrete examples and prompts

[0786] Specific examples

[0787] User-Provided Data:

[0788] Ad Material A: 5,000 clicks, 2.5% conversion rate

[0789] Ad Material B: 3,000 clicks, conversion rate 3.0%

[0790] Market Trend Data: Online Shopping Increases Among Young People

[0791] Prompt Sentence Examples

[0792] "Based on the performance data of advertising campaigns over the past six months, please analyze the most effective advertising materials and market trends, and propose hypotheses and measures for the next campaign."

[0793] In this way, the server can propose quick and specific improvement actions to the user, accelerating the PDCA cycle.

[0794] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0795] Step 1:

[0796] Collection of performance data

[0797] Users upload promotional campaign performance data into the system in CSV file format.

[0798] Input: A CSV file containing performance data such as clicks, conversion rates, and sales.

[0799] The server receives the uploaded CSV file and stores it in a database.

[0800] Output: Actual data stored in a database.

[0801] Step 2:

[0802] Data Preprocessing

[0803] The server cleanses the stored performance data, specifically by filling in missing values ​​and removing outliers.

[0804] Input: Actual data stored in the database.

[0805] The server will fill in missing values ​​with the average value, remove outliers, and standardize the data format. For example, it will standardize the date format to "YYYY-MM-DD."

[0806] Output: Cleansed actual data.

[0807] Step 3:

[0808] Analyzing the data

[0809] The server analyzes the performance data using a generative artificial intelligence model (e.g., GPT-4).

[0810] Input: Cleansed actual data.

[0811] The server compares advertising materials and market trends with performance data to identify effective patterns and trends, such as identifying whether certain advertising materials result in higher conversion rates among certain market segments.

[0812] Output: Effective patterns and trends as a result of the analysis.

[0813] Step 4:

[0814] Hypothesis and proposed measures

[0815] The server generates a hypothesis based on the analysis results, for example, "Advertising material B shows a high conversion rate among young people."

[0816] Input: Effective patterns and trends as a result of analysis.

[0817] The server then proposes specific measures based on the hypothesis, such as "Use advertising material B, which targets younger generations, in the next campaign."

[0818] Output: Hypothesis and specific measures.

[0819] Step 5:

[0820] Suggestion of improvement actions

[0821] The server then presents specific improvement actions to the user based on the proposed measures, such as "Use advertising material B in the next campaign and target younger generations."

[0822] Input: Hypothesis and specific measures.

[0823] The server generates and presents remedial actions to the user.

[0824] Output: Specific improvement actions presented to the user.

[0825] Step 6:

[0826] Taking improvement actions

[0827] The user checks the proposed improvement actions and reflects them in the next promotional measures. Specifically, the user plans and executes a new campaign using advertising material B.

[0828] Input: The specific improvement actions suggested to the user.

[0829] The user runs a new campaign based on the improvement actions.

[0830] Output: The new campaign that was executed.

[0831] (Application example 1)

[0832] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0833] Traditional promotional measures require a great deal of time and effort in each phase of the PDCA cycle, and require rapid response, particularly in analyzing advertising creatives and market trends, formulating hypotheses, and implementing measures. However, there is a lack of systems to efficiently carry out these processes, which results in long lead times for improvement actions, making it difficult to implement effective promotional measures. In addition, there is no way to monitor advertising campaign performance in real time and quickly implement improvement actions, which makes it difficult to maximize advertising effectiveness.

[0834] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0835] In this invention, the server includes a means for accelerating the PDCA cycle of promotional measures using generative AI, a means for inputting results such as acquisitions into the generative AI, and a means for analyzing the results in comparison with advertising creatives and market trends. This makes it possible to propose hypotheses and measures based on the analysis results, shortening the lead time until improvement actions are taken. In addition, by including a means for monitoring the performance of advertising campaigns in real time and quickly taking improvement actions as an application installed on a smartphone, it is possible to maximize advertising effectiveness.

[0836] "Generative AI" refers to artificial intelligence that generates new information and content based on data.

[0837] "Promotional measures" are plans and activities to increase awareness of and promote sales of a product or service.

[0838] The PDCA cycle is a method for continuously improving business processes and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0839] "Results" refers to the data and results obtained as a result of past activities and measures.

[0840] "Advertising creative" refers to the elements that make up an advertisement, such as its content, design, and message.

[0841] "Market trends" refer to changes and trends in consumer preferences and behavior, competitor trends, etc. in a particular market.

[0842] A hypothesis is a prediction or premise made to find an explanation or solution to a phenomenon or problem.

[0843] "Countermeasures" refer to specific actions or measures taken to address a specific problem or issue.

[0844] "Lead time" is the time it takes from the start to the end of a process.

[0845] A "smartphone" is a highly functional mobile phone that can connect to the Internet and use applications.

[0846] An "application" is software designed to provide a specific function or service.

[0847] "Performance" refers to the results or effects of a particular activity or measure.

[0848] "Real-time" refers to the ability to process and display ongoing events and data immediately.

[0849] An "improvement action" is a specific improvement measure implemented to solve a specific problem or issue.

[0850] The system for carrying out this invention uses generative AI to accelerate the PDCA cycle of promotional measures. A specific embodiment of this system will be described below.

[0851] System Configuration

[0852] The system includes a server, a smartphone, a generative AI model, a database, and an internet connection. The server runs the generative AI model and accesses the database to retrieve advertising performance data. The smartphone runs an application that allows users to monitor advertising campaign performance in real time and take action to improve.

[0853] Program processing

[0854] The server processes the data in the following steps:

[0855] 1. Data loading:

[0856] The server reads advertising performance data from the database, including click-through rates (CTR), conversion rates (CVR), advertising costs, etc.

[0857] 2. Data Analysis:

[0858] The server uses a generative AI model to analyze advertising performance data, specifically by comparing it with advertising creative and market trends, and then propose hypotheses and countermeasures.

[0859] 3. Improvement suggestions:

[0860] The server uses a generative AI model to suggest improvement actions based on the analysis results, which are then communicated to the user via a smartphone application.

[0861] Hardware and software used

[0862] Hardware: Servers, smartphones

[0863] Software: Python, OpenAI API, Pandas

[0864] Data processing and calculation

[0865] The server uses Python to read advertising performance data from the database and processes the data using Pandas. It then analyzes the data using a generative AI model (OpenAI API) to generate hypotheses and countermeasures. The generated hypotheses and countermeasures are then notified to the user via a smartphone application.

[0866] Specific examples

[0867] For example, assume that the following advertising performance data exists:

[0868] Click-through rate: 2.5%

[0869] Conversion rate: 1.2%

[0870] Advertising cost: 500,000 yen

[0871] Based on this data, the server inputs the following prompt sentence into the generative AI model:

[0872] Example prompt sentence:

[0873] Based on the advertising performance data below, please analyze the advertising creative and market trends and suggest areas for improvement.

[0874] Advertising performance data:

[0875] Click-through rate: 2.5%

[0876] Conversion rate: 1.2%

[0877] Advertising cost: 500,000 yen

[0878] Based on this prompt, the generative AI model analyzes advertising creatives and market trends and suggests improvements. The proposed improvements are notified to the user via a smartphone application, allowing them to quickly take action.

[0879] In this way, the system can accelerate the PDCA cycle of promotional initiatives and maximize the performance of advertising campaigns.

[0880] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0881] Step 1:

[0882] The server reads ad performance data from the database. As input, it requires database connection information and a query. The server executes the query to obtain data such as click-through rate (CTR), conversion rate (CVR), and ad cost, and stores it in a Pandas data frame. The output is a data frame containing ad performance data.

[0883] Step 2:

[0884] The server generates a prompt to input the acquired advertising performance data into the generative AI model. Advertising performance data is required as input. The server converts the data frame into text format and generates a prompt like the following: "Based on the following advertising performance data, please analyze the advertising creative and market trends and suggest improvements. Advertising performance data: Click-through rate: 2.5%, Conversion rate: 1.2%, Advertising cost: 500,000 yen." The output is the generated prompt.

[0885] Step 3:

[0886] The server uses a generative AI model (OpenAI API) to input a prompt text and analyze ad creatives and market trends. The prompt text is required as input. The server sends the prompt text to the OpenAI API and receives a response from the generative AI model. The output is the analysis results from the generative AI model.

[0887] Step 4:

[0888] The server generates hypotheses and proposed countermeasures based on the analysis results from the generative AI model. The analysis results from the generative AI model are required as input. The server analyzes the analysis results and proposes specific improvement actions. The output is proposed hypotheses and countermeasures.

[0889] Step 5:

[0890] The server notifies the smartphone application of the proposed hypotheses and countermeasures. The proposed hypotheses and countermeasures are required as input. The server sends a notification to the smartphone application, prompting the user to take an improvement action. The output is the improvement action notified to the user.

[0891] Step 6:

[0892] The user checks the notified improvement actions through a smartphone application and executes them as necessary. The input requires a notification from the server. The user checks the notification and executes the improvement action according to the application's instructions. The output is the executed improvement action.

[0893] In this way, the system can accelerate the PDCA cycle of promotional initiatives and maximize the performance of advertising campaigns.

[0894] Example 2

[0895] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0896] Conventional promotional measures have the problem that each step of the PDCA cycle takes a lot of time and effort, making it difficult to make quick improvements. In addition, it was difficult to effectively utilize performance data on advertising creatives and market trend data, making it difficult to propose optimal measures. This resulted in the promotional effect not being maximized and resources being wasted.

[0897] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0898] In this invention, the server includes: a means for a user to upload past advertising creative performance data and market trend data to the server; a means for the server to save the uploaded data in a database and perform preprocessing; a means for the server to input the preprocessed data into a generative AI model and send a prompt sentence to the generative AI model; a means for the server to generate proposals and return them to the server; a means for the server to display the proposals to the user; a means for the user to implement the proposed measures and feed back the results to the server; and a means for the server to re-input the feedback data into the generative AI model and reflect it in the next proposal. This speeds up the PDCA cycle, reduces labor hours, and enables optimal measures to be proposed that effectively utilize advertising creative performance data and market trend data.

[0899] "Generative AI" is artificial intelligence that makes new suggestions and predictions based on past data and market trends.

[0900] The PDCA cycle is a method for continuously improving business operations and processes by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0901] "Promotional measures" are specific activities and strategies for increasing awareness of products and services and promoting sales.

[0902] "Performance data" refers to numerical data that indicates the effectiveness of advertising creatives and promotional measures, and includes the number of clicks, conversion rates, costs, etc.

[0903] "Market trend data" is data that shows current market trends and consumer behavior patterns.

[0904] A "server" is a computer system that stores, processes, and manages data.

[0905] A "database" is a system for efficiently storing, searching, and managing data.

[0906] "Preprocessing" refers to the process of converting data into a format suitable for analysis or model input.

[0907] A "generative AI model" is an artificial intelligence model that generates new suggestions and predictions based on input data.

[0908] A "prompt" is text that gives specific instructions or questions to a generative AI model.

[0909] The "proposal content" refers to the new advertising creative direction and measures for market trends generated by the generative AI model.

[0910] "Feedback" refers to data that is re-entered into the system to reflect the results of the measures implemented in the next proposal.

[0911] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures and reduce man-hours. This system consists of multiple components, including a user, a server, and a generative AI model.

[0912] First, the user uploads past advertising creative performance data and market trend data to the server. This data is generally provided in CSV or Excel file format. The server then stores the uploaded data in a database. For the database, a relational database management system (RDBMS) such as MySQL or PostgreSQL is used.

[0913] The server then preprocesses the stored data. This includes imputing missing values ​​and normalizing the data. These processes can be performed using the Python Pandas library. The preprocessed data is then converted into a format that is easy for the generative AI model to understand. Specifically, it is often converted into JSON format.

[0914] The server inputs the preprocessed data into a generative AI model. For example, natural language processing (NLP) models such as GPT-4 or BERT are used as generative AI models. The server inputs a prompt to the generative AI model. An example of a prompt is, "Based on the performance data of advertising creatives over the past six months and current market trend data, please suggest a direction for the next advertising campaign."

[0915] Based on the input data and prompt text, the generative AI model proposes the next direction for advertising creative and measures to address market trends. The generated proposals are returned to the server in text format. The server receives the proposals returned by the generative AI model and displays them to the user. For display, data visualization tools such as Tableau or Power BI are used.

[0916] The user implements the proposed measures and feeds the results back to the server. The feedback data is then saved in the database and reflected in the next proposal. The server then inputs the feedback data back into the generative AI model, accelerating the PDCA cycle.

[0917] This system will speed up the PDCA cycle, reduce labor costs, and enable the proposal of optimal measures that effectively utilize advertising creative performance data and market trend data.

[0918] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0919] Step 1:

[0920] Users upload historical advertising creative performance data and market trend data to the server.

[0921] Input: Past ad creative performance data (e.g., CSV file), market trend data (e.g., Excel file)

[0922] Output: Data stored on the server

[0923] Specific operation: A user selects a CSV or Excel file through the web interface and clicks the upload button. The server receives the uploaded file and stores it in the database.

[0924] Step 2:

[0925] The server stores the uploaded data in a database and performs preprocessing.

[0926] Input: Uploaded data (CSV file, Excel file)

[0927] Output: Preprocessed data (JSON format)

[0928] How it works: The server reads the data using Python's Pandas library, imputes missing values, normalizes the data, and converts the preprocessed data into a JSON format that is easy for the generative AI model to understand.

[0929] Step 3:

[0930] The server inputs the preprocessed data into the generative AI model and sends the prompt sentence to the generative AI model.

[0931] Input: Preprocessed data (JSON format), prompt statement

[0932] Output: Proposals from the generative AI model (text format)

[0933] Specific operation: The server calls the generative AI model's API and sends the preprocessed data and a prompt. An example of a prompt is, "Based on the performance data of the ad creative over the past six months and current market trend data, please suggest the direction of our next ad campaign."

[0934] Step 4:

[0935] Based on the input data and prompt text, the generative AI model suggests the next direction for advertising creative and measures to address market trends.

[0936] Input: Preprocessed data (JSON format), prompt statement

[0937] Output: Proposal content (text format)

[0938] How it works: The generative AI model uses its internal algorithms to analyze data and generate optimal ad creative direction and countermeasures for market trends. The generated proposals are returned to the server in text format.

[0939] Step 5:

[0940] The server receives the suggestions returned by the generative AI model and displays them to the user.

[0941] Input: Proposal content from the generative AI model (text format)

[0942] Output: The suggestions displayed to the user

[0943] Specific operation: The server receives the proposals, creates a dashboard using a data visualization tool (e.g., Tableau, Power BI), and displays it to the user.

[0944] Step 6:

[0945] The user implements the proposed measures and feeds back the results to the server.

[0946] Input: Implementation result data (e.g. Excel file)

[0947] Output: Feedback data stored on the server

[0948] Specific operation: The user compiles the results of the implemented measures in an Excel file and uploads it to the server. The server receives the uploaded file and stores it in a database.

[0949] Step 7:

[0950] The server then inputs the feedback data back into the generative AI model and reflects it in the next proposal.

[0951] Input: Feedback data (JSON format)

[0952] Output: Data reflected in the next proposal

[0953] Specific operation: The server preprocesses the feedback data and re-inputs it into the generative AI model, which then takes the feedback data into account to generate the next proposal.

[0954] (Application example 2)

[0955] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0956] Conventional promotional measures had the problem of taking a lot of time and effort to analyze advertising creative performance data and market trend data and propose hypotheses and countermeasures. There was also a lack of means to speed up the PDCA cycle and efficiently implement improvement actions. Furthermore, there was no established process for efficient data input using smart devices, or for implementing and providing feedback on proposals using generative AI models.

[0957] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for accelerating the PDCA cycle of promotional measures using a generative AI; means for loading the generative AI with actual results such as acquisitions; means for analyzing the results in comparison with advertising creatives and market trends; means for proposing hypotheses and measures based on the analysis results; means for shortening the lead time to improvement actions; means for inputting performance data and market trend data of past advertising creatives using an application installed on a smart device and inputting prompt statements into the generative AI model; and means for implementing measures in response to the advertising creative direction and market trends proposed by the generative AI model and providing feedback on the results. This makes it possible to speed up the PDCA cycle of promotional measures and significantly reduce labor costs.

[0958] "Generative AI" is artificial intelligence that generates new advertising creatives and strategies based on past data and market trends.

[0959] "Promotional measures" are plans and activities to increase awareness of products and services and promote sales.

[0960] The PDCA cycle is a method for continuously improving business operations by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[0961] "Performance data" refers to data used to measure the effectiveness of advertising creatives, including click-through rates and conversion rates.

[0962] "Market trend data" is data that indicates current market trends and consumer interests.

[0963] A "smart device" is a portable electronic device that can connect to the Internet, including smartphones and smart glasses.

[0964] A "prompt" is an instruction given to generative AI, and serves as the basis for generating the direction and measures for advertising creatives.

[0965] "Feedback" is the process of inputting the results of implemented measures back into the generative AI and reflecting them in the next proposal.

[0966] The system for carrying out this invention uses generative AI to accelerate the PDCA cycle of promotional measures. A specific embodiment of this system will be described below.

[0967] Hardware and software used

[0968] Hardware: Smartphones, smart glasses, servers

[0969] Software: Generative AI models (e.g., OpenAI GPT-4), database management systems (e.g., MySQL, PostgreSQL), data analysis tools (e.g., Tableau, Google Analytics)

[0970] Data processing and calculation

[0971] 1. Data Collection:

[0972] The server retrieves historical performance data of advertising creatives from a database, and also retrieves market trend data from an external API (e.g., Google Trends API).

[0973] 2. Data Preprocessing:

[0974] The server cleans the acquired data, imputes missing values, normalizes the data, and converts it into a format suitable for analysis.

[0975] 3. Data Analysis:

[0976] The server uses data analysis tools to evaluate the performance of past advertising creatives and analyzes market trend data to identify current trends.

[0977] 4. Utilizing generative AI models:

[0978] The server inputs the following prompt sentences into the generative AI model to generate the next direction for advertising creative and measures for market trends.

[0979] Prompt Sentence Examples

[0980] Based on past ad creative performance data and current market trend data, please propose the next ad creative direction and measures to address market trends.

[0981] 5. Implementation and feedback of proposals:

[0982] Using an application installed on a smart device, users can implement the advertising creative direction and market trend measures suggested by the generative AI model. The results are then fed back into the generative AI model, accelerating the PDCA cycle.

[0983] Specific examples

[0984] Example 1: Smartphone application

[0985] Users open the smartphone app and input past ad creative performance data and market trend data. The app uses a generative AI model to suggest the direction of the next ad creative. Users can then review and implement the suggestions, completing the PDCA cycle.

[0986] Example 2: Smart Glasses Application

[0987] The marketer creating the ad creative wears the smart glasses and uses voice commands to access past performance data and market trend data. The smart glasses use a generative AI model to suggest the direction of the next ad creative in real time. The marketer then reviews the suggestions and implements them, completing the PDCA cycle.

[0988] In this way, it is possible to speed up the PDCA cycle of promotional measures and significantly reduce labor costs.

[0989] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0990] Step 1:

[0991] The server retrieves past performance data of ad creatives from a database. The inputs are database connection information and a query, and the output is performance data of ad creatives. Specifically, the server connects to the database and executes the specified query to retrieve the data.

[0992] Step 2:

[0993] The server retrieves market trend data from an external API. The input requires the API endpoint and authentication information, and the output is market trend data. Specifically, the server sends a request to the API and receives market trend data as a response.

[0994] Step 3:

[0995] The server cleans the acquired data and fills in missing values. It takes advertising creative performance data and market trend data as input, and obtains cleaned data as output. Specifically, the server identifies missing values ​​in the data and fills them in appropriately.

[0996] Step 4:

[0997] The server normalizes the data and converts it into a format suitable for analysis. It takes cleaned data as input and normalized data as output. Specifically, the server scales and encodes the data.

[0998] Step 5:

[0999] The server uses a data analysis tool to evaluate the performance of past advertising creatives. Normalized data is required as input, and performance evaluation results are obtained as output. Specifically, the server inputs data into the data analysis tool and obtains the evaluation results.

[1000] Step 6:

[1001] The server analyzes the market trend data to identify the current trend. The input required is normalized market trend data, and the output is trend analysis results. Specifically, the server uses data analysis tools to identify the trend.

[1002] Step 7:

[1003] The server inputs prompt statements into the generative AI model and generates the direction of the next ad creative and measures for market trends. The required inputs are performance evaluation results, trend analysis results, and prompt statements, and the output is the direction of the generated ad creative and measures. Specifically, the server inputs prompt statements into the generative AI model and obtains the generated results.

[1004] Step 8:

[1005] Using an application installed on a smart device, users implement the advertising creative direction and market trend countermeasures suggested by the generative AI model. The inputs are the generated advertising creative direction and countermeasures, and the output is the implementation results. Specifically, users operate the application to implement the suggested countermeasures.

[1006] Step 9:

[1007] The user then feeds back the results of the measures they have implemented to the generative AI model. The input is the implementation results, and the output is the fed-back data. Specifically, the user uses an application to input the implementation results and feeds them back to the generative AI model.

[1008] Example 3

[1009] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1010] The PDCA cycle for conventional promotional measures had long lead times and made it difficult to take prompt action to improve. Furthermore, there was a lack of concrete methods for effectively utilizing the analysis results using generative AI, making it difficult to reduce man-hours and prioritize effective measures.

[1011] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes a means for the user to input a prompt sentence, a means for the terminal to send the prompt sentence to the server, a means for the server to preprocess the prompt sentence, a means for the server to input the prompt sentence to the generative AI model, a means for the generative AI model to predict the effectiveness of a countermeasure, a means for the server to receive the prediction result, a means for the server to analyze the prediction result and determine the priority of improvement actions, a means for the server to return the priority to the user, and a means for the user to execute the improvement action based on the priority. This enables shortening of lead time and realization of a rapid PDCA cycle.

[1012] "Generative AI" is an artificial intelligence technology that generates data based on user input and performs analysis and predictions.

[1013] "Promotional measures" are marketing activities aimed at increasing awareness of products and services and promoting sales.

[1014] The PDCA cycle is a method for achieving continuous improvement in business operations and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[1015] "Lead time" is the time required from the start to the end of a process.

[1016] A "prompt" is text that contains instructions or questions that a user inputs to a generative AI.

[1017] A "server" is a computer system that provides services to other computers on a network.

[1018] A "terminal" is a device that a user uses to communicate with a server.

[1019] "Preprocessing" refers to the initial process of converting data into a format suitable for analysis or prediction.

[1020] A "generative AI model" is a specific implementation of generative AI, an algorithm that generates data and makes predictions based on user input.

[1021] "Prediction results" are the results of analysis and predictions that a generative AI model outputs based on user input.

[1022] "Improvement actions" are specific measures or policies implemented based on the prediction results.

[1023] "Priority" refers to the order of importance or urgency of execution among multiple improvement actions.

[1024] This invention relates to a system that uses generative AI to accelerate the PDCA cycle of promotional measures and shorten lead times. Specific embodiments of this system are described below.

[1025] First, the user uses a terminal to input a prompt. A prompt is text used to give instructions or ask questions to the generative AI. For example, a prompt such as "Please predict the effect of increasing sales if a new marketing strategy is implemented" can be input.

[1026] The terminal sends the prompt text entered by the user to the server. The HTTPS protocol is used for communication to ensure data security. The server then preprocesses the received prompt text. Specifically, it normalizes the text and removes unnecessary spaces.

[1027] Next, the server inputs the preprocessed prompt sentence into a generative AI model. The generative AI model runs on a high-performance server equipped with an NVIDIA GPU. For example, GPT-4 is used as the generative AI model. This generative AI model predicts the effectiveness of countermeasures based on the prompt sentence. For example, it predicts the effect of increasing sales based on past data and similar cases.

[1028] The server receives prediction results from the generative AI model. The prediction results are returned as numerical or text data. The server analyzes the prediction results and prioritizes improvement actions. For example, it ranks marketing strategies in order of their effectiveness in increasing sales.

[1029] The server returns the determined priorities to the user, who then displays them to the terminal. The user then takes action to improve the situation based on the priorities received from the server. For example, the user may first implement the marketing strategies predicted to be most effective.

[1030] This system significantly shortens lead times and enables a faster PDCA cycle. In addition, by effectively utilizing the analysis results using generative AI, it is possible to reduce man-hours and prioritize effective countermeasures.

[1031] As a concrete example, consider the case where a user inputs a prompt such as, "Please predict the sales increase effect of introducing a new marketing strategy." The device sends this prompt to the server, which then predicts the sales increase effect using a generative AI model. Based on the prediction results, the server determines the priority of the marketing strategies and returns it to the user. The user then implements the marketing strategy predicted to be most effective based on this priority.

[1032] In this way, the PDCA cycle of promotional measures utilizing generative AI can be accelerated and the lead time can be shortened. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1033] Step 1:

[1034] The user inputs a prompt sentence, such as "Please predict the sales increase effect of introducing a new marketing strategy" through the terminal interface. The input prompt sentence is temporarily stored in the terminal's memory.

[1035] Step 2:

[1036] The terminal sends the prompt text to the server. The terminal sends the prompt text entered by the user to the server using the HTTPS protocol. The input is the prompt text, and the output is a notification to the server that transmission has been completed.

[1037] Step 3:

[1038] The server preprocesses the prompt text. The server normalizes the received prompt text and removes unnecessary spaces and special characters. The input is the prompt text and the output is the preprocessed prompt text. Specific operations include text cleaning and tokenization.

[1039] Step 4:

[1040] The server inputs a prompt sentence to the generative AI model. The server then inputs a preprocessed prompt sentence to the generative AI model. The input is the preprocessed prompt sentence, and the output is a notification to the generative AI model that the input has been completed. Specific operations include API calls and sending data to the model.

[1041] Step 5:

[1042] The generative AI model predicts the effectiveness of countermeasures. The generative AI model predicts the effectiveness of countermeasures based on the prompt sentence. The input is the preprocessed prompt sentence, and the output is the predicted result. Specific operations include the model's inference process.

[1043] Step 6:

[1044] The server receives the prediction results. The server receives the prediction results from the generative AI model. The input is the prediction results, and the output is saving the prediction results to the server. Specific operations include saving the results to a database and analyzing the results.

[1045] Step 7:

[1046] The server analyzes the prediction results and determines the priorities of improvement actions. The server analyzes the prediction results and determines the priorities of improvement actions. The input is the prediction results and the output is a priority list. Specific operations include analysis using an algorithm and generation of rankings.

[1047] Step 8:

[1048] The server returns the priority to the user. The server returns the determined priority to the user. The input is a priority list, and the output is a notification to the user that transmission is complete. Specific operations include data format conversion and the transmission process.

[1049] Step 9:

[1050] The user executes improvement actions based on the priorities. The user executes improvement actions based on the priorities received from the server. The input is a priority list, and the output is the executed improvement actions. Specific operations include the introduction of marketing strategies and the execution of measures.

[1051] (Application example 3)

[1052] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1053] In conventional promotional measures and factory robot operations, there is a demand for accelerating the PDCA cycle and shortening lead times, but there is a lack of efficient ways to achieve this. Furthermore, there is no system in place that can collect factory robot operation data in real time, predict the effects of improvement measures using generative AI, and determine priorities, making it difficult to maximize production efficiency. A new system is needed to solve these issues.

[1054] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1055] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time to improvement actions, means for collecting operation data of robots in the factory in real time and predicting the effects of the proposed improvement measures using generative AI, and means for determining the priorities of improvement actions based on the prediction results and notifying the factory manager. This makes it possible to improve the efficiency of promotional measures and maximize the production efficiency of factory robots.

[1056] "Generative AI" refers to artificial intelligence that generates new information and predictions based on data.

[1057] The PDCA cycle is a method for improving business operations by repeating the four steps of Plan, Do, Check, and Act.

[1058] "Lead time" is the time it takes for a product or service to be delivered after it is ordered.

[1059] A "factory robot" is a mechanical device that performs work automatically in a factory.

[1060] "Operational data" refers to data that records the operating status of a machine or system.

[1061] "Improvement measures" refer to specific means or methods for solving current problems.

[1062] "Priority" refers to determining the order in which multiple items or tasks should be done first.

[1063] A "factory manager" is a position that involves the operation and management of a factory.

[1064] "Promotional measures" are specific activities or plans aimed at promoting products or services.

[1065] "Advertising creative" refers to the content, design, and expression of the advertisement.

[1066] A "market trend" is a change in consumer preferences and behavior in a particular market.

[1067] A "hypothesis" is a tentative explanation or prediction about a phenomenon or problem.

[1068] A "proposal" is a presentation of a solution or new idea to a problem.

[1069] The following system configuration will be described as an embodiment of the present invention.

[1070] The server uses generative AI to provide a means to accelerate the PDCA cycle of promotional measures. Specifically, the server loads performance data such as acquisition data into the generative AI, and analyzes it in comparison with advertising creatives and market trends. Based on the analysis results, the server proposes hypotheses and countermeasures, shortening the lead time until improvement actions are taken.

[1071] Furthermore, the server collects operational data from robots in the factory in real time and uses generative AI to predict the effectiveness of proposed improvement measures. Based on the prediction results, it determines the priority of improvement actions and provides a means to notify factory managers. This makes it possible to maximize factory production efficiency and shorten lead times.

[1072] This system uses the following hardware and software:

[1073] Hardware: Smartphone (iOS or Android), factory robot

[1074] Software: Python, Pandas, Scikit-learn, Matplotlib

[1075] The server reads the factory robot's operation data from a CSV file and splits the data into features and targets. It splits the data into training data and test data, and uses a random forest regression model to learn the training data. It makes predictions on the test data and evaluates the predicted effects. It calculates the importance of features and determines the priority of improvement actions. It displays the results in a graph and notifies the factory manager.

[1076] As a specific example, assume that the operation data of a factory robot is in the following format:

[1077] time, speed, load, temperature, target

[1078] 0, 1.2, 50, 30, 0.8

[1079] 1, 1.5, 55, 32, 0.9

[1080] 2, 1.1, 52, 31, 0.85

[1081] ...

[1082] An example of a prompt sentence to input to the generative AI model is as follows:

[1083] "Based on the operation data of a factory robot, predict the effect of improvement measures and determine priorities. The data is in the following format: time, speed, load, temperature, target. Output the prediction results and priorities."

[1084] Using this prompt, the generative AI model can make appropriate predictions and prioritization decisions.

[1085] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1086] Step 1:

[1087] The server reads the factory robot's operation data from a CSV file. It requires a CSV file as input and obtains data in a data frame format as output. This data contains columns for time, speed, load, temperature, and target. Specifically, it reads the CSV file using Python's Pandas library.

[1088] Step 2:

[1089] The server splits the loaded data into features (X) and targets (y). Data in a data frame format is required as input, and a dataset of features and targets is obtained as output. Specifically, it separates the target column from the data frame and treats the remaining columns as features.

[1090] Step 3:

[1091] The server splits the features and targets into training data and test data. As input, it requires a dataset of features and targets, and as output, it obtains training data and test data. Specifically, it splits the data using the train_test_split function of Scikit-learn.

[1092] Step 4:

[1093] The server uses a random forest regression model to learn the training data. The training data is required as input, and the trained model is obtained as output. Specifically, the model is created using Scikit-learn's RandomForestRegressor class, and training is performed using the fit method.

[1094] Step 5:

[1095] The server uses the trained model to make predictions on test data. Test data is required as input, and prediction results are obtained as output. Specific operations include making predictions using the predict method.

[1096] Step 6:

[1097] The server evaluates the effect prediction by comparing the predicted results with the actual values. The predicted results and actual values ​​are required as input, and an evaluation index (e.g., mean squared error) is obtained as output. Specifically, the evaluation is performed using the mean_squared_error function of Scikit-learn.

[1098] Step 7:

[1099] The server calculates the importance of features and determines the priority of improvement actions. A trained model is required as input, and the importance and ranking of features are obtained as output. Specifically, the server obtains the importance using the feature_importances_ attribute of the model and determines the ranking based on the importance.

[1100] Step 8:

[1101] The server displays the results in a graph and notifies the factory manager. The input required is the feature importance and its ranking, and the output is a visual graph and a notification message. Specifically, the server creates the graph using the Matplotlib library and sends the results to the factory manager through a notification system.

[1102] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1103] "Example 1"

[1104] One embodiment of the present invention is a system that combines generative AI and an emotion engine. This system recognizes a user's emotions and proposes promotional measures based on those emotions. Specifically, if the system recognizes that the user is feeling joy or excitement, it proposes promotional measures such as new product release information or special discount information to the user. If the system recognizes that the user is feeling dissatisfied or anger, it proposes promotional measures such as support information for resolving the user's problem or an apology message to the user.

[1105] "Example 2"

[1106] Another embodiment of the present invention is a system that suggests improvement actions according to a user's emotions based on the emotion recognition results of an emotion engine. Specifically, if it is recognized that a user is feeling confused or perplexed, it suggests improvement actions such as providing the user with information on how to use a product or details of a service. Also, if it is recognized that a user is feeling relieved or satisfied, it suggests improvement actions such as suggesting the user to purchase an additional product or upgrade a service.

[1107] "Example 3"

[1108] Furthermore, another embodiment of the present invention is a system that generates advertising creatives according to a user's emotions based on the emotion recognition results of an emotion engine. Specifically, if it is recognized that the user is feeling joy or excitement, it generates advertising creatives using bright colors and lively music that enhance the user's emotions. On the other hand, if it is recognized that the user is feeling relief or satisfaction, it generates advertising creatives using calm colors and tranquil music that deepen the user's emotions.

[1109] The processing flow of each embodiment will be described below.

[1110] "Example 1"

[1111] Step 1: Generative AI collects user behavior data.

[1112] Step 2: Based on the collected data, the generative AI predicts the user's emotions.

[1113] Step 3: The emotion engine receives the predictions from the generative AI and recognizes the user's emotions.

[1114] Step 4: Based on the emotions recognized by the emotion engine, appropriate promotional measures are proposed.

[1115] "Example 2"

[1116] Step 1: Generative AI collects user behavior data.

[1117] Step 2: Based on the collected data, the generative AI predicts the user's emotions.

[1118] Step 3: The emotion engine receives the predictions from the generative AI and recognizes the user's emotions.

[1119] Step 4: The emotion engine suggests appropriate improvement actions based on the emotions it recognizes. "Example 3"

[1120] Step 1: Generative AI collects user behavior data.

[1121] Step 2: Based on the collected data, the generative AI predicts the user's emotions.

[1122] Step 3: The emotion engine receives the predictions from the generative AI and recognizes the user's emotions.

[1123] Step 4: The emotion engine generates appropriate ad creatives based on the emotions it recognizes.

[1124] Example 1

[1125] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1126] Conventional promotional measures require a lot of time and effort for data analysis, hypothesis formulation, and countermeasure proposals at each phase of the PDCA cycle, making it difficult to take prompt action to improve. Furthermore, promotional measures do not take into account user emotions, which means that user satisfaction cannot be increased.

[1127] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1128] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for inputting results such as acquisitions into the generative AI, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, and means for recognizing user emotions and proposing promotional measures according to those emotions. This enables the implementation of prompt and effective promotional measures and appropriate responses according to user emotions.

[1129] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[1130] The PDCA cycle is a business improvement method consisting of four phases: Plan, Do, Check, and Act.

[1131] "Results" are data and results obtained as a result of past activities and measures.

[1132] "Advertising creative" is a general term that refers to elements such as the content, design, and message of an advertisement.

[1133] "Market trends" refer to changes in consumer preferences and behavior, trends in fashion, etc. in a particular market.

[1134] A "hypothesis" is an explanation or proposal based on a prediction or conjecture about a particular phenomenon or problem.

[1135] "Countermeasures" are specific actions or measures taken to address a specific problem or issue.

[1136] "Lead time" is the time it takes from the start to the end of a process.

[1137] "Emotion recognition" is a technology that analyzes and identifies emotions from a user's facial expressions, voice, etc.

[1138] "Promotional measures" are specific activities and strategies aimed at promoting products and services.

[1139] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures. This system works in cooperation with three parties: a server, a terminal, and a user.

[1140] Hardware and software used

[1141] Hardware: Servers, terminals

[1142] Software: Generative AI models (e.g., GPT-4), emotion engines (e.g., Affectiva)

[1143] Data collection and preprocessing

[1144] The server retrieves user performance data (e.g., ad clicks, purchase history) from a database. It also retrieves data on ad creatives and market trends from external APIs. The retrieved data is cleansed and missing values ​​and outliers are processed. The data is standardized and converted into a format suitable for generative AI models.

[1145] Data analysis

[1146] The server then inputs the preprocessed data into a generative AI model, which compares the performance data with ad creatives and market trends to extract patterns and trends. For example, it might discover that a particular ad creative matches a particular market trend, resulting in a higher click-through rate.

[1147] Hypothesis and proposed measures

[1148] The server makes hypotheses based on the analysis results of the generative AI model. For example, it may make a hypothesis that "advertising creatives that match specific market trends are effective." The server then proposes countermeasures based on the hypothesis. For example, it may propose a countermeasure such as "use creatives that match current market trends in the next advertising campaign."

[1149] Emotion recognition and promotional measures

[1150] While the user is browsing a website, the device uses a camera and microphone to collect the user's facial expressions and voice. The emotion engine analyzes the collected data and recognizes the user's emotions. For example, if the user is smiling, it is recognized as feeling happy. The server proposes promotional measures appropriate for the user based on the emotion engine's results. For example, if the user is feeling happy or excited, it may display information about new product releases or special discounts. If the user is feeling frustrated or angry, it may display support information for resolving the problem or an apology message.

[1151] Specific examples

[1152] Measuring the effectiveness of advertising campaigns

[1153] The server retrieves click counts and purchase history from advertising campaigns over the past year from a database. The retrieved data is then cleansed, missing values ​​are filled, and outliers are corrected. The preprocessed data is then input into a generative AI model, which analyzes which advertising creative was most effective. Based on the analysis results, a hypothesis is made for the next advertising campaign, and a countermeasure is proposed, such as "creative content tailored to specific market trends will be effective."

[1154] Promotions based on user emotions

[1155] While the user is browsing a page introducing a new product, the device uses a camera and microphone to collect the user's facial expressions and voice. The emotion engine analyzes the collected data and recognizes that the user is excited. The server displays special discount information based on the emotion engine's results. The user checks the displayed discount information and considers purchasing.

[1156] Prompt Sentence Examples

[1157] "Analyze the most effective ad creatives based on data from advertising campaigns over the past year."

[1158] "Analyze user emotions and propose appropriate promotional measures to users who are happy."

[1159] The above is an embodiment of the present invention.

[1160] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1161] Step 1: Data collection

[1162] The server retrieves user performance data (e.g., ad clicks, purchase history) from a database, as well as ad creative and market trend data from external APIs. These data serve as input for subsequent analysis.

[1163] Step 2: Data Preprocessing

[1164] The server cleanses the acquired data and processes missing and outliers. Specifically, it imputes missing values ​​with the mean and replaces outliers with the median. It then standardizes the data and converts it into a format suitable for generative AI models. The preprocessed data becomes the input data for analysis.

[1165] Step 3: Data analysis

[1166] The server inputs the preprocessed data into a generative AI model. The generative AI model compares the performance data with ad creatives and market trends to extract patterns and trends. For example, it may discover that click-through rates are higher when specific ad creatives match specific market trends. The analysis results become output data for hypothesis generation.

[1167] Step 4: Propose hypotheses and countermeasures

[1168] The server formulates hypotheses based on the analysis results of the generative AI model. For example, it may formulate a hypothesis that "advertising creatives tailored to specific market trends are effective." The server then proposes countermeasures based on the hypothesis. For example, it may propose a countermeasure such as "In the next advertising campaign, use creatives that are tailored to current market trends." The proposed countermeasures become output data for the execution phase.

[1169] Step 5: Emotion Recognition

[1170] The device uses a camera and microphone to collect the user's facial expressions and voice while they are browsing a website. The emotion engine analyzes the collected data and recognizes the user's emotions. For example, if the user is smiling, it will recognize that they are feeling happy. The recognized emotion data becomes input data for proposing promotional measures.

[1171] Step 6: Propose promotional measures

[1172] The server proposes promotional measures appropriate for the user based on the results of the emotion engine. For example, if the user feels happy or excited, it displays information about new product releases or special discounts. If the user feels dissatisfied or angry, it displays support information for resolving the problem or an apology message. The proposed promotional measures become output data for the user.

[1173] (Application example 1)

[1174] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1175] The traditional PDCA cycle for promotional measures required a lot of time and effort to collect data, analyze it, propose hypotheses, and implement improvement actions. It was also difficult to propose promotional measures that took user emotions into account, making it difficult to achieve effective marketing. This made it difficult to implement promotional measures quickly and effectively, reducing the efficiency of corporate marketing activities.

[1176] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1177] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for inputting results such as acquisitions into the generative AI, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, means for recognizing user emotions, and means for proposing promotional measures according to the user emotions, thereby enabling the implementation of promotional measures quickly and effectively.

[1178] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[1179] "Promotional measures" are specific marketing activities aimed at increasing awareness of products and services and promoting sales.

[1180] The PDCA cycle is a method for continuously improving business processes and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[1181] "Results" are data and results obtained as a result of past activities and measures.

[1182] "Advertising creative" refers to the elements that make up an advertisement, such as its content, design, and message.

[1183] "Market trends" are changes or trends in consumer preferences and behavior, competitor trends, etc. in a particular market.

[1184] A "hypothesis" is a prediction or conjecture about a particular phenomenon or problem, and is a premise whose validity is confirmed through verification.

[1185] A "measure" is a specific action or measure to address a specific problem or issue.

[1186] "Lead time" is the time required from the start to the end of a process.

[1187] "Emotion recognition" is a technology that determines a user's emotions from facial expressions, voice, text, etc.

[1188] A "promotional measure proposal" is a proposal for specific marketing activities that should be carried out to achieve a specific objective.

[1189] A system for carrying out the present invention uses generative AI to accelerate the PDCA cycle of promotional measures. Specific embodiments of the system will be described below.

[1190] System Configuration

[1191] The system includes a server, a user device, and a generative AI model. The server is responsible for collecting data, analyzing it, proposing hypotheses, and implementing improvement actions. The user device is used to recognize user emotions and receive promotional proposals.

[1192] Hardware and software used

[1193] Server: A high-performance computer equipped with a database, generative AI models, and an emotion recognition engine.

[1194] User devices: smartphones, tablets, computers, etc.

[1195] Generative AI model: Uses OpenAI API.

[1196] Emotion Recognition Engine: Uses the EmotionRecognizer library.

[1197] Data collection and analysis

[1198] The server collects performance data from ad campaigns, including clicks, conversion rates, and user feedback. This data is then fed into a generative AI model and analyzed against ad creative and market trends.

[1199] Hypothesis and proposed measures

[1200] The generative AI model proposes hypotheses and countermeasures based on collected data and market trends, thereby shortening the lead time for improving promotional measures.

[1201] Emotion recognition and promotional measures

[1202] The user device recognizes emotions using user input data (e.g., feedback and comments). The emotion recognition engine identifies the user's emotions in real time and proposes promotional measures based on those emotions. For example, if the user is dissatisfied, it can provide special discount information or support information for resolving the problem.

[1203] Specific examples

[1204] User input data: "I don't find recent ads very appealing."

[1205] Perceived Emotion: Frustration

[1206] Proposed promotion: "Test new ad creative designs. Offer special discount information."

[1207] Prompt Sentence Examples

[1208] Advertising campaign performance data: {ad_data}

[1209] User emotion: {emotion}

[1210] Based on this, please propose the best promotional measures.

[1211] In this way, by implementing a system that combines generative AI and an emotion recognition engine, it becomes possible to implement promotional measures quickly and effectively.

[1212] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1213] Step 1:

[1214] The server collects performance data of the advertising campaign, specifically, data such as the number of clicks, conversion rate, user feedback, etc. from a database. The input is the performance data of the advertising campaign, and the output is a set of collected data.

[1215] Step 2:

[1216] The server loads the collected performance data into the generative AI model. Specifically, the data is input using OpenAI's API and analyzed against advertising creatives and market trends. The input is the collected performance data, and the output is the analysis results from the generative AI model.

[1217] Step 3:

[1218] The server proposes hypotheses and countermeasures based on the analysis results of the generative AI model. Specifically, it generates optimal promotional measures based on the analysis results obtained from the generative AI model. The input is the analysis results of the generative AI model, and the output is proposed hypotheses and countermeasures.

[1219] Step 4:

[1220] The server transmits the hypotheses and proposed measures to the user terminal. Specifically, it notifies the user terminal of the proposed promotion measures. The input is the hypotheses and proposed measures, and the output is a notification to the user terminal.

[1221] Step 5:

[1222] The user terminal collects user input data. Specifically, it acquires user feedback and comments. The input is the user input data, and the output is the collected user data.

[1223] Step 6:

[1224] The user device inputs the collected user data into an emotion recognition engine to recognize the user's emotions. Specifically, it uses the EmotionRecognizer library to determine the user's emotions. The input is the user's input data, and the output is the recognized emotion.

[1225] Step 7:

[1226] The server proposes promotional measures based on the recognized user emotions. Specifically, it uses a generative AI model to generate optimal promotional measures according to the user emotions. The input is the recognized emotions, and the output is the proposed promotional measures.

[1227] Step 8:

[1228] The server transmits the proposed promotional measures to the user terminal. Specifically, it notifies the user terminal of the promotional measures according to the user's emotions. The input is the proposed promotional measures, and the output is the notification to the user terminal.

[1229] Example 2

[1230] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1231] Conventional promotional measures have a long cycle of planning, execution, evaluation, and improvement, which requires a lot of man-hours. It is also difficult to propose appropriate improvement actions based on user emotions, which means that the user experience is not fully improved. This results in the effectiveness of advertising creative not being maximized, and marketing efficiency is declining.

[1232] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1233] In this invention, the server includes means for accelerating the planning, execution, evaluation, and improvement cycle of promotional measures using generative artificial intelligence, means for loading past performance data, means for analyzing the data in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time to improvement actions, means for collecting user emotion data and analyzing it with an emotion engine, and means for proposing improvement actions according to the user's emotions based on the analysis results.This speeds up the PDCA cycle of promotional measures, reduces labor costs, and makes it possible to propose appropriate improvement actions according to the user's emotions.

[1234] "Generative AI" is AI that has the ability to generate new advertising creatives and strategies based on past data and market trends.

[1235] "Promotional measures" are plans and activities to increase awareness of and promote sales of a product or service.

[1236] The "plan-do-evaluate-improve cycle" is a series of processes for planning promotional measures, implementing them, evaluating the results, and improving them.

[1237] "Performance data" refers to data on the results and performance of promotional measures taken in the past.

[1238] "Advertising creative" refers to the creative elements of an advertisement, such as its content, design, and message.

[1239] "Market trends" refers to information such as current market conditions and trends, and consumer behavior patterns.

[1240] A "hypothesis" is a premise for making predictions or explanations about a particular problem or phenomenon.

[1241] A "measure" is a specific action or plan to address a particular problem or issue.

[1242] "Lead time" is the time it takes from planning to executing a promotional measure.

[1243] "Emotion data" refers to data relating to emotions obtained from the user's facial expressions, voice, etc.

[1244] An "emotion engine" is a technology that analyzes a user's emotional data and recognizes those emotions.

[1245] "Improvement actions" are specific improvement measures implemented based on user sentiment and the results of promotional measures.

[1246] This invention is a system that uses generative artificial intelligence to accelerate the planning, execution, evaluation, and improvement cycle of promotional measures and proposes improvement actions based on user emotions. Specific embodiments of this system are described below.

[1247] System configuration

[1248] Hardware and Software

[1249] 1. Server

[1250] The server collects past performance data from a database and analyzes the data using a generative artificial intelligence model. Specifically, it uses a cloud-based database service (e.g., Amazon RDS or Google Cloud SQL).

[1251] The server runs a generative artificial intelligence model (e.g., OpenAI's GPT-4) and proposes hypotheses and countermeasures based on advertising creative and market trends.

[1252] 2. Terminal

[1253] The device collects the user's facial expression and voice data and sends it to the emotion engine. Specifically, the device collects data using a camera and microphone.

[1254] The device uses an emotion engine (e.g., Microsoft Azure's Emotion API) to analyze the collected data and recognize the user's emotions.

[1255] 3. Users

[1256] The user checks the direction and measures of the advertising creative proposed by the server and implements them, specifically by setting up the advertising campaign and producing the creative.

[1257] The user receives suggestions from the server and takes necessary improvement actions, for example, watching a video on how to use the product.

[1258] Data processing and calculation

[1259] 1. Data collection and preprocessing

[1260] The server collects historical ad creative performance data and market trend data from the database, and the collected data is preprocessed using Python's Pandas library, which performs missing value imputation and data normalization.

[1261] 2. Analysis using generative AI models

[1262] The server inputs the pre-processed data into a generative AI model for analysis, which then generates the next direction for advertising creative and measures to address market trends.

[1263] 3. Emotion Data Analysis

[1264] The device sends the user's facial expressions and voice data to the emotion engine, which analyzes their emotions. The analysis results are sent to the server, which then suggests improvement actions based on the user's emotions.

[1265] Specific examples

[1266] Ad creative suggestion system using generative AI

[1267] Example prompt: "Based on ad creative performance data from the past six months and current market trend data, please suggest a direction for our next advertising campaign."

[1268] Improvement action suggestion system using emotion engine

[1269] Example prompt: "If you identify a user as confused, what remedial action would you suggest for that user?"

[1270] This enables the system to speed up the PDCA cycle of promotional measures, reduce labor costs, and suggest appropriate improvement actions based on user emotions.

[1271] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1272] Step 1:

[1273] Data collection

[1274] The server collects historical ad creative performance data and market trend data from a database. Specifically, it retrieves the data using a cloud-based database service (e.g., Amazon RDS or Google Cloud SQL). The input is the database query results, and the output is the collected dataset.

[1275] Step 2:

[1276] Data Preprocessing

[1277] The server preprocesses the collected data. Specifically, it uses the Python Pandas library to impute missing values ​​and normalize the data. The input is the collected dataset, and the output is the preprocessed dataset.

[1278] Step 3:

[1279] Analysis using generative AI models

[1280] The server inputs the preprocessed data into a generative AI model (e.g., OpenAI's GPT-4) for analysis. The input is the preprocessed dataset, and the output is the analysis results, which are the direction of advertising creative and measures against market trends. Specifically, the server inputs a prompt statement into the generative AI model and obtains the analysis results.

[1281] Step 4:

[1282] Proposal Generation

[1283] The server generates specific ad creative proposals based on the analysis results obtained from the generative AI model. The input is the analysis results of the generative AI model, and the output is specific ad creative proposals. Specific actions include generating new ad copy and visual ideas.

[1284] Step 5:

[1285] Implementing the proposal

[1286] The user confirms and implements the direction and measures of the advertising creative proposed by the server. The input is the proposal from the server, and the output is the implemented advertising campaign. Specific operations include setting up the advertising campaign and producing the creative.

[1287] Step 6:

[1288] Feedback of results

[1289] The server again collects performance data of the executed advertising creative and feeds it back to the generative AI model. The input is the performance data of the executed advertising campaign, and the output is an updated generative AI model. Specifically, the collected data is preprocessed again and input into the generative AI model.

[1290] Step 7:

[1291] Collecting user emotion data

[1292] The device collects the user's facial expression and voice data and sends it to the emotion engine. The input is the user's emotion data collected through the camera and microphone, and the output is the data sent to the emotion engine. Specifically, the device collects data using the camera and microphone.

[1293] Step 8:

[1294] Emotional Data Analysis

[1295] The server uses an emotion engine (for example, Microsoft Azure's Emotion API) to analyze the collected data and recognize the user's emotions. The input is the emotion data sent from the device, and the output is the user's emotional state as an analysis result. Specifically, the data is input into the emotion engine and the analysis results are obtained.

[1296] Step 9:

[1297] Proposal of improvement actions

[1298] The server proposes improvement actions based on the user's emotions based on the recognition results from the emotion engine. The input is the analysis result of the emotion engine, and the output is a proposal for a specific improvement action. As a specific operation, if the user is confused, it will suggest providing a video on how to use the product.

[1299] Step 10:

[1300] Implementing the proposal

[1301] The user receives the suggestion from the server and performs the necessary action. The input is the suggestion for improvement action from the server, and the output is the implemented improvement action. The specific action is to watch a video on how to use the product.

[1302] Step 11:

[1303] Feedback of results

[1304] The server collects the results of the user's implementation of the suggestions and feeds them back to the emotion engine. The input is the result of the implemented improvement action, and the output is an updated emotion engine. Specifically, the collected data is preprocessed again and input into the emotion engine.

[1305] (Application example 2)

[1306] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1307] In conventional promotional measures, the process of analyzing performance data of advertising creatives and market trend data, and proposing the direction of the next advertising creative and measures to address market trends, takes time and effort, so there is a need to speed up the PDCA cycle. Also, there is the issue of how difficult it is to properly recognize user emotions and propose corresponding improvement actions.

[1308] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1309] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, means for recognizing user emotions using an emotion recognition engine, and means for proposing improvement actions based on the recognized emotions. This makes it possible to speed up the PDCA cycle of promotional measures and quickly propose appropriate improvement actions according to the user's emotions.

[1310] "Generative AI" is artificial intelligence that generates new advertising creatives and strategies based on past data and market trends.

[1311] The PDCA cycle is a method for improving business operations and processes by repeating four steps: Plan, Do, Check, and Act.

[1312] "Ad creative" refers to the visual and content elements of an advertisement, such as its design, content, and message.

[1313] "Market trends" are data and trends that indicate the current market situation, such as consumer preferences and behavior in a particular market, and the actions of competitors.

[1314] An "emotion recognition engine" is a technology for recognizing and analyzing emotions from a user's facial expressions, voice, etc.

[1315] "Improvement actions" are specific improvements or measures proposed based on the analysis results and user sentiment.

[1316] "Lead time" refers to the time it takes from the start to the end of a process.

[1317] A "hypothesis" is a premise or way of thinking that allows one to make predictions or explanations about a particular problem or phenomenon.

[1318] "Countermeasures" are specific actions or measures taken to address a specific problem or issue.

[1319] The system for implementing this invention uses generative AI to accelerate the PDCA cycle of promotional measures, recognizes user emotions, and proposes appropriate improvement actions. Specific embodiments of this system are described below.

[1320] The server uses generative AI to read past ad creative performance data and market trend data to accelerate the PDCA cycle of promotional initiatives. This makes it possible to propose the direction of the next ad creative and measures to address market trends. Specifically, it uses OpenAI's API to analyze the data and generate hypotheses and measures.

[1321] The server also uses an emotion recognition engine to recognize the user's emotions. It receives the user's image and voice data as input and analyzes their emotions using EmotionRecognizer. Based on the recognized emotions, the server suggests appropriate improvement actions. For example, if the user is confused, it provides instructions on how to use the product, and if the user is satisfied, it suggests additional purchases.

[1322] The specific hardware configuration of this system includes a server, a user device (such as a smartphone or PC), and a camera and microphone to run the emotion recognition engine. The software uses OpenAI's API, EmotionRecognizer, and a database management system.

[1323] As a concrete example, let's say the performance data for an ad creative shows that "click-through rates are declining." Based on this data, the prompt to suggest the direction of the next ad creative is as follows:

[1324] Ad creative performance data: Click-through rates are declining. Please suggest the next direction for ad creative and measures to address market trends.

[1325] By inputting this prompt into OpenAI's API, the next direction for advertising creative and measures to address market trends are suggested, accelerating the PDCA cycle for promotional measures and enabling the rapid proposal of appropriate improvement actions based on user emotions.

[1326] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1327] Step 1:

[1328] The server collects historical ad creative performance data and market trend data, including retrieving data from a database and / or external APIs. The input is the ad creative performance data and market trend data, and the output is a consolidated dataset of these data.

[1329] Step 2:

[1330] The server inputs the collected data into generative AI to generate the next direction for advertising creative and countermeasures for market trends. Specifically, it uses OpenAI's API to analyze the data and generate hypotheses and countermeasures. The input is the integrated dataset, and the output is the generated hypotheses and countermeasures.

[1331] Step 3:

[1332] Based on the generated hypotheses and measures, the server creates a specific action plan to accelerate the PDCA cycle of promotional measures. The input is the generated hypotheses and measures, and the output is a specific action plan.

[1333] Step 4:

[1334] The user device recognizes the user's emotions using an emotion recognition engine. Specifically, it acquires the user's image and voice data through a camera and microphone, and analyzes the emotions using EmotionRecognizer. The input is the user's image and voice data, and the output is the recognized emotion.

[1335] Step 5:

[1336] The server then suggests appropriate improvement actions based on the recognized emotions. For example, if the user is confused, it provides product usage instructions, and if the user is satisfied, it suggests additional purchases. The input is the recognized emotions, and the output is the suggested improvement actions.

[1337] Step 6:

[1338] The server sends the proposed improvement action to the user terminal and notifies the user. The input is the proposed improvement action, and the output is the notification to the user.

[1339] Step 7:

[1340] The user checks the notified improvement actions and executes them as necessary. The input is the notified improvement action, and the output is the result of the executed improvement action.

[1341] This will speed up the PDCA cycle for promotional measures and make it possible to quickly propose appropriate improvement actions based on user emotions.

[1342] Example 3

[1343] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1344] Conventional promotional measures have a long lead time for the PDCA cycle, making it difficult to take quick action to improve. In addition, the creation of advertising creatives based on user emotions is done manually, which is inefficient and makes it difficult to provide advertisements that match user emotions.

[1345] The specification process by the specification processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes a means for accelerating the PDCA cycle of promotional measures using a generative AI, a means for inputting the results of acquisitions and the like into the generative AI, a means for analyzing the results in comparison with advertising creatives and market trends, a means for proposing hypotheses and measures based on the analysis results, a means for shortening the lead time until improvement actions, a means for recognizing user emotions using an emotion engine, and a means for generating advertising creatives according to the user emotions. This enables the PDCA cycle to be executed quickly and advertising creatives that are in line with the user's emotions to be provided.

[1346] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[1347] The PDCA cycle is a method for continuously improving business operations and processes by repeating the four steps of Plan, Do, Check, and Act.

[1348] "Lead time" is the time required from the start to the end of a process.

[1349] An "emotion engine" is a technology that analyzes a user's text and voice data and recognizes their emotions.

[1350] "Advertising creative" refers to elements that appeal to users visually or audibly, such as the content, design, and music of the advertisement.

[1351] "Market trends" refer to changes in consumer preferences and behaviors in a particular market, or trends therein.

[1352] A "hypothesis" refers to a premise or prediction that is used to propose an explanation or solution to a phenomenon or problem.

[1353] "Countermeasures" refer to specific actions or measures taken to address a specific problem or issue.

[1354] "Achievements" refers to results or accomplishments achieved in the past.

[1355] "User" refers to an individual or organization that uses the system or service.

[1356] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures and generate advertising creatives that respond to user emotions. A specific embodiment of this system is described below.

[1357] The server uses a generative AI model to accelerate the PDCA cycle of promotional measures. Specifically, the server receives data sent by users and analyzes it using a generative AI model (e.g., a general-purpose natural language processing model). Based on the analysis results, it proposes hypotheses and countermeasures and determines the priority of improvement actions. At this time, it also references past data stored in a database (e.g., a relational database management system).

[1358] Furthermore, the server uses an emotion engine to recognize the user's emotions. It analyzes text and voice data entered by the user through a device (e.g., a smartphone or PC) to obtain emotion recognition results. For example, an emotion analysis API can be used as the emotion engine.

[1359] Based on the emotion recognition results, the server uses a generative AI model to generate ad creatives. For example, if the user is feeling happy or excited, the server generates an ad using bright colors and lively music. On the other hand, if the user is feeling relieved or satisfied, the server generates an ad using calm colors and tranquil music.

[1360] As a concrete example, consider the case where a user posts a review of a new product. The user uses a device to input a review such as, "This product is very easy to use and has a great design." The server receives the review and uses a generative AI model to generate a prediction such as, "Improving the design of this product will increase user satisfaction by 20%." Next, the server uses an emotion engine to recognize the user's emotions and determine that the user is excited. Finally, the server uses the generative AI model to input a prompt such as, "Generate an ad creative that will enhance this user's excitement," and generates an ad featuring bright colors and lively music.

[1361] Example prompt sentence:

[1362] "Analyze user reviews, recognize emotions, and generate ad creatives that reflect those emotions."

[1363] "Predict the effectiveness of proposed measures and prioritize remedial actions"

[1364] In this way, the server, terminal, and user work together to execute each processing step of the system, enabling the rapid execution of the PDCA cycle and the provision of advertising creatives that are in line with the user's emotions. The flow of the specific processing in Example 3 will be described with reference to Figure 21.

[1365] Step 1:

[1366] The user inputs data through a device. The user uses a smartphone or PC to input reviews, feedback, and other text data. For example, a user might write a review of a new product, saying, "This product is very easy to use and has a great design." The input data is sent to the server via an HTTP request.

[1367] Step 2:

[1368] The server receives input data and predicts the effectiveness of countermeasures using a generative AI model. The server receives data sent by users and analyzes the data using a generative AI model (for example, a natural language processing model). Specifically, it analyzes the content of user reviews and generates a predicted result, such as "Improving the design of this product will increase user satisfaction by 20%." The input is the user review, and the output is the predicted result.

[1369] Step 3:

[1370] The server determines the priority of improvement actions based on the prediction results. The server determines the priority of improvement actions based on the prediction results. In doing so, it also refers to past data stored in a database (e.g., a relational database management system). For example, if it predicts that design improvements will be most effective, the server sets a priority such as "implement design improvements as the top priority." The inputs are the prediction results and past data, and the output is the priority.

[1371] Step 4:

[1372] The server uses an emotion engine to recognize the user's emotions. The server uses an emotion engine (for example, an emotion analysis API) to recognize the user's emotions. It analyzes the text data entered by the user and obtains emotion recognition results. For example, from a review that says, "This product is very easy to use and has a great design," it is recognized that the user is feeling happy or excited. The input is the user's review, and the output is the emotion recognition result.

[1373] Step 5:

[1374] The server generates ad creatives based on the emotion recognition results. The server generates ad creatives using a generative AI model based on the emotion recognition results. For example, if a user is feeling happy or excited, it generates an ad using bright colors and lively music. Specifically, a prompt such as "Generate ad creatives that will enhance this user's excitement" is input into the generative AI model. The inputs are the emotion recognition results and the prompt, and the output is the ad creatives.

[1375] Step 6:

[1376] The server sends the generated ad creative to the terminal. The server sends the generated ad creative to the user's terminal. For example, if the generated ad is in video format, the video data is sent to the terminal. The input is the ad creative, and the output is transmission to the terminal.

[1377] Step 7:

[1378] The user reviews the ad creative on their device. The user reviews the ad creative received on their device. For example, they may watch a video ad on their smartphone and provide feedback on the content. The input is the ad creative, and the output is the user's feedback.

[1379] (Application example 3)

[1380] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1381] Conventional promotional measures have the problem that the PDCA cycle takes time to execute, making it difficult to take quick action to improve. In addition, it is not possible to generate advertising creatives that correspond to the user's emotions, making it difficult to deliver effective advertisements.

[1382] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means. In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time until improvement actions, means for recognizing user emotions using an emotion engine, and means for generating advertising creatives in accordance with the user emotions. This enables rapid execution of the PDCA cycle and effective ad delivery in accordance with user emotions.

[1383] "Generative AI" is an artificial intelligence technology that generates new information and content based on data.

[1384] The PDCA cycle is a method for improving business processes and projects by repeating the four steps of Plan, Do, Check, and Act.

[1385] "Lead time" is the time it takes from the start to the end of a process.

[1386] An "emotion engine" is a technology that recognizes emotions from a user's facial expressions, voice, etc.

[1387] "Advertising creative" is a term that refers to elements such as the content, design, and message of an advertisement.

[1388] A "market trend" is a tendency that indicates changes in consumer preferences and behavior in a particular market.

[1389] A "hypothesis" is a premise or assumption used to explain or predict a phenomenon or problem.

[1390] A "measure" is a specific action or measure to address a specific problem or issue.

[1391] "User emotion" refers to a psychological state such as joy, excitement, relief, or satisfaction that a user feels in response to a particular situation or stimulus.

[1392] The system for implementing this invention uses generative AI to accelerate the PDCA cycle of promotional measures and generate advertising creatives that respond to user emotions. A specific embodiment of this system is described below.

[1393] System Program

[1394] This system is implemented using the following hardware and software.

[1395] Hardware: Smartphone camera and microphone

[1396] Software: OpenCV, Keras, playsound, PIL (Python Imaging Library)

[1397] Processing flow

[1398] The server captures video in real time using the smartphone's camera. It uses OpenCV to detect faces and converts the detected facial area to grayscale. Next, it predicts the user's emotion from the facial area using an emotion recognition model pre-trained with Keras. Depending on the predicted emotion, it generates an appropriate advertising creative (image) using PIL and plays the corresponding music using playsound. Finally, the advertising creative is displayed on the smartphone screen.

[1399] Specific examples

[1400] For example, if a user is using a smartphone and the camera captures the user's face, and the emotion recognition model predicts that the user is "happy," an advertisement with bright images and lively music will be displayed.

[1401] Prompt Sentence Examples

[1402] An example of a prompt sentence to input to the generative AI model is as follows:

[1403] Develop a smartphone application that recognizes user emotions in real time and generates and displays advertising creatives according to those emotions. If the user is happy, display an advertisement using bright colors and lively music, and if the user is relieved, display an advertisement using calm colors and gentle music.

[1404] In this way, the server can quickly execute the PDCA cycle and effectively deliver advertisements according to the user's emotions.

[1405] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1406] Step 1:

[1407] The server captures video in real time using the smartphone camera, the input is video data from the smartphone camera, and the output is real-time video frames, which are used in subsequent processing steps.

[1408] Step 2:

[1409] The server uses OpenCV to detect faces from video frames. The input is the video frame acquired in step 1, and the output is the coordinate information of the detected face area. Specifically, the video frame is converted to grayscale and the face area is identified by applying a face detection algorithm.

[1410] Step 3:

[1411] The server converts the detected face area to grayscale and performs preprocessing to input it into the emotion recognition model. The input is the coordinate information of the face area obtained in step 2 and the video frame, and the output is preprocessed face image data. Specifically, the face area is cut out, resized to 48x48 pixels, and normalized.

[1412] Step 4:

[1413] The server uses an emotion recognition model pre-trained with Keras to predict the user's emotion from the pre-processed facial image data. The input is the facial image data obtained in step 3, and the output is the emotion prediction result. Specifically, the facial image data is input into the emotion recognition model and the probability distribution of emotions is obtained.

[1414] Step 5:

[1415] The server generates advertising creatives according to the predicted emotions. The input is the emotion prediction result obtained in step 4, and the output is the generated advertising creatives (images) and music files. Specifically, if the emotion is "joy," it selects brightly colored images and lively music, and if the emotion is "relief," it selects calmly colored images and tranquil music.

[1416] Step 6:

[1417] The server displays the generated ad creative on the smartphone screen and plays the corresponding music. The input is the ad creative and music file obtained in step 5, and the output is the ad displayed to the user and the music played. Specifically, PIL is used to display the image and plays the music using playsound.

[1418] In this way, the server can generate and display advertising creatives in real time according to the user's emotions, achieving effective advertising delivery.

[1419] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1420] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1421] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are listed.

[1422] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1423] [Third embodiment]

[1424] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1425] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1426] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1427] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1428] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1429] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1430] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1431] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1432] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1433] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1434] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1435] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1436] "Example 1"

[1437] The present invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures. Specifically, the generative AI is made to read the results of acquisitions, etc., and

[1438] We analyze the results in light of current trends and market trends. Based on the analysis results, we propose hypotheses and measures, shortening the lead time for improvement actions. This will result in a reduction of 10 hours / month x 12 months x 4 measures = 480 hours / year of man-hours.

[1439] "Example 2"

[1440] As a concrete example, generative AI reads performance data from past ad creatives and market trend data. Based on this data, it proposes the direction of the next ad creative and measures to address market trends. The proposed measures are implemented as part of the PDCA cycle, and the results are fed back to the generative AI. This speeds up the PDCA cycle and reduces labor costs.

[1441] "Example 3"

[1442] Furthermore, generative AI predicts the effectiveness of proposed measures and prioritizes improvement actions based on these predictions, significantly shortening lead times and enabling faster PDCA cycles.

[1443] The processing flow of each embodiment will be described below.

[1444] "Example 1"

[1445] Step 1: Load the generative AI with past achievements, such as acquisitions. This allows the AI ​​to learn based on past data.

[1446] Step 2: Analyze against ad creative and market trends. This analysis will help you understand the current market situation and how your ads are performing.

[1447] Step 3: Based on the analysis results, hypotheses and countermeasures are proposed, which then lead to the creation of the next action plan.

[1448] Step 4: Reduce the lead time to improvement actions, which speeds up the PDCA cycle and reduces man-hours.

[1449] "Example 2"

[1450] Step 1: The generative AI reads past performance data of advertising creatives and market trend data, allowing the AI ​​to learn from past data.

[1451] Step 2: Based on this data, the generative AI proposes the next direction for advertising creative and measures to address market trends. This allows the next action plan to be formulated. Step 3: The proposed measures are implemented as part of the PDCA cycle, and the results are fed back to the generative AI. This speeds up the PDCA cycle and reduces labor costs.

[1452] "Example 3"

[1453] Step 1: Generative AI predicts the effectiveness of the proposed measures, which in turn predicts the effectiveness of the next action plan.

[1454] Step 2: The generative AI prioritizes improvement actions based on the prediction results, enabling the most effective actions to be implemented first.

[1455] Step 3: This significantly reduces lead times and enables a faster PDCA cycle.

[1456] Example 1

[1457] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1458] The traditional plan-do-check-act (PDCA) cycle for promotional measures required a great deal of time and effort for each step, such as collecting data, analyzing it, formulating hypotheses, and implementing measures, making it difficult to respond quickly. Furthermore, data cleansing and analysis required specialized knowledge, making efficient operation difficult. This delayed prompt improvement actions to maximize the effectiveness of promotional measures, potentially resulting in a decline in market competitiveness.

[1459] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1460] In this invention, the server includes means for accelerating the planning, execution, evaluation, and improvement cycle of promotional measures using generative artificial intelligence, means for loading performance data, means for analyzing the data against advertising materials and market trends, means for cleansing the performance data, filling in missing values, and removing outliers, means for analyzing the performance data using generative artificial intelligence and identifying effective patterns and trends, means for proposing specific measures based on hypotheses, and means for presenting specific improvement actions based on the proposed measures. This enables the PDCA cycle of promotional measures to be executed quickly and efficiently, reducing labor hours and improving market competitiveness.

[1461] "Generative AI" refers to AI that has the ability to generate new information and patterns based on data.

[1462] "Promotional measures" refer to plans and activities to increase awareness of and promote sales of products and services.

[1463] The "Plan-Do-Check-Act cycle (PDCA cycle)" is a method for achieving continuous improvement in business processes and projects by repeating the four steps of Plan, Do, Check, and Act.

[1464] "Performance data" refers to data obtained as a result of past activities and measures, including the number of clicks, conversion rates, and sales.

[1465] "Advertising materials" refers to the visuals, text, and other creative elements used in an advertising campaign.

[1466] "Market trends" refers to current market conditions and trends, such as consumer behavior and competitor actions in a particular market.

[1467] "Cleansing" refers to the process of preparing data for analysis, such as filling in missing values ​​and removing outliers.

[1468] "Patterns and trends" refer to certain regularities or characteristics found in data.

[1469] A "hypothesis" is a prediction or assumption made based on data analysis.

[1470] "Countermeasures" refer to specific actions or measures implemented based on a hypothesis.

[1471] "Improvement actions" refer to specific improvement measures implemented based on analysis results and hypotheses.

[1472] This invention is a system that uses generative artificial intelligence to accelerate the plan-do-check-act (PDCA) cycle of promotional measures. Specific embodiments of this system are described below.

[1473] System configuration

[1474] Hardware and Software Configuration

[1475] The server provides computational resources for running a generative artificial intelligence model (e.g., OpenAI's GPT-4). The server includes a database, a data processing unit, and a generative artificial intelligence model.

[1476] The terminal provides an interface for users to access the system, upload data, check analysis results, etc. Terminals include PCs, tablets, smartphones, etc.

[1477] Program processing

[1478] Data collection and preprocessing

[1479] Users upload promotional campaign performance data into the system in CSV file format, including the number of clicks, conversion rate, sales, and other data.

[1480] The server receives the uploaded data and stores it in a database, where it is cleansed, missing values ​​are filled, and outliers are removed.

[1481] Analyzing the data

[1482] The server uses a generative artificial intelligence model to analyze performance data, specifically matching it with advertising materials and market trends to identify effective patterns and trends.

[1483] The server identifies that particular advertising materials have demonstrated high conversion rates in particular market segments.

[1484] Hypothesis and proposed measures

[1485] The server generates a hypothesis based on the analysis results, for example, "Advertising material B shows a high conversion rate among young people."

[1486] The server then proposes specific measures based on the hypothesis, such as "Use advertising material B, which targets younger generations, in the next campaign."

[1487] Suggestion of improvement actions

[1488] The server then presents specific improvement actions to the user based on the proposed measures, such as "Use advertising material B in the next campaign and target younger generations."

[1489] The user checks the proposed improvement actions and reflects them in the next promotional measures. Specifically, the user plans and executes a new campaign using advertising material B.

[1490] Examples of concrete examples and prompts

[1491] Specific examples

[1492] User-Provided Data:

[1493] Ad Material A: 5,000 clicks, 2.5% conversion rate

[1494] Ad Material B: 3,000 clicks, conversion rate 3.0%

[1495] Market Trend Data: Online Shopping Increases Among Young People

[1496] Prompt Sentence Examples

[1497] "Based on the performance data of advertising campaigns over the past six months, please analyze the most effective advertising materials and market trends, and propose hypotheses and measures for the next campaign."

[1498] In this way, the server can propose quick and specific improvement actions to the user, accelerating the PDCA cycle.

[1499] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1500] Step 1:

[1501] Collection of performance data

[1502] Users upload promotional campaign performance data into the system in CSV file format.

[1503] Input: A CSV file containing performance data such as clicks, conversion rates, and sales.

[1504] The server receives the uploaded CSV file and stores it in a database.

[1505] Output: Actual data stored in a database.

[1506] Step 2:

[1507] Data Preprocessing

[1508] The server cleanses the stored performance data, specifically by filling in missing values ​​and removing outliers.

[1509] Input: Actual data stored in the database.

[1510] The server will fill in missing values ​​with the average value, remove outliers, and standardize the data format. For example, it will standardize the date format to "YYYY-MM-DD."

[1511] Output: Cleansed actual data.

[1512] Step 3:

[1513] Analyzing the data

[1514] The server analyzes the performance data using a generative artificial intelligence model (e.g., GPT-4).

[1515] Input: Cleansed actual data.

[1516] The server compares advertising materials and market trends with performance data to identify effective patterns and trends, such as identifying whether certain advertising materials result in higher conversion rates among certain market segments.

[1517] Output: Effective patterns and trends as a result of the analysis.

[1518] Step 4:

[1519] Hypothesis and proposed measures

[1520] The server generates a hypothesis based on the analysis results, for example, "Advertising material B shows a high conversion rate among young people."

[1521] Input: Effective patterns and trends as a result of analysis.

[1522] The server then proposes specific measures based on the hypothesis, such as "Use advertising material B, which targets younger generations, in the next campaign."

[1523] Output: Hypothesis and specific measures.

[1524] Step 5:

[1525] Suggestion of improvement actions

[1526] The server then presents specific improvement actions to the user based on the proposed measures, such as "Use advertising material B in the next campaign and target younger generations."

[1527] Input: Hypothesis and specific measures.

[1528] The server generates and presents remedial actions to the user.

[1529] Output: Specific improvement actions presented to the user.

[1530] Step 6:

[1531] Taking improvement actions

[1532] The user checks the proposed improvement actions and reflects them in the next promotional measures. Specifically, the user plans and executes a new campaign using advertising material B.

[1533] Input: The specific improvement actions suggested to the user.

[1534] The user runs a new campaign based on the improvement actions.

[1535] Output: The new campaign that was executed.

[1536] (Application example 1)

[1537] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1538] Traditional promotional measures require a great deal of time and effort in each phase of the PDCA cycle, and require rapid response, particularly in analyzing advertising creatives and market trends, formulating hypotheses, and implementing measures. However, there is a lack of systems to efficiently carry out these processes, which results in long lead times for improvement actions, making it difficult to implement effective promotional measures. In addition, there is no way to monitor advertising campaign performance in real time and quickly implement improvement actions, which makes it difficult to maximize advertising effectiveness.

[1539] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1540] In this invention, the server includes a means for accelerating the PDCA cycle of promotional measures using generative AI, a means for inputting results such as acquisitions into the generative AI, and a means for analyzing the results in comparison with advertising creatives and market trends. This makes it possible to propose hypotheses and measures based on the analysis results, shortening the lead time until improvement actions are taken. In addition, by including a means for monitoring the performance of advertising campaigns in real time and quickly taking improvement actions as an application installed on a smartphone, it is possible to maximize advertising effectiveness.

[1541] "Generative AI" refers to artificial intelligence that generates new information and content based on data.

[1542] "Promotional measures" are plans and activities to increase awareness of and promote sales of a product or service.

[1543] The PDCA cycle is a method for continuously improving business processes and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[1544] "Results" refers to the data and results obtained as a result of past activities and measures.

[1545] "Advertising creative" refers to the elements that make up an advertisement, such as its content, design, and message.

[1546] "Market trends" refer to changes and trends in consumer preferences and behavior, competitor trends, etc. in a particular market.

[1547] A hypothesis is a prediction or premise made to find an explanation or solution to a phenomenon or problem.

[1548] "Countermeasures" refer to specific actions or measures taken to address a specific problem or issue.

[1549] "Lead time" is the time it takes from the start to the end of a process.

[1550] A "smartphone" is a highly functional mobile phone that can connect to the Internet and use applications.

[1551] An "application" is software designed to provide a specific function or service.

[1552] "Performance" refers to the results or effects of a particular activity or measure.

[1553] "Real-time" refers to the ability to process and display ongoing events and data immediately.

[1554] An "improvement action" is a specific improvement measure implemented to solve a specific problem or issue.

[1555] The system for carrying out this invention uses generative AI to accelerate the PDCA cycle of promotional measures. A specific embodiment of this system will be described below.

[1556] System Configuration

[1557] The system includes a server, a smartphone, a generative AI model, a database, and an internet connection. The server runs the generative AI model and accesses the database to retrieve advertising performance data. The smartphone runs an application that allows users to monitor advertising campaign performance in real time and take action to improve.

[1558] Program processing

[1559] The server processes the data in the following steps:

[1560] 1. Data loading:

[1561] The server reads advertising performance data from the database, including click-through rates (CTR), conversion rates (CVR), advertising costs, etc.

[1562] 2. Data Analysis:

[1563] The server uses a generative AI model to analyze advertising performance data, specifically by comparing it with advertising creative and market trends, and then propose hypotheses and countermeasures.

[1564] 3. Improvement suggestions:

[1565] The server uses a generative AI model to suggest improvement actions based on the analysis results, which are then communicated to the user via a smartphone application.

[1566] Hardware and software used

[1567] Hardware: Servers, smartphones

[1568] Software: Python, OpenAI API, Pandas

[1569] Data processing and calculation

[1570] The server uses Python to read advertising performance data from the database and processes the data using Pandas. It then analyzes the data using a generative AI model (OpenAI API) to generate hypotheses and countermeasures. The generated hypotheses and countermeasures are then notified to the user via a smartphone application.

[1571] Specific examples

[1572] For example, assume that the following advertising performance data exists:

[1573] Click-through rate: 2.5%

[1574] Conversion rate: 1.2%

[1575] Advertising cost: 500,000 yen

[1576] Based on this data, the server inputs the following prompt sentence into the generative AI model:

[1577] Example prompt sentence:

[1578] Based on the advertising performance data below, please analyze the advertising creative and market trends and suggest areas for improvement.

[1579] Advertising performance data:

[1580] Click-through rate: 2.5%

[1581] Conversion rate: 1.2%

[1582] Advertising cost: 500,000 yen

[1583] Based on this prompt, the generative AI model analyzes advertising creatives and market trends and suggests improvements. The proposed improvements are notified to the user via a smartphone application, allowing them to quickly take action.

[1584] In this way, the system can accelerate the PDCA cycle of promotional initiatives and maximize the performance of advertising campaigns.

[1585] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1586] Step 1:

[1587] The server reads ad performance data from the database. As input, it requires database connection information and a query. The server executes the query to obtain data such as click-through rate (CTR), conversion rate (CVR), and ad cost, and stores it in a Pandas data frame. The output is a data frame containing ad performance data.

[1588] Step 2:

[1589] The server generates a prompt to input the acquired advertising performance data into the generative AI model. Advertising performance data is required as input. The server converts the data frame into text format and generates a prompt like the following: "Based on the following advertising performance data, please analyze the advertising creative and market trends and suggest improvements. Advertising performance data: Click-through rate: 2.5%, Conversion rate: 1.2%, Advertising cost: 500,000 yen." The output is the generated prompt.

[1590] Step 3:

[1591] The server uses a generative AI model (OpenAI API) to input a prompt text and analyze ad creatives and market trends. The prompt text is required as input. The server sends the prompt text to the OpenAI API and receives a response from the generative AI model. The output is the analysis results from the generative AI model.

[1592] Step 4:

[1593] The server generates hypotheses and proposed countermeasures based on the analysis results from the generative AI model. The analysis results from the generative AI model are required as input. The server analyzes the analysis results and proposes specific improvement actions. The output is proposed hypotheses and countermeasures.

[1594] Step 5:

[1595] The server notifies the smartphone application of the proposed hypotheses and countermeasures. The proposed hypotheses and countermeasures are required as input. The server sends a notification to the smartphone application, prompting the user to take an improvement action. The output is the improvement action notified to the user.

[1596] Step 6:

[1597] The user checks the notified improvement actions through a smartphone application and executes them as necessary. The input requires a notification from the server. The user checks the notification and executes the improvement action according to the application's instructions. The output is the executed improvement action.

[1598] In this way, the system can accelerate the PDCA cycle of promotional initiatives and maximize the performance of advertising campaigns.

[1599] Example 2

[1600] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1601] Conventional promotional measures have the problem that each step of the PDCA cycle takes a lot of time and effort, making it difficult to make quick improvements. In addition, it was difficult to effectively utilize performance data on advertising creatives and market trend data, making it difficult to propose optimal measures. This resulted in the promotional effect not being maximized and resources being wasted.

[1602] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1603] In this invention, the server includes: a means for a user to upload past advertising creative performance data and market trend data to the server; a means for the server to save the uploaded data in a database and perform preprocessing; a means for the server to input the preprocessed data into a generative AI model and send a prompt sentence to the generative AI model; a means for the server to generate proposals and return them to the server; a means for the server to display the proposals to the user; a means for the user to implement the proposed measures and feed back the results to the server; and a means for the server to re-input the feedback data into the generative AI model and reflect it in the next proposal. This speeds up the PDCA cycle, reduces labor hours, and enables optimal measures to be proposed that effectively utilize advertising creative performance data and market trend data.

[1604] "Generative AI" is artificial intelligence that makes new suggestions and predictions based on past data and market trends.

[1605] The PDCA cycle is a method for continuously improving business operations and processes by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[1606] "Promotional measures" are specific activities and strategies for increasing awareness of products and services and promoting sales.

[1607] "Performance data" refers to numerical data that indicates the effectiveness of advertising creatives and promotional measures, and includes the number of clicks, conversion rates, costs, etc.

[1608] "Market trend data" is data that shows current market trends and consumer behavior patterns.

[1609] A "server" is a computer system that stores, processes, and manages data.

[1610] A "database" is a system for efficiently storing, searching, and managing data.

[1611] "Preprocessing" refers to the process of converting data into a format suitable for analysis or model input.

[1612] A "generative AI model" is an artificial intelligence model that generates new suggestions and predictions based on input data.

[1613] A "prompt" is text that gives specific instructions or questions to a generative AI model.

[1614] The "proposal content" refers to the new advertising creative direction and measures for market trends generated by the generative AI model.

[1615] "Feedback" refers to data that is re-entered into the system to reflect the results of the measures implemented in the next proposal.

[1616] This invention is a system that uses generative AI to accelerate the PDCA cycle of promotional measures and reduce man-hours. This system consists of multiple components, including a user, a server, and a generative AI model.

[1617] First, the user uploads past advertising creative performance data and market trend data to the server. This data is generally provided in CSV or Excel file format. The server then stores the uploaded data in a database. For the database, a relational database management system (RDBMS) such as MySQL or PostgreSQL is used.

[1618] The server then preprocesses the stored data. This includes imputing missing values ​​and normalizing the data. These processes can be performed using the Python Pandas library. The preprocessed data is then converted into a format that is easy for the generative AI model to understand. Specifically, it is often converted into JSON format.

[1619] The server inputs the preprocessed data into a generative AI model. For example, natural language processing (NLP) models such as GPT-4 or BERT are used as generative AI models. The server inputs a prompt to the generative AI model. An example of a prompt is, "Based on the performance data of advertising creatives over the past six months and current market trend data, please suggest a direction for the next advertising campaign."

[1620] Based on the input data and prompt text, the generative AI model proposes the next direction for advertising creative and measures to address market trends. The generated proposals are returned to the server in text format. The server receives the proposals returned by the generative AI model and displays them to the user. For display, data visualization tools such as Tableau or Power BI are used.

[1621] The user implements the proposed measures and feeds the results back to the server. The feedback data is then saved in the database and reflected in the next proposal. The server then inputs the feedback data back into the generative AI model, accelerating the PDCA cycle.

[1622] This system will speed up the PDCA cycle, reduce labor costs, and enable the proposal of optimal measures that effectively utilize advertising creative performance data and market trend data.

[1623] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1624] Step 1:

[1625] Users upload historical advertising creative performance data and market trend data to the server.

[1626] Input: Past ad creative performance data (e.g., CSV file), market trend data (e.g., Excel file)

[1627] Output: Data stored on the server

[1628] Specific operation: A user selects a CSV or Excel file through the web interface and clicks the upload button. The server receives the uploaded file and stores it in the database.

[1629] Step 2:

[1630] The server stores the uploaded data in a database and performs preprocessing.

[1631] Input: Uploaded data (CSV file, Excel file)

[1632] Output: Preprocessed data (JSON format)

[1633] How it works: The server reads the data using Python's Pandas library, imputes missing values, normalizes the data, and converts the preprocessed data into a JSON format that is easy for the generative AI model to understand.

[1634] Step 3:

[1635] The server inputs the preprocessed data into the generative AI model and sends the prompt sentence to the generative AI model.

[1636] Input: Preprocessed data (JSON format), prompt statement

[1637] Output: Proposals from the generative AI model (text format)

[1638] Specific operation: The server calls the generative AI model's API and sends the preprocessed data and a prompt. An example of a prompt is, "Based on the performance data of the ad creative over the past six months and current market trend data, please suggest the direction of our next ad campaign."

[1639] Step 4:

[1640] Based on the input data and prompt text, the generative AI model suggests the next direction for advertising creative and measures to address market trends.

[1641] Input: Preprocessed data (JSON format), prompt statement

[1642] Output: Proposal content (text format)

[1643] How it works: The generative AI model uses its internal algorithms to analyze data and generate optimal ad creative direction and countermeasures for market trends. The generated proposals are returned to the server in text format.

[1644] Step 5:

[1645] The server receives the suggestions returned by the generative AI model and displays them to the user.

[1646] Input: Proposal content from the generative AI model (text format)

[1647] Output: The suggestions displayed to the user

[1648] Specific operation: The server receives the proposals, creates a dashboard using a data visualization tool (e.g., Tableau, Power BI), and displays it to the user.

[1649] Step 6:

[1650] The user implements the proposed measures and feeds back the results to the server.

[1651] Input: Implementation result data (e.g. Excel file)

[1652] Output: Feedback data stored on the server

[1653] Specific operation: The user compiles the results of the implemented measures in an Excel file and uploads it to the server. The server receives the uploaded file and stores it in a database.

[1654] Step 7:

[1655] The server then inputs the feedback data back into the generative AI model and reflects it in the next proposal.

[1656] Input: Feedback data (JSON format)

[1657] Output: Data reflected in the next proposal

[1658] Specific operation: The server preprocesses the feedback data and re-inputs it into the generative AI model, which then takes the feedback data into account to generate the next proposal.

[1659] (Application example 2)

[1660] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1661] Conventional promotional measures had the problem of taking a lot of time and effort to analyze advertising creative performance data and market trend data and propose hypotheses and countermeasures. There was also a lack of means to speed up the PDCA cycle and efficiently implement improvement actions. Furthermore, there was no established process for efficient data input using smart devices, or for implementing and providing feedback on proposals using generative AI models.

[1662] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for accelerating the PDCA cycle of promotional measures using a generative AI; means for loading the generative AI with actual results such as acquisitions; means for analyzing the results in comparison with advertising creatives and market trends; means for proposing hypotheses and measures based on the analysis results; means for shortening the lead time to improvement actions; means for inputting performance data and market trend data of past advertising creatives using an application installed on a smart device and inputting prompt statements into the generative AI model; and means for implementing measures in response to the advertising creative direction and market trends proposed by the generative AI model and providing feedback on the results. This makes it possible to speed up the PDCA cycle of promotional measures and significantly reduce labor costs.

[1663] "Generative AI" is artificial intelligence that generates new advertising creatives and strategies based on past data and market trends.

[1664] "Promotional measures" are plans and activities to increase awareness of products and services and promote sales.

[1665] The PDCA cycle is a method for continuously improving business operations by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[1666] "Performance data" refers to data used to measure the effectiveness of advertising creatives, including click-through rates and conversion rates.

[1667] "Market trend data" is data that indicates current market trends and consumer interests.

[1668] A "smart device" is a portable electronic device that can connect to the Internet, including smartphones and smart glasses.

[1669] A "prompt" is an instruction given to generative AI, and serves as the basis for generating the direction and measures for advertising creatives.

[1670] "Feedback" is the process of inputting the results of implemented measures back into the generative AI and reflecting them in the next proposal.

[1671] The system for carrying out this invention uses generative AI to accelerate the PDCA cycle of promotional measures. A specific embodiment of this system will be described below.

[1672] Hardware and software used

[1673] Hardware: Smartphones, smart glasses, servers

[1674] Software: Generative AI models (e.g., OpenAI GPT-4), database management systems (e.g., MySQL, PostgreSQL), data analysis tools (e.g., Tableau, Google Analytics)

[1675] Data processing and calculation

[1676] 1. Data Collection:

[1677] The server retrieves historical performance data of advertising creatives from a database, and also retrieves market trend data from an external API (e.g., Google Trends API).

[1678] 2. Data Preprocessing:

[1679] The server cleans the acquired data, imputes missing values, normalizes the data, and converts it into a format suitable for analysis.

[1680] 3. Data Analysis:

[1681] The server uses data analysis tools to evaluate the performance of past advertising creatives and analyzes market trend data to identify current trends.

[1682] 4. Utilizing generative AI models:

[1683] The server inputs the following prompt sentences into the generative AI model to generate the next direction for advertising creative and measures for market trends.

[1684] Prompt Sentence Examples

[1685] Based on past ad creative performance data and current market trend data, please propose the next ad creative direction and measures to address market trends.

[1686] 5. Implementation and feedback of proposals:

[1687] Using an application installed on a smart device, users can implement the advertising creative direction and market trend measures suggested by the generative AI model. The results are then fed back into the generative AI model, accelerating the PDCA cycle.

[1688] Specific examples

[1689] Example 1: Smartphone application

[1690] Users open the smartphone app and input past ad creative performance data and market trend data. The app uses a generative AI model to suggest the direction of the next ad creative. Users can then review and implement the suggestions, completing the PDCA cycle.

[1691] Example 2: Smart Glasses Application

[1692] The marketer creating the ad creative wears the smart glasses and uses voice commands to access past performance data and market trend data. The smart glasses use a generative AI model to suggest the direction of the next ad creative in real time. The marketer then reviews the suggestions and implements them, completing the PDCA cycle.

[1693] In this way, it is possible to speed up the PDCA cycle of promotional measures and significantly reduce labor costs.

[1694] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1695] Step 1:

[1696] The server retrieves past performance data of ad creatives from a database. The inputs are database connection information and a query, and the output is performance data of ad creatives. Specifically, the server connects to the database and executes the specified query to retrieve the data.

[1697] Step 2:

[1698] The server retrieves market trend data from an external API. The input requires the API endpoint and authentication information, and the output is market trend data. Specifically, the server sends a request to the API and receives market trend data as a response.

[1699] Step 3:

[1700] The server cleans the acquired data and fills in missing values. It takes advertising creative performance data and market trend data as input, and obtains cleaned data as output. Specifically, the server identifies missing values ​​in the data and fills them in appropriately.

[1701] Step 4:

[1702] The server normalizes the data and converts it into a format suitable for analysis. It takes cleaned data as input and normalized data as output. Specifically, the server scales and encodes the data.

[1703] Step 5:

[1704] The server uses a data analysis tool to evaluate the performance of past advertising creatives. Normalized data is required as input, and performance evaluation results are obtained as output. Specifically, the server inputs data into the data analysis tool and obtains the evaluation results.

[1705] Step 6:

[1706] The server analyzes the market trend data to identify the current trend. The input required is normalized market trend data, and the output is trend analysis results. Specifically, the server uses data analysis tools to identify the trend.

[1707] Step 7:

[1708] The server inputs prompt statements into the generative AI model and generates the direction of the next ad creative and measures for market trends. The required inputs are performance evaluation results, trend analysis results, and prompt statements, and the output is the direction of the generated ad creative and measures. Specifically, the server inputs prompt statements into the generative AI model and obtains the generated results.

[1709] Step 8:

[1710] Using an application installed on a smart device, users implement the advertising creative direction and market trend countermeasures suggested by the generative AI model. The inputs are the generated advertising creative direction and countermeasures, and the output is the implementation results. Specifically, users operate the application to implement the suggested countermeasures.

[1711] Step 9:

[1712] The user then feeds back the results of the measures they have implemented to the generative AI model. The input is the implementation results, and the output is the fed-back data. Specifically, the user uses an application to input the implementation results and feeds them back to the generative AI model.

[1713] Example 3

[1714] Next, a third embodiment of the third embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1715] The PDCA cycle for conventional promotional measures had long lead times and made it difficult to take prompt action to improve. Furthermore, there was a lack of concrete methods for effectively utilizing the analysis results using generative AI, making it difficult to reduce man-hours and prioritize effective measures.

[1716] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes a means for the user to input a prompt sentence, a means for the terminal to send the prompt sentence to the server, a means for the server to preprocess the prompt sentence, a means for the server to input the prompt sentence to the generative AI model, a means for the generative AI model to predict the effectiveness of a countermeasure, a means for the server to receive the prediction result, a means for the server to analyze the prediction result and determine the priority of improvement actions, a means for the server to return the priority to the user, and a means for the user to execute the improvement action based on the priority. This enables shortening of lead time and realization of a rapid PDCA cycle.

[1717] "Generative AI" is an artificial intelligence technology that generates data based on user input and performs analysis and predictions.

[1718] "Promotional measures" are marketing activities aimed at increasing awareness of products and services and promoting sales.

[1719] The PDCA cycle is a method for achieving continuous improvement in business operations and projects by repeating the four steps of planning (Plan), executing (Do), evaluating (Check), and improving (Act).

[1720] "Lead time" is the time required from the start to the end of a process.

[1721] A "prompt" is text that contains instructions or questions that a user inputs to a generative AI.

[1722] A "server" is a computer system that provides services to other computers on a network.

[1723] A "terminal" is a device that a user uses to communicate with a server.

[1724] "Preprocessing" refers to the initial process of converting data into a format suitable for analysis or prediction.

[1725] A "generative AI model" is a specific implementation of generative AI, an algorithm that generates data and makes predictions based on user input.

[1726] "Prediction results" are the results of analysis and predictions that a generative AI model outputs based on user input.

[1727] "Improvement actions" are specific measures or policies implemented based on the prediction results.

[1728] "Priority" refers to the order of importance or urgency of execution among multiple improvement actions.

[1729] This invention relates to a system that uses generative AI to accelerate the PDCA cycle of promotional measures and shorten lead times. Specific embodiments of this system are described below.

[1730] First, the user uses a terminal to input a prompt. A prompt is text used to give instructions or ask questions to the generative AI. For example, a prompt such as "Please predict the effect of increasing sales if a new marketing strategy is implemented" can be input.

[1731] The terminal sends the prompt text entered by the user to the server. The HTTPS protocol is used for communication to ensure data security. The server then preprocesses the received prompt text. Specifically, it normalizes the text and removes unnecessary spaces.

[1732] Next, the server inputs the preprocessed prompt sentence into a generative AI model. The generative AI model runs on a high-performance server equipped with an NVIDIA GPU. For example, GPT-4 is used as the generative AI model. This generative AI model predicts the effectiveness of countermeasures based on the prompt sentence. For example, it predicts the effect of increasing sales based on past data and similar cases.

[1733] The server receives prediction results from the generative AI model. The prediction results are returned as numerical or text data. The server analyzes the prediction results and prioritizes improvement actions. For example, it ranks marketing strategies in order of their effectiveness in increasing sales.

[1734] The server returns the determined priorities to the user, who then displays them to the terminal. The user then takes action to improve the situation based on the priorities received from the server. For example, the user may first implement the marketing strategies predicted to be most effective.

[1735] This system significantly shortens lead times and enables a faster PDCA cycle. In addition, by effectively utilizing the analysis results using generative AI, it is possible to reduce man-hours and prioritize effective countermeasures.

[1736] As a concrete example, consider the case where a user inputs a prompt such as, "Please predict the sales increase effect of introducing a new marketing strategy." The device sends this prompt to the server, which then predicts the sales increase effect using a generative AI model. Based on the prediction results, the server determines the priority of the marketing strategies and returns it to the user. The user then implements the marketing strategy predicted to be most effective based on this priority.

[1737] In this way, the PDCA cycle of promotional measures utilizing generative AI can be accelerated and the lead time can be shortened. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1738] Step 1:

[1739] The user inputs a prompt sentence, such as "Please predict the sales increase effect of introducing a new marketing strategy" through the terminal interface. The input prompt sentence is temporarily stored in the terminal's memory.

[1740] Step 2:

[1741] The terminal sends the prompt text to the server. The terminal sends the prompt text entered by the user to the server using the HTTPS protocol. The input is the prompt text, and the output is a notification to the server that transmission has been completed.

[1742] Step 3:

[1743] The server preprocesses the prompt text. The server normalizes the received prompt text and removes unnecessary spaces and special characters. The input is the prompt text and the output is the preprocessed prompt text. Specific operations include text cleaning and tokenization.

[1744] Step 4:

[1745] The server inputs a prompt sentence to the generative AI model. The server then inputs a preprocessed prompt sentence to the generative AI model. The input is the preprocessed prompt sentence, and the output is a notification to the generative AI model that the input has been completed. Specific operations include API calls and sending data to the model.

[1746] Step 5:

[1747] The generative AI model predicts the effectiveness of countermeasures. The generative AI model predicts the effectiveness of countermeasures based on the prompt sentence. The input is the preprocessed prompt sentence, and the output is the predicted result. Specific operations include the model's inference process.

[1748] Step 6:

[1749] The server receives the prediction results. The server receives the prediction results from the generative AI model. The input is the prediction results, and the output is saving the prediction results to the server. Specific operations include saving the results to a database and analyzing the results.

[1750] Step 7:

[1751] The server analyzes the prediction results and determines the priorities of improvement actions. The server analyzes the prediction results and determines the priorities of improvement actions. The input is the prediction results and the output is a priority list. Specific operations include analysis using an algorithm and generation of rankings.

[1752] Step 8:

[1753] The server returns the priority to the user. The server returns the determined priority to the user. The input is a priority list, and the output is a notification to the user that transmission is complete. Specific operations include data format conversion and the transmission process.

[1754] Step 9:

[1755] The user executes improvement actions based on the priorities. The user executes improvement actions based on the priorities received from the server. The input is a priority list, and the output is the executed improvement actions. Specific operations include the introduction of marketing strategies and the execution of measures.

[1756] (Application example 3)

[1757] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1758] In conventional promotional measures and factory robot operations, there is a demand for accelerating the PDCA cycle and shortening lead times, but there is a lack of efficient ways to achieve this. Furthermore, there is no system in place that can collect factory robot operation data in real time, predict the effects of improvement measures using generative AI, and determine priorities, making it difficult to maximize production efficiency. A new system is needed to solve these issues.

[1759] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1760] In this invention, the server includes means for accelerating the PDCA cycle of promotional measures using generative AI, means for loading the generative AI with results such as acquisitions, means for analyzing the results in comparison with advertising creatives and market trends, means for proposing hypotheses and measures based on the analysis results, means for shortening the lead time to improvement actions, means for collecting operation data of robots in the factory in real time and predicting the effects of the proposed improvement measures using generative AI, and means for determining the priorities of improvement actions based on the prediction results and notifying the factory manager. This makes it possible to improve the efficiency of promotional measures and maximize the production efficiency of factory robots.

[1761] "Generative AI" refers to artificial intelligence that generates new information and predictions based on data.

[1762] The PDCA cycle is a method for improving business operations by repeating the four steps of Plan, Do, Check, and Act.

[1763] "Lead time" is the time it takes for a product or service to be delivered after it is ordered.

[1764] A "factory robot" is a mechanical device that performs work automatically in a factory.

[1765] "Operational data" refers to data that records the operating status of a machine or system.

[1766] "Improvement measures" refer to specific means or methods for solving current problems.

[1767] "Priority" refers to determining the order in which multiple items or tasks should be done first.

[1768] A "factory manager" is a position that involves the operation and management of a factory.

[1769] "Promotional measures" are specific activities or plans aimed at promoting products or services.

[1770] "Advertising creative" refers to the content, design, and expression of the advertisement.

[1771] A "market trend" is a change in consumer preferences and behavior in a particular market.

[1772] A "hypothesis" is a tentative explanation or prediction about a phenomenon or problem.

[1773] A "proposal" is a presentation of a solution or new idea to a problem.

[1774] The following system configuration will be described as an embodiment of the present invention.

[1775] The server uses generative AI to provide a means to accelerate the PDCA cycle of promotional measures. Specifically, the server loads performance data such as acquisition data into the generative AI, and analyzes it in comparison with advertising creatives and market trends. Based on the analysis results, the server proposes hypotheses and countermeasures, shortening the lead time until improvement actions are taken.

[1776] Furthermore, the server collects operational data from robots in the factory in real time and uses generative AI to predict the effectiveness of proposed improvement measures. Based on the prediction results, it determines the priority of improvement actions and provides a means to notify ...

Claims

1. A means for acquiring performance data of promotional measures; A means for analyzing the performance data using generative AI and proposing measures based on the analysis results; A means for suggesting an improvement action based on the countermeasure; A system including:

2. The system according to claim 1 , wherein the means for proposing countermeasures analyzes the performance data and proposes the countermeasures using a prompt sentence instructing the proposal of a countermeasure and the generative AI.

3. means for recognizing an emotion of a user using an emotion recognition engine; The system of claim 1 , wherein the means for suggesting remediation actions is further configured to suggest the remediation actions based on a recognized emotion.

Images