system

Abstract

We provide the system. [Solution] The present invention relates to a system comprising a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit. Based on customer attribute information and behavioral history, the system uses generative AI to automatically segment customers, automatically generates and distributes marketing messages and offers optimized for each segment or individual customer, and further collects and learns response data after distribution in real time to optimize the content and timing of future communications. This reduces the burden on staff while continuously and efficiently achieving optimal communication for each customer, enabling improved customer satisfaction, reduced churn rates, and the maintenance of long-term customer relationships.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a persona chatbot control method executed by at least one processor, the method comprising the steps of: receiving a user utterance; adding the user utterance to a prompt including an instruction sentence associated with a description relating to a chatbot character; encoding the prompt; and inputting the encoded prompt into a language model to generate a chatbot utterance that responds to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Traditionally, building long-term, positive relationships with customers and improving customer satisfaction requires meticulous communication tailored to each customer's attributes and behavioral patterns. However, manually analyzing vast amounts of customer data and creating and delivering optimal messages and offers for each customer is extremely inefficient and difficult to achieve due to human resource and time constraints. As a result, companies often resort to uniform information provision and one-sided campaign announcements, failing to adequately meet customer interests and needs, leading to problems such as decreased customer satisfaction, increased churn rates, and even customer churn to competitors. This invention aims to fundamentally solve the problems of conventional technology by automatically segmenting customers using generative AI based on customer behavior history and attribute information, automatically generating and delivering personalized messages and offers optimized for each segment and individual customer, and further collecting and learning customer response data in real time to optimize the content and timing of future communications. This significantly reduces the burden on staff while continuously and efficiently achieving optimal communication for each customer, enabling improved customer satisfaction, reduced churn rates, and the maintenance of long-term customer relationships. [Means for solving the problem]

[0005] The present invention comprises a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit. The customer data acquisition unit extracts data over a predetermined period from a customer database including customer attribute information, past usage history, and past response history. The customer segment classification unit uses a machine learning algorithm to classify customers into multiple segments based on the extracted data. The personalized message generation unit uses a generation AI to analyze past usage trends and response trends for each segment and automatically generates marketing messages and offers optimized for individual customers. The distribution management unit distributes the generated messages and offers. This system automatically selects the optimal channel and delivery timing for each customer from multiple channels such as email, social media, and app notifications, and delivers the message automatically. A response data collection unit collects response data such as whether the message was opened, clicked, or whether the offer was used in real time and stores it in a database. An optimization learning unit analyzes the collected response data and feeds it back as learning data for a generative AI, which is then used to optimize the content and delivery timing of future messages. By automatically executing this series of processes, the system reduces the burden on staff while continuously and efficiently achieving optimal communication for each customer, thereby improving customer satisfaction, reducing churn rates, and maintaining long-term customer relationships. [Brief explanation of the drawing]

[0006] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Modes for carrying out the invention]

[0007] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0008] First, let's explain the terminology used in the following explanation.

[0009] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), and APU (Accelerated Processing Unit).

[0010] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0011] In the following embodiments, the numbered storage is one or a plurality of non-volatile storage devices that store various programs, various parameters, and the like. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (for example, hard disks), magnetic tapes, and the like.

[0012] In the following embodiments, the numbered communication I / F (Interface) is an interface including a communication processor, an antenna, and the like. The communication I / F manages communication between a plurality of 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), and the like.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it may be only A, only B, or a combination of A and B. Further, in the present specification, when three or more items are connected and expressed by "and / or", the same concept as that for "A and / or B" is applied.

[0014] [First Embodiment]

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

[0016] As shown in FIG. 1, the data processing system 10 includes a data processing apparatus 12 and a smart device 14. An example of the data processing apparatus 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 the "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), and the like.

[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, the camera 42, and the communication I / F 44 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 receives user input from contact of an indicator by detecting the contact of the indicator (for example, a pen, a finger, or the like). The microphone 38B receives voice-based user input by detecting the user's voice. A control unit 46A in the processor 46 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, a 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

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

[0023] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0024] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0025] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output processing program 60 executed on the RAM 48. The smart device 14 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0026] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device having the data generation model 58. The data processing device 12 may also be a server device or a terminal device owned by a user (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0027] (Example 1)

[0028] The flow of a specific process in Example 1 will now be described. Each part of the system described below is implemented by a data processing device 12 and a smart device 14. The data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0029] (Embodiments for carrying out the present invention)

[0030] Embodiments of the present invention will be described in more specific and detail. The system is realized through the cooperation of a server and a terminal and comprises a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit. Each of these units is mainly implemented on the server, but some functions are also distributed on the terminal side.

[0031] First, the customer data acquisition unit operates on a server and extracts various data from the customer database. For example, it comprehensively acquires attribute information such as customer ID, name, age, gender, residential area, occupation, family structure, hobbies and preferences, birthday, and anniversary; past purchase history (purchase date and time, product category, product name, purchase amount, purchase channel); service usage history (usage date and time, type of service used, frequency of use, location of use); inquiry history (inquiry date and time, inquiry content, response history); and past campaign response history (email open date and time, link click date and time, SNS reaction type, app notification read date and time, coupon usage date and time, survey responses, event participation history). For example, it accumulates a chronological record of multifaceted customer behavior, such as what products a customer has purchased in the past, when and through which channel; how they responded to which campaigns; how many times they have made what kind of inquiries; and how they have answered various surveys. On the device side, the latest attribute information entered when using the app or website, as well as in-app activity logs (for example, product browsing history, cart addition history, wish list registration history, in-app search history, read / unread status of push notifications, in-app banner clicks, in-app coupon acquisition / use, in-app survey responses, in-app chat history, etc.) are sent to the server in real time and reflected in the server-side database sequentially.

[0032] The customer segmentation unit operates on a server and generates multidimensional features based on extracted customer data. For example, it extracts features such as monthly purchase frequency, average purchase amount, purchase ratio of specific categories, service usage frequency, inquiry frequency, campaign response rate, survey response rate, event participation rate, app usage frequency, and SNS reaction frequency. Using these features as input, clustering algorithms (e.g., k-means, hierarchical clustering, Gaussian mixture models, self-organizing maps, supervised learning classification models, etc.) are applied to classify customers into multiple segments such as "frequent high-spending customers," "low-frequency customers," "new customers," "inactive customers," "customers with a preference for specific categories," "event-participating customers," "coupon-using customers," "high-frequency inquiry customers," and "active survey responders." For example, it is possible to segment customers who have made purchases five or more times a month in the past six months, whose average purchase amount exceeds 10,000 yen, and whose purchase ratio of the health food category is 70% or more as "health food VIP customers." The segment classification results, along with each customer's attribute information and behavioral history, are recorded in a management database on the server and referenced in subsequent processing.

[0033] The personalized message generation unit operates on a server and utilizes a generative AI as an essential component. For each segment or individual customer, multidimensional data such as past usage trends, purchase history, inquiry content, campaign response history, survey responses, event participation history, and in-app behavior logs are provided as input data to the generative AI. The generative AI uses, for example, a large-scale language model to estimate customer interests and purchasing trends, and based on the estimation results, automatically generates optimal product suggestions, service information, coupons, campaign information, event invitations, survey requests, and support information in natural language for each customer. For example, a customer who has purchased many beauty-related products in the past and answered "interested in skin whitening" in a survey will be given a message such as "Announcement of a campaign to commemorate the launch of a new skin whitening product" or "Invitation to a skin whitening care experience event." A customer who has frequently purchased electronic devices in the past and often views new product information within the app will be given a message such as "Announcement of pre-order for a new smartphone" or "Accessory set discount offer." Furthermore, generative AI dynamically adjusts the style and expression of messages according to the customer's age group, gender, past response trends, and survey responses. For example, it uses casual language for younger generations, polite language for seniors, empathetic language for parents, and language that emphasizes efficiency and time-saving for business people, achieving fine-grained personalization. In addition, generative AI can use image generation models and banner generation models to automatically generate optimal visual materials for each customer (for example, product images that match the customer's interest categories, banners that match the season or event, and colors and designs that match the customer's attributes), and create messages that combine text and visuals. For example, it can automatically generate visuals that match the season or event, such as banners supporting new beginnings in the spring, birthday cake images for birthdays, and limited-edition gift images for Christmas.

[0034] The distribution management unit operates on a server and automatically delivers generated messages and offers from multiple channels, including email, social media, app notifications, web push notifications, SMS, and direct mail, selecting the optimal channel and delivery timing for each customer. For example, based on past delivery history, it selects email for customers with high email open rates, direct messages on social media for customers with many social media reactions, in-app notifications for customers who frequently use apps, SMS for customers where the immediacy of SMS is effective, and paper announcements for customers who respond well to direct mail. Regarding delivery timing, based on past response data, the generation AI predicts the time of day and event timing most likely to elicit a response from each customer, such as "weekday evenings," "Saturday mornings," "end of the month," "birthday," or "the day before an event," and delivers at the optimal time. On the device side, it implements features such as receiving app notifications and web push notifications, message viewing interfaces, coupon acquisition and usage interfaces, survey response interfaces, and event participation application interfaces.

[0035] The response data collection unit operates in conjunction with both the server and the terminal. On the server side, it collects data in real time, such as whether emails were opened, whether links were clicked, whether offers were used, reactions on social media, website access history, survey responses, and event participation status. On the terminal side, it detects user actions such as whether app notifications were read, whether in-app links were clicked, coupon usage, survey responses, chat history, and in-app event participation, and sends this data to the server. For example, if a customer uses a coupon within the app, that usage information is immediately sent to the server and recorded in the database. Similarly, survey responses and event participation applications are also sent from the terminal to the server and stored as individual customer behavior data.

[0036] The optimization learning unit operates on a server, analyzing collected response data and feeding it back as training data for the generative AI. For example, it statistically analyzes response rates to specific message expressions and offer content, effectiveness measurements for each delivery channel and timing, response trends by customer attributes and segment, correlation analysis between survey responses and response rates, and relationship analysis between event participation rates and campaign utilization rates. This data is then reflected in the generative AI's parameters, prompt design, segment classification algorithm weighting, delivery channel selection logic, and delivery timing prediction model. For instance, if messages emphasizing discount rates received a high response in the past, the system prioritizes generating discount-focused messages for similar customer segments. Conversely, if new product information received a low response, the system switches to other appeal axes, enabling dynamic optimization. Furthermore, it can personalize messages to reflect customer preferences, such as prioritizing future event announcements for customers who indicated interest in attending events in a survey.

[0037] Thus, in the embodiments of the present invention, a series of processes from customer data acquisition to segment classification, personalized message generation, delivery, response collection, and optimization learning are mainly performed automatically and integrally on the server, while the terminal side is responsible for customer interface and acquisition of real-time behavioral data. This significantly reduces the burden on staff while continuously and efficiently achieving optimal communication for each customer. Furthermore, by utilizing generative AI as an essential component and automatically generating messages and offers optimized for each customer, a high level of personalization beyond conventional uniform information provision is achieved, enabling improved customer satisfaction, reduced churn rates, and the maintenance of long-term customer relationships. Moreover, the system configuration can be flexibly adapted to various specific use cases, such as first-time limited offers for new customers, campaigns to encourage dormant customers to return, related product suggestions for customers with specific category preferences, special message delivery tailored to events such as birthdays and anniversaries, personalized support guidance based on survey responses, VIP invitations based on event participation history, reminder notifications based on in-app behavior logs, and SNS-only campaign announcements based on SNS reaction trends.

[0038] (System Configuration)

[0039] The system according to this embodiment includes a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit. The customer data acquisition unit comprehensively acquires a multifaceted history of customer behavior in chronological order, including attribute information such as customer ID, name, age, gender, residential area, occupation, family structure, hobbies and preferences, birthday, and anniversary; past purchase history such as purchase date and time, product category, product name, purchase amount, and purchase channel; service usage history such as usage date and time, service type used, frequency of use, and location of use; inquiry history such as inquiry date and time, inquiry content, and response history; and past campaign response history such as email open date and time, link click date and time, SNS reaction type, app notification read date and time, coupon usage date and time, survey response content, and event participation history. For example, the acquisition targets also include the history of when customers viewed products in the app, when they added products to their cart, when they registered products to their wish list, the keyword history searched in the app, read / unread information for push notifications, in-app banner click history, in-app coupon acquisition / usage history, in-app survey response history, and in-app chat inquiry history. Furthermore, the system collects data from every customer touchpoint, including website access logs, page viewing history, form submission history, event registration history, and response history to direct mail, and stores it in a database on the server.

[0040] The Customer Segmentation Classification Unit extracts features such as monthly purchase frequency, average purchase amount, specific category purchase ratio, service usage frequency, inquiry frequency, campaign response rate, survey response rate, event participation rate, app usage frequency, and SNS reaction frequency based on multidimensional data obtained from the Customer Data Acquisition Unit, and classifies customers into multiple segments using clustering algorithms and supervised learning models. For example, customers who have made purchases five or more times a month in the past six months, have an average purchase amount exceeding 10,000 yen, and whose purchase ratio in the health food category is 70% or more can be classified as "Health Food VIP Customers," while customers with no purchase history in the past three months and a low campaign email open rate can be classified as "Dormant Customers." It is also possible to classify customers with a high event participation history and high survey response rate as "Event Participant Customers," and customers with a high coupon usage frequency as "Coupon-Using Customers." Furthermore, segmentation can be performed using a variety of axes, such as classifying customers with many SNS reactions as "SNS Active Customers" and customers with a high inquiry frequency as "Support-Focused Customers."

[0041] The personalized message generation unit receives multidimensional data such as past usage trends, purchase history, inquiry content, campaign response history, survey responses, event participation history, and in-app behavior logs as input data for the generative AI, for each segment classified by the customer segment classification unit, or for each individual customer. Using the generative AI, it estimates the customer's interests and purchasing trends, and based on the estimation results, automatically generates optimal product suggestions, service information, coupons, campaign information, event invitations, survey requests, and support information in natural language for each customer. For example, a customer who has purchased many beauty-related products in the past and answered "interested in skin whitening" in a survey will be given a message such as "Announcement of a campaign to commemorate the launch of a new skin whitening product" or "Invitation to a skin whitening care experience event." A customer who frequently purchases electronic devices and often views new product information within the app will be given a message such as "Announcement of advance reservation for a new smartphone" or "Accessory set discount offer." Furthermore, the generative AI dynamically adjusts the style and expression of messages according to the customer's age group, gender, past response trends, and survey responses, achieving fine-grained personalization. For example, it uses casual language for younger generations, polite language for seniors, empathetic language for parents, and language that emphasizes efficiency and time-saving for business people. In addition, the generative AI uses image generation models and banner generation models to automatically generate the most suitable visual materials for each customer. For example, it can automatically generate banners supporting new beginnings in the spring, birthday cake images for birthdays, limited gift images for Christmas, and event logos and participation certificate images for events, creating messages that combine text and visuals according to the season and event. Specific examples of prompt texts to be fed into a generative AI include: "A woman in her 30s, who has purchased health foods 10 times in the past year, with an average purchase amount of 8,000 yen, and who is interested in dieting according to a survey. Generate an announcement for a new diet product launch campaign in a friendly tone." "A man in his 40s, with a history of numerous purchases in the electronics category, who frequently views new product information within apps. Generate a concise announcement for pre-ordering a new smartphone, incorporating specialized content." "A student in his 20s, with a history of numerous event participations, and who is active on social media."Examples of assignments include: "Generate an invitation message for the next event using casual language suitable for social media," and "Generate a follow-up invitation to a health seminar for a woman in her 60s who indicated a health-conscious attitude in a survey and has previously attended health seminars, using polite and respectful language."

[0042] The distribution management department automatically delivers messages and offers generated by the personalized message generation department to each customer via multiple channels, including email, social media, app notifications, web push notifications, SMS, and direct mail, selecting the most suitable channel and delivery timing for each customer. For example, based on past delivery history, email is sent to customers with high email open rates, direct messages on social media are sent to customers with high social media engagement, in-app notifications are sent to customers with high app usage frequency, SMS is used for customers where the immediacy of SMS is effective, and paper-based announcements are sent to customers who respond well to direct mail. Regarding delivery timing, based on past response data, the generation AI predicts the time of day and event timings when each customer is most likely to respond, such as weekday evenings, Saturday mornings, the end of the month, their birthday, or the day before an event, and delivers at the optimal time. For example, the system automatically determines the optimal delivery timing for each channel, such as app notifications during commuting hours or lunch breaks, emails on weekend mornings, and social media messages just before an event. Furthermore, it can implement cross-channel delivery strategies by combining multiple channels, for example, sending an email first and then sending a reminder via social media if it remains unopened.

[0043] The response data collection unit collects customer response data after delivery by coordinating on both the server and the terminal. On the server side, it collects in real time whether emails were opened, links were clicked, offers were used, reactions on social media, website access history, survey responses, and event participation status. On the terminal side, it detects user actions such as whether app notifications were read, in-app links were clicked, coupons were used, surveys were answered, chat history was recorded, and in-app event participation was performed, and sends this information to the server. For example, if a customer uses a coupon within the app, the usage information is immediately sent to the server and recorded in the database. Survey responses and event participation applications are also sent from the terminal to the server and stored as individual customer behavior. Furthermore, it comprehensively collects a variety of response data, including banner clicks and form submissions on websites, comments and shares on social media, and completion of video viewing within the app.

[0044] The optimization learning unit analyzes collected response data and feeds it back as training data for the generative AI. For example, it statistically analyzes response rates to specific message expressions and offer content, effectiveness measurements for each delivery channel and timing, response trends by customer attributes and segments, correlation analysis between survey responses and response rates, and relationship analysis between event participation rates and campaign utilization rates. This data is then reflected in the generative AI's parameters, prompt design, segment classification algorithm weighting, delivery channel selection logic, and delivery timing prediction model. For instance, if messages emphasizing discount rates received a high response in the past, the system prioritizes generating discount-focused messages for similar customer segments. Conversely, if new product information received little response, the system switches to other appeal axes, enabling dynamic optimization. Furthermore, it can personalize messages that reflect customer preferences, such as prioritizing the delivery of future event announcements to customers who indicated interest in participating in events in surveys. In addition, it can automatically conduct A / B tests based on response data, continuously learning and applying the most effective messages and delivery strategies.

[0045] Thus, the system according to this embodiment organically links a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit, and utilizes a generative AI as an essential component to automatically and continuously realize optimized communication for each customer, enabling improved customer satisfaction, reduced churn rates, and the maintenance of long-term customer relationships.

[0046] (Implementation Steps)

[0047] Step 1: Comprehensive collection of customer data The customer data acquisition unit operates on the server and comprehensively acquires a multifaceted history of customer behavior in chronological order, including attribute information such as customer ID, name, age, gender, residential area, occupation, family structure, hobbies and preferences, birthday, and anniversaries, as well as past purchase history (purchase date and time, product category, product name, purchase amount, purchase channel), service usage history (usage date and time, type of service used, frequency of use, location of use), inquiry history (inquiry date and time, inquiry content, response history), and past campaign response history (email open date and time, link click date and time, SNS reaction type, app notification read date and time, coupon usage date and time, survey responses, event participation history). Furthermore, the latest attribute information entered when using the app or website, as well as in-app behavior logs (product browsing history, cart addition history, wish list registration history, in-app search history, push notification read / unread status, in-app banner clicks, in-app coupon acquisition / use, in-app survey responses, in-app chat history, etc.) are sent to the server in real time and reflected sequentially in the server-side database.

[0048] Step 2: Multidimensional classification of customer segments The customer segmentation unit operates on the server, extracting features from collected customer data such as monthly purchase frequency, average purchase amount, purchase ratio of specific categories, service usage frequency, inquiry frequency, campaign response rate, survey response rate, event participation rate, app usage frequency, and SNS reaction frequency. It then classifies customers into multiple segments using clustering algorithms and supervised learning models. For example, it segments customers into categories such as VIP customers for health foods, inactive customers, event-participating customers, coupon-using customers, SNS-active customers, and support-focused customers, and records the segment classification results in a management database.

[0049] Step 3: Generating personalized messages using AI The personalized message generation unit operates on the server and provides multidimensional data such as past usage trends, purchase history, inquiry content, campaign response history, survey responses, event participation history, and in-app behavior logs as input data for the generating AI, for each segment or for each individual customer. The generating AI is used to estimate the customer's interests and purchasing trends, and based on the estimation results, it automatically generates optimal product suggestions, service information, coupons, campaign information, event invitations, survey requests, support information, etc., in natural language for each customer. Specific examples of prompt texts to be fed into the generating AI include: "A woman in her 30s, who has purchased health foods 10 times in the past year, with an average purchase amount of 8,000 yen, and who expressed interest in dieting in a survey. Generate a friendly tone for announcing a new diet product launch campaign." "A man in his 40s, with a history of numerous purchases in the electronics category, who frequently views new product information within apps. Generate a concise yet technical announcement for pre-ordering a new smartphone." "A student in his 20s, with a history of numerous event participations and active engagement on social media. Generate a casual invitation message for the next event, suitable for social media." "A woman in her 60s, who indicated a health-conscious attitude in a survey and has previously attended health seminars. Generate a polite and respectful message to remind her about a health seminar." Furthermore, the generating AI also utilizes image generation and banner generation models to automatically generate visuals appropriate to the season and event, such as a banner supporting new beginnings in the spring, a birthday cake image for birthdays, and a limited-edition gift image for Christmas, creating messages that combine text and visuals.

[0050] Step 4: Automated delivery on the optimal channel and timing. The distribution management unit operates on a server and automatically delivers generated messages and offers from multiple channels, including email, social media, app notifications, web push notifications, SMS, and direct mail, selecting the optimal channel and delivery timing for each customer. For example, it selects email for customers with high email open rates, direct messages on social media for customers with high social media engagement, in-app notifications for customers who frequently use apps, SMS for customers where immediacy is effective, and paper mail for customers who respond well to direct mail. Regarding delivery timing, based on past response data, the AI ​​generates messages to predict the time of day and event timings when each customer is most likely to respond, such as weekday evenings, Saturday mornings, the end of the month, their birthday, or the day before an event, and delivers at the optimal time.

[0051] Step 5: Real-time collection of diverse response data The response data collection unit works in conjunction with both the server and the terminal to comprehensively collect diverse response data in real time, including whether emails were opened, whether links were clicked, whether offers were used, reactions on social media, website access history, survey responses, event registration status, app notification read status, in-app link clicks, coupon usage, survey responses, chat history, in-app event participation, banner clicks and form submissions on websites, comments and shares on social media, and completion of in-app video viewing. This data is then recorded in the server's database.

[0052] Step 6: Optimization learning and continuous improvement using generative AI The optimization learning unit operates on the server, analyzing collected response data and feeding it back as training data for the generating AI. For example, it statistically analyzes response rates to specific message expressions and offer content, effectiveness measurements for each delivery channel and timing, response trends by customer attributes and segment, correlation analysis between survey responses and response rates, and relationship analysis between event participation rates and campaign utilization rates. This data is then reflected in the generating AI's parameters, prompt design, segment classification algorithm weighting, delivery channel selection logic, and delivery timing prediction model. Specific examples of prompts to be fed into the generating AI include: "For customer segments that have shown a high response to messages emphasizing discount rates over the past three months, prioritize generating discount-focused messages in the future," "For customer segments that showed little response to new product information, generate messages emphasizing other appeals (e.g., exclusivity, testimonials, rankings)," and "For customers who indicated interest in participating in events in the survey, generate messages prioritizing future event announcements." Furthermore, A / B test results and changes in response rates are also incorporated as training data to continuously learn and apply the most effective messages and delivery strategies.

[0053] (Specific use cases)

[0054] For example, let's consider a case where a business primarily dealing in health foods introduces the system of the present invention and conducts marketing activities aimed at promoting continued use by existing customers and improving the retention rate of new customers. The customer data acquisition unit comprehensively stores multifaceted behavioral and attribute data of customers in a database on the server, including customer ID, age, gender, residential area, past health food purchase history (purchase date and time, product category, purchase amount, purchase channel), app usage history (number of times products are viewed in the app, cart addition history, coupon acquisition and usage history), survey responses (health orientation, interest in dieting, presence or absence of allergies), event participation history (attendance at health seminars, online consultation usage history), and campaign response history via email and social media (open date and time, click date and time, type of social media reaction). The customer segment classification unit classifies, for example, customers who have purchased health foods three or more times a month in the past six months, whose average purchase amount exceeds 8,000 yen, and who answered "interested in dieting" in a survey as "high-frequency customers with a dieting orientation," and customers who have no purchase history in the past three months and have a low open rate of campaign emails as "dormant customers." Furthermore, customers who have a history of attending health seminars and who answered "actively engaged in maintaining their health" in a survey are classified as "actively engaged in maintaining their health," and customers who frequently view new product information within the app are classified as "customers interested in new products."

[0055] The personalized message generation unit inputs multidimensional data such as past purchase history, in-app behavior logs, survey responses, event participation history, and campaign response history into the generation AI for each segment or for each individual customer. For example, it generates messages such as "announcement of a campaign to commemorate the launch of a new diet product" or "an invitation to a diet support event" for "high-frequency customers with a diet focus," and "announcement of a coupon to encourage return" or "announcement of a repurchase benefit campaign" for "dormant customers." It also generates messages such as "announcement of a new lineup of health maintenance support products" or "announcement of re-participation in a health seminar" for "customers interested in new products," and "announcement of advance reservation for a new product" or "announcement of a new product experience event" for "customers interested in new products." The generation AI uses casual language for younger customers and polite language for senior customers, depending on the customer's age group and past response trends. Furthermore, it utilizes an image generation model to automatically generate visual materials tailored to customer attributes and interests, such as image files for new diet products or event banners for health seminars, and combines them with text to construct the delivered message.

[0056] The distribution management department automatically selects the optimal channel and delivery timing for each customer from multiple channels, including email, social media, app notifications, web push notifications, and SMS. For example, new product announcements are sent via email to "health-conscious customers" who have a high email open rate, invitations to experience events are sent via direct message on social media to "customers interested in new products" who have many reactions on social media, and coupon information is sent via in-app notifications to "high-frequency customers with a diet focus" who frequently use the app. The distribution timing is also determined by generating AI, which predicts the time of day when each customer is most likely to respond, such as weekday evenings, weekend mornings, or the day before an event, based on past response data, and delivers at the optimal time.

[0057] The response data collection unit collects diverse response data in real time, such as whether emails were opened after delivery, whether links were clicked, whether offers were used, SNS reactions, app notification read status, in-app coupon usage, survey responses, event registrations, and website banner clicks, and records it in the server's database. The optimization learning unit analyzes the collected response data, statistically analyzing things like click-through rates and coupon usage rates for "new diet product announcements," event participation rates for "health seminar reminder notices," and repurchase rates for "recovery promotion coupon announcements," and reflects this analysis in the generation AI parameters, prompt design, segment classification algorithm, delivery channel selection logic, and delivery timing prediction model.

[0058] Specific examples of prompt texts to be fed into the generating AI include: "A woman in her 30s, who has purchased health foods 10 times in the past year, with an average purchase amount of 8,000 yen, and who expressed interest in dieting in a survey. Generate a friendly tone for announcing a new diet product launch campaign." "A man in his 40s, with a history of numerous purchases in the electronics category, who frequently views new product information within apps. Generate a concise yet technical announcement for pre-ordering a new smartphone." "An inactive customer with no purchase history in the past 3 months and a low campaign email open rate. Generate a return-promotion coupon announcement that encourages repeat use." "A customer actively engaged in health maintenance, with a history of attending health seminars, who answered in a survey that they are actively engaged in health maintenance. Generate a polite and respectful message for a health seminar re-announcement."

[0059] Thus, the system of the present invention automatically generates personalized messages and offers based on diverse customer behavior and attribute data, utilizing generative AI, delivers them through the optimal channel and timing, and continuously optimizes by collecting and learning response data in real time, thereby improving customer satisfaction, reducing churn rates, and maintaining long-term customer relationships.

[0060] (Application Example 1) The flow of a specific process in Application Example 1 will be explained below. Each part of the system described below is implemented by a data processing device 12 and a smart device 14. The data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0061] (Embodiments for carrying out the present invention)

[0062] Embodiments of the present invention will be described in more specific and detail. The present invention is a system that automatically generates and delivers product suggestions and marketing communications optimized for each customer on an e-commerce site, and learns and optimizes their effectiveness in real time, and comprises a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a delivery management unit, a response data collection unit, and an optimization learning unit.

[0063] The customer data acquisition unit operates on the server and collects attribute information such as customer ID, name, age, gender, residential area, membership rank, point balance, family structure, hobbies and preferences, birthday, anniversary, past purchase history (purchase date and time, product category, product name, purchase amount, purchase channel, purchase frequency, purchase cycle), browsing history (product page viewing date and time, number of views by category, viewing time, browsing device), cart addition history (product name, addition date and time, non-purchase period, number of times cart was left untouched), favorite registration history (product name, registration date and time, unregistration date and time), review posting history (posting date and time, evaluation content, review content, number of likes on the review), campaign response history (email open date and time, link click date and time, coupon acquisition / use date and time, The system comprehensively and chronologically stores a vast amount of multifaceted and extensive customer behavior and attribute data in a database on the server. This includes campaign participation history, inquiry history (date and time of inquiry, content, response status, inquiry channel), survey responses (interests, purchase intent, improvement requests, satisfaction ratings), app usage history (login frequency, notification read status, in-app search history, in-app banner click history, in-app coupon acquisition and usage history, in-app survey responses, in-app chat history), website usage history (date and time of access, page transition history, form input history, event participation application history), and SNS integration history (SNS account integration status, SNS share history, SNS reaction history). For example, it records in detail various purchasing behaviors and interest trends, such as customers who frequently browse fashion category products and add them to their cart multiple times but do not make a purchase, customers who have previously purchased high-priced items in the home appliance category and wish to receive restock notifications, customers who actively post reviews and receive many likes from other users, customers who frequently use coupons and have a high campaign participation rate, customers who frequently search for new product information within the app and frequently click banners, and customers who frequently share products on social media.

[0064] Based on this data, the Customer Segmentation Department extracts multidimensional features such as monthly purchase frequency, average purchase amount, purchase ratio of specific categories, cart abandonment rate, cart abandonment period, review posting frequency, average review rating, coupon usage rate, app usage frequency, campaign response rate, banner click-through rate, survey response rate, SNS sharing frequency, inquiry frequency, and event participation rate. Using clustering algorithms (k-means, hierarchical clustering, Gaussian mixture models, self-organizing maps, etc.) and supervised learning models (decision trees, random forests, support vector machines, neural networks, etc.), they classify customers along multiple axes into categories such as "high-frequency purchasing customers," "new customers," "dormant customers," "customers with a preference for specific categories," "customers who abandon their carts," "customers who actively post reviews," "customers who are inclined to use coupons," "app active customers," "SNS active customers," "event-participating customers," and "high-frequency inquiry customers." For example, customers who have made 10 or more purchases in the past three months and whose average purchase amount exceeds 10,000 yen are classified as "high-frequency, high-value customers," while customers who have no purchase history in the past year and have a low email open rate are classified as "dormant customers." Customers whose browsing and purchasing ratio in the fashion category exceeds 80% are classified as "fashion-oriented customers," and customers with multiple items in their cart that have not been purchased for more than 7 days are classified as "cart-abandoned customers." Customers who post 5 or more reviews per month and have an average rating of 4.5 or higher are classified as "active review-posting customers," and customers whose coupon usage rate exceeds 80% are classified as "coupon-using customers." Customers who use the app more than 10 times per week are classified as "app-active customers," and customers who share products on social media 5 or more times a month are classified as "SNS-active customers." Customers who have participated in 3 or more events are classified as "event-participating customers," and customers who make inquiries 3 or more times a month are classified as "high-frequency inquiry customers," and so on, segmenting is carried out using a variety of criteria.

[0065] The personalized message generation unit inputs past behavioral history and attribute information for each segment or individual customer into the generation AI, and automatically generates optimal product suggestions, campaign announcements, coupon distribution messages, reminder notifications, review requests, restock announcements, seasonal event announcements, banner ads, in-app pop-up messages, social media posts, survey requests, event invitations, support announcements, and more in natural language. For example, for "fashion-conscious customers," it generates "recommended new fashion items," "seasonal limited-time coordination feature announcements," and "fashion coordination diagnosis announcements." For "customers who abandon their carts," it generates "reminder notifications about low stock of items in the cart," "coupon announcements for cart items only," and "double points campaign announcements for cart purchases." For "high-frequency, high-spending customers," it generates "VIP-only pre-sale announcements," "point-up campaign announcements," and "limited-time gift announcements." For "customers who actively post reviews," it generates "review-posting thank-you coupon announcements," "new product review requests," and "review ranking participation announcements." For "coupon-oriented customers," it generates "exclusive coupon distribution announcements" and "free shipping campaign announcements with coupon use." For "app-active customers," it generates "app-exclusive new feature announcements" and "in-app event invitations." For "SNS-active customers," it generates "point reward announcements for SNS sharing" and "SNS-exclusive campaign announcements." For "event-participating customers," it generates "next event invitations" and "event participation benefit announcements." For "high-frequency inquiry customers," it generates "enhanced support announcements" and "dedicated chat support announcements." The generation AI achieves detailed personalization based on customer age group, past response trends, survey responses, purchase cycle, usage channels, and past message response rates, using casual language for younger generations, polite language for seniors, language emphasizing efficiency and time-saving for business people, and language that evokes empathy for parents.Furthermore, by utilizing image generation models and banner generation models, the system can automatically generate images such as new product images, seasonal event banner images, limited-edition gift packaging images, event participation certificate images, and illustrations for in-app pop-ups, and combine them with text to create delivery messages and advertising materials. For example, it can automatically generate visuals appropriate to the season and event, such as a banner to support new beginnings in the spring, a birthday cake image for birthdays, a limited-edition gift image for Christmas, and event logos and participation certificate images when events are held.

[0066] The distribution management department automatically selects the optimal channel and delivery timing for each customer from multiple channels, including email, app notifications, web push notifications, SMS, SNS direct messages, and postal direct mail. For example, it selects in-app notifications for customers who frequently use the app, email for customers with high email open rates, SMS for timely reminder notifications, SNS direct messages for customers who frequently interact on SNS, and paper announcements for customers who respond well to postal direct mail. Based on past response data, the generation AI predicts the time of day and event timing when each customer is most likely to respond, such as weekday evenings, weekend mornings, just before a sale starts, within 24 hours of adding items to the cart, on their birthday, or the day before an event, and delivers at the optimal time. Furthermore, it can implement cross-channel distribution strategies by combining multiple channels, for example, by sending an email first and then sending a reminder via app notification or SMS if it is not opened, or by distributing campaign information on SNS and then re-notifying within the app if there is no response.

[0067] The response data collection unit collects response data in real time, including whether an email was opened after delivery, whether it was clicked, whether a coupon was used, whether a product was purchased, whether a review was posted, whether an app notification was read, website access history, banner click history, whether a survey was answered, whether an event was registered, whether social media shares were made, and whether chat support was used. This data is then recorded in the server's database. For example, if a customer uses a coupon within the app, that usage information is immediately sent to the server and recorded in the database. Survey responses and event registration details are also sent from the device to the server and stored as individual customer behavior. A wide variety of response data is comprehensively collected, including banner clicks and form submissions on websites, comments and shares on social media, and completion of video viewing within the app.

[0068] The optimization learning unit analyzes collected response data and feeds it back as training data for the generative AI. For example, it statistically analyzes response rates to specific message expressions and offer content, effectiveness measurements for each delivery channel and timing, response trends by customer attributes and segments, correlation analysis between survey responses and response rates, relationship analysis between event participation rates and campaign utilization rates, review submission rates for review requests, purchase rates of items in the cart for cart abandonment reminder notifications, and click-through rates for banner ads. This data is then reflected in the generative AI's parameters, prompt design, segment classification algorithm weighting, delivery channel selection logic, and delivery timing prediction model. For instance, if messages emphasizing discount rates received a high response in the past, the system prioritizes generating discount-focused messages for similar customer segments. Conversely, if new product information received a low response, it switches to other appeal axes (exclusivity, testimonials, rankings, review excerpts, etc.), enabling dynamic optimization. Personalization reflecting customer preferences is also achieved, such as prioritizing the delivery of future event announcements to customers who indicated interest in participating in events in surveys. Based on response data, it's possible to automatically conduct A / B tests and continuously learn and apply the most effective messaging and delivery strategies.

[0069] Thus, in the embodiments of the present invention, a series of processes from customer data acquisition to segment classification, personalized message generation, delivery, response collection, and optimization learning are mainly performed automatically and integrally on a server, while the terminal side is responsible for customer interface and acquisition of real-time behavioral data. This significantly reduces the burden on staff while enabling continuous and efficient optimal communication for each customer. Furthermore, by utilizing generative AI as an essential component and automatically generating messages and offers optimized for each customer, a high level of personalization that goes beyond conventional uniform information provision is achieved, enabling improved customer satisfaction, reduced churn rates, and the maintenance of long-term customer relationships. Furthermore, the system can be configured to flexibly accommodate a variety of specific use cases, such as first-time customer offers, campaigns to encourage dormant customers to return, product recommendations for customers with specific category preferences, special message delivery tailored to events like birthdays and anniversaries, personalized support guidance based on survey responses, VIP invitations based on event participation history, reminder notifications based on in-app activity logs, SNS-exclusive campaign announcements based on SNS reaction trends, requests for active review-posting customers to participate in review rankings, inventory warning notifications for customers who have abandoned their carts, exclusive coupon distribution to coupon-using customers, and app-exclusive event announcements to active app users.

[0070] (System Configuration)

[0071] The system according to this embodiment includes a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit. The customer data acquisition unit includes attribute information such as customer ID, name, age, gender, residential area, membership rank, point balance, family structure, hobbies and preferences, birthday, and anniversary; past purchase history including purchase date and time, product category, product name, purchase amount, purchase channel, purchase frequency, and purchase cycle; browsing history including product page viewing date and time, number of views by category, viewing time, and browsing device; cart addition history including product name, addition date and time, non-purchase period, and number of times the cart was left untouched; favorite registration history including product name, registration date and time, and unregistration date and time; review posting history including posting date and time, evaluation content, review content, and number of likes on the review; and campaign response history including email opening date and time, link click date and time, coupon acquisition and usage date and time, and campaign The system comprehensively and chronologically stores a vast amount of multifaceted customer behavior and attribute data in a database on the server. This includes campaign participation history, inquiry history (inquiry date and time, content, response status, inquiry channel), survey responses (interests, purchase intent, improvement requests, satisfaction ratings), app usage history (login frequency, notification read status, in-app search history, in-app banner click history, in-app coupon acquisition / usage history, in-app survey responses, in-app chat history), website usage history (access date and time, page transition history, form input history, event participation application history), and SNS integration history (SNS account integration status, SNS share history, SNS reaction history). For example, it records various purchasing behaviors and interest trends in detail, such as customers who frequently browse fashion category products and add them to their cart multiple times but do not make a purchase, customers who have previously purchased high-priced items in the home appliance category and wish to receive restock notifications, customers who actively post reviews and receive many likes from other users, customers who frequently use coupons and have a high campaign participation rate, customers who frequently search for new product information within the app and frequently click banners, and customers who frequently share products on social media.

[0072] The Customer Segment Classification Unit extracts multidimensional features such as monthly purchase frequency, average purchase amount, specific category purchase ratio, cart abandonment rate, cart abandonment period, review posting frequency, average review rating, coupon usage rate, app usage frequency, campaign response rate, banner click-through rate, survey response rate, SNS sharing frequency, inquiry frequency, and event participation rate, and uses clustering algorithms and supervised learning models to classify them along multiple axes into categories such as "high-frequency purchasing customers," "new customers," "dormant customers," "customers with a preference for specific categories," "customers who abandon their carts," "customers who actively post reviews," "customers who are inclined to use coupons," "app active customers," "SNS active customers," "event-participating customers," and "high-frequency inquiry customers." For example, customers who have made 10 or more purchases in the past three months and whose average purchase amount exceeds 10,000 yen are classified as "high-frequency, high-value customers," while customers who have no purchase history in the past year and have a low email open rate are classified as "dormant customers." Customers whose browsing and purchasing ratio in the fashion category exceeds 80% are classified as "fashion-oriented customers," and customers with multiple items in their cart that have not been purchased for more than 7 days are classified as "cart-abandoned customers." Customers who post 5 or more reviews per month and have an average rating of 4.5 or higher are classified as "active review-posting customers," and customers whose coupon usage rate exceeds 80% are classified as "coupon-using customers." Customers who use the app more than 10 times per week are classified as "app-active customers," and customers who share products on social media 5 or more times a month are classified as "SNS-active customers." Customers who have participated in 3 or more events are classified as "event-participating customers," and customers who make inquiries 3 or more times a month are classified as "high-frequency inquiry customers," and so on, segmenting is carried out using a variety of criteria.

[0073] The personalized message generation unit inputs past behavioral history and attribute information for each segment or individual customer into the generation AI, and automatically generates optimal product suggestions, campaign announcements, coupon distribution messages, reminder notifications, review requests, restock announcements, seasonal event announcements, banner ads, in-app pop-up messages, social media posts, survey requests, event invitations, support announcements, and more in natural language. For example, for "fashion-conscious customers," it generates "recommended new fashion items," "seasonal limited-time coordination feature announcements," and "fashion coordination diagnosis announcements." For "customers who abandon their carts," it generates "reminder notifications about low stock of items in the cart," "coupon announcements for cart items only," and "double points campaign announcements for cart purchases." For "high-frequency, high-spending customers," it generates "VIP-only pre-sale announcements," "point-up campaign announcements," and "limited-time gift announcements." For "customers who actively post reviews," it generates "review-posting thank-you coupon announcements," "new product review requests," and "review ranking participation announcements." For "coupon-oriented customers," it generates "exclusive coupon distribution announcements" and "free shipping campaign announcements with coupon use." For "app-active customers," it generates "app-exclusive new feature announcements" and "in-app event invitations." For "SNS-active customers," it generates "point reward announcements for SNS sharing" and "SNS-exclusive campaign announcements." For "event-participating customers," it generates "next event invitations" and "event participation benefit announcements." For "high-frequency inquiry customers," it generates "enhanced support announcements" and "dedicated chat support announcements." The generating AI achieves detailed personalization based on the customer's age group, past response trends, survey responses, purchase cycle, usage channels, and past message response rates, using casual language for younger generations, polite language for seniors, language emphasizing efficiency and time-saving for business people, and language that evokes empathy for parents.Furthermore, by utilizing image generation models and banner generation models, the system automatically generates images for new products, banners for seasonal events, packaging for limited-edition gifts, event participation certificates, and illustrations for in-app pop-ups, and combines these with text to create delivery messages and advertising materials. Specific examples of prompt text to be fed into the generation AI include: "Female in her 20s, has purchased items in the fashion category 8 times in the past 6 months, and has items left in her cart. Generate a friendly tone for recommending new fashion items," "Male in his 40s, has a history of numerous purchases in the home appliance category, and has items awaiting restock. Generate a restock notification text incorporating expert content," "Inactive customer, no purchase history in the past year, coupon-oriented. Generate a coupon notification text to encourage re-entry and reuse," and "Active review-posting customer, has posted 5 reviews in the past 3 months. Generate a new product review request text expressing gratitude."

[0074] The distribution management department automatically selects the optimal channel and delivery timing for each customer from multiple channels, including email, app notifications, web push notifications, SMS, SNS direct messages, and postal direct mail. For example, it selects in-app notifications for customers who frequently use the app, email for customers with high email open rates, SMS for timely reminder notifications, SNS direct messages for customers who frequently interact on SNS, and paper announcements for customers who respond well to postal direct mail. Based on past response data, the AI ​​generates predictions the time of day and event timings when customers are most likely to respond, such as weekday evenings, weekend mornings, just before a sale starts, within 24 hours of adding items to the cart, on their birthday, or the day before an event, and delivers at the optimal time. Furthermore, it can implement cross-channel distribution strategies by combining multiple channels, for example, by sending an email first and then sending a reminder via app notification or SMS if it is not opened, or by distributing campaign information on SNS and then re-notifying within the app if there is no response.

[0075] The response data collection unit collects response data in real time, including whether an email was opened after delivery, whether it was clicked, whether a coupon was used, whether a product was purchased, whether a review was posted, whether an app notification was read, website access history, banner click history, whether a survey was answered, whether an event was registered, whether social media shares were made, and whether chat support was used. This data is then recorded in the server's database. For example, if a customer uses a coupon within the app, that usage information is immediately sent to the server and recorded in the database. Survey responses and event registration details are also sent from the device to the server and stored as individual customer behavior. A wide variety of response data is comprehensively collected, including banner clicks and form submissions on websites, comments and shares on social media, and completion of video viewing within the app.

[0076] The optimization learning unit analyzes collected response data and feeds it back as training data for the generation AI. For example, it statistically analyzes response rates to specific message expressions and offer content, effectiveness measurements for each delivery channel and timing, response trends by customer attributes and segments, correlation analysis between survey responses and response rates, relationship analysis between event participation rates and campaign utilization rates, review submission rates for review requests, purchase rates of items in the cart for cart abandonment reminder notifications, and click-through rates for banner ads. This data is then reflected in the generation AI's parameters, prompt design, segment classification algorithm weighting, delivery channel selection logic, and delivery timing prediction model. For instance, if messages emphasizing discount rates received a high response in the past, the system prioritizes generating discount-focused messages for similar customer segments. Conversely, if new product information received little response, the system switches to other appeal axes, enabling dynamic optimization. Personalization reflecting customer preferences is also possible, such as prioritizing the delivery of future event announcements to customers who indicated interest in participating in events in surveys. It is also possible to automatically conduct A / B tests based on response data to continuously learn and apply the most effective messages and delivery strategies.

[0077] Thus, the system according to this embodiment organically links a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a distribution management unit, a response data collection unit, and an optimization learning unit, and utilizes a generation AI as an essential component to automatically and continuously realize optimized communication for each customer, enabling improved customer satisfaction, reduced churn rates, and the maintenance of long-term customer relationships.

[0078] (Implementation Steps)

[0079] Step 1: Comprehensive collection of customer data The customer data acquisition unit operates on the server and collects attribute information such as customer ID, name, age, gender, residential area, membership rank, point balance, family structure, hobbies and preferences, birthday, anniversaries, past purchase history (purchase date and time, product category, product name, purchase amount, purchase channel, purchase frequency, purchase cycle), browsing history (product page viewing date and time, number of views by category, viewing time, browsing device), cart addition history (product name, addition date and time, non-purchase period, number of times cart was left untouched), favorite registration history (product name, registration date and time, unregistration date and time), review posting history (posting date and time, evaluation content, review content, number of likes on the review), campaign response history (email open date and time, link click date and time, coupon acquisition / use date and time, The system comprehensively stores a vast amount of multifaceted and extensive customer behavior and attribute data in chronological order in a database on the server. This includes campaign participation history, inquiry history (date and time of inquiry, content, response status, inquiry channel), survey responses (interests, purchase intent, improvement requests, satisfaction ratings), app usage history (login frequency, notification read status, in-app search history, in-app banner click history, in-app coupon acquisition and usage history, in-app survey responses, in-app chat history), website usage history (date and time of access, page transition history, form input history, event participation application history), and SNS integration history (SNS account integration status, SNS share history, SNS reaction history).

[0080] Step 2: Customer segmentation using multidimensional features The Customer Segmentation Department extracts multidimensional features from the collected data, such as monthly purchase frequency, average purchase amount, specific category purchase ratio, cart abandonment rate, cart abandonment period, review posting frequency, average review rating, coupon usage rate, app usage frequency, campaign response rate, banner click-through rate, survey response rate, SNS sharing frequency, inquiry frequency, and event participation rate. Using clustering algorithms and supervised learning models, they classify these customers along multiple axes into categories such as "high-frequency purchasing customers," "new customers," "dormant customers," "customers with a preference for specific categories," "customers who abandon their carts," "customers who actively post reviews," "customers who are inclined to use coupons," "app-active customers," "SNS-active customers," "event-participating customers," and "customers with high inquiry frequency."

[0081] Step 3: Generating personalized messages using AI The personalized message generation unit inputs past behavioral history and attribute information for each segment or individual customer into the generation AI, and automatically generates optimal product suggestions, campaign announcements, coupon distribution messages, reminder notifications, review requests, restock announcements, seasonal event announcements, banner ads, in-app pop-up messages, social media posts, survey requests, event invitations, support messages, and more in natural language. Specific examples of prompts to be fed into the generation AI include: "Female in her 20s, has purchased 8 items in the fashion category in the past six months, has items left in her cart. Generate a friendly tone for recommending new fashion items," "Male in his 40s, has a history of numerous purchases in the home appliance category, has items awaiting restock. Generate a restock announcement incorporating expert content," "Inactive customer, no purchase history in the past year, coupon-oriented. Generate a coupon announcement to encourage reactivation and reuse," and "Active review-posting customer, has posted 5 reviews in the past three months. Generate a new product review request expressing gratitude."

[0082] Step 4: Automated delivery on the optimal channel and timing. The distribution management department automatically selects the optimal channel and delivery timing for each customer from multiple channels, including email, app notifications, web push notifications, SMS, SNS direct messages, and postal direct mail. For example, it selects in-app notifications for customers who frequently use the app, email for customers with high email open rates, SMS for timely reminder notifications, SNS direct messages for customers who frequently interact on social media, and paper notices for customers who respond well to postal direct mail. Based on past response data, the AI ​​generates predictions the time of day and event timings when customers are most likely to respond, such as weekday evenings, weekend mornings, just before a sale starts, within 24 hours of adding items to the cart, on their birthday, or the day before an event, and delivers at the optimal time.

[0083] Step 5: Real-time collection of diverse response data The response data collection unit collects response data in real time, including whether the message was opened after delivery, whether it was clicked, whether a coupon was used, whether a product was purchased, whether a review was posted, whether app notifications were read, website access history, banner click history, whether a survey was answered, whether an event was registered, whether social media shares were made, and whether chat support was used, and records this data in the server's database.

[0084] Step 6: Optimization learning and continuous improvement using generative AI The optimization learning unit analyzes the collected response data and feeds it back as training data for the generation AI. For example, it statistically analyzes response rates to specific message expressions and offer content, effectiveness measurements for each delivery channel and timing, response trends by customer attribute and segment, correlation analysis between survey responses and response rates, relationship analysis between event participation rates and campaign utilization rates, review submission rates for review requests, purchase rates of items in the cart for cart abandonment reminder notifications, and click-through rates for banner ads. This data is then reflected in the generation AI's parameters, prompt design, segment classification algorithm weighting, delivery channel selection logic, and delivery timing prediction model. Specific examples of prompts to be fed into the generation AI include: "For customer segments that have shown a high response to messages emphasizing discount rates over the past three months, prioritize generating discount-focused messages in the future," "For customer segments that showed little response to new product information, generate messages emphasizing other appeals (e.g., exclusivity, testimonials, rankings)," and "For customers who indicated interest in participating in events in the survey, generate messages that prioritize future event announcements."

[0085] (Specific use cases)

[0086] For example, when applying the system of the present invention to an e-commerce site, the customer data acquisition unit comprehensively collects customer ID, age, gender, residential area, membership rank, point balance, past purchase history (purchase date and time, product category, product name, purchase amount, purchase channel, purchase frequency), browsing history (product page viewing date and time, number of views by category, viewing time), cart addition history (product name, addition date and time, non-purchase period, number of times cart was left untouched), favorite registration history (product name, registration date and time), review posting history (posting date and time, evaluation content, review content, number of likes on the review), campaign response history (email opening date and time, link click date and time, coupon acquisition / use date and time), inquiry history (inquiry date and time, content, response status), survey response content (interests, purchase intent, improvement requests), app usage history (login frequency, notification read status, in-app search history, in-app banner click history, in-app coupon acquisition / use history), website usage history (access date and time, page transition history, form input history), SNS linking history (SNS account linking status, SNS share history, SNS reaction history), etc., and stores them in a database on the server. Based on this data, the Customer Segmentation Department classifies customers on multiple axes, such as: "High-Frequency High-Value Customers" who have made 10 or more purchases in the past three months with an average purchase amount exceeding 10,000 yen; "Dormant Customers" who have no purchase history in the past year and have a low email open rate; "Fashion-Oriented Customers" who view and purchase more than 80% of fashion items; "Abandoned Cart Customers" who have multiple items in their cart that have not been purchased for more than seven days after being added; "Active Review Posting Customers" who post 5 or more reviews per month with an average rating of 4.5 or higher; "Coupon-Oriented Customers" who use coupons more than 80% of the time; "App Active Customers" who use the app more than 10 times per week; "SNS Active Customers" who share products on SNS more than 5 times a month; "Event-Participating Customers" who have participated in events more than 3 times; and "High-Frequency Inquiry Customers" who make inquiries more than 3 times a month.

[0087] The personalized message generation unit inputs past behavioral history and attribute information for each segment or individual customer into the generation AI. For example, "fashion-oriented customers" might receive "recommended new fashion items," "seasonal limited coordination features," or "fashion coordination diagnosis." Customers who abandon their carts might receive "low stock reminder notifications for items in their cart," "exclusive coupons for items in their cart," or "double points campaigns for cart purchases." High-frequency, high-spending customers might receive "VIP-only pre-sale announcements," "point-up campaign announcements," "exclusive gift offers," or "review accumulation notifications." For "highly active reviewers," the system generates messages such as "thank you coupon for review submission," "request for new product reviews," and "invitation to participate in review rankings." For "coupon-oriented customers," it generates messages such as "exclusive coupon distribution announcement" and "free shipping campaign announcement with coupon use." For "app-active customers," it generates messages such as "new app-exclusive feature announcement" and "invitation to in-app events." For "SNS-active customers," it generates messages such as "point reward announcement for SNS sharing" and "SNS-exclusive campaign announcement." For "event-participating customers," it generates messages such as "invitation to the next event" and "event participation benefits announcement." For "high-frequency inquiry customers," it generates messages such as "enhanced support announcement" and "dedicated chat support announcement." The generating AI achieves detailed personalization based on the customer's age group, past response trends, survey responses, purchase cycle, channels used, and past message response rates, using casual language for younger generations, polite language for seniors, language emphasizing efficiency and time-saving for business people, and language that evokes empathy for parents. Furthermore, by utilizing image generation models and banner generation models, the system automatically generates images such as new product images, seasonal event banner images, limited-edition gift packaging images, event participation certificate images, and illustrations for in-app pop-ups, and combines these with text to create delivery messages and advertising materials.

[0088] The distribution management department automatically selects the optimal channel and delivery timing for each customer from multiple channels, including email, app notifications, web push notifications, SMS, SNS direct messages, and postal direct mail. For example, it selects in-app notifications for customers who frequently use the app, email for customers with high email open rates, SMS for timely reminder notifications, SNS direct messages for customers who frequently interact on social media, and paper notices for customers who respond well to postal direct mail. Based on past response data, the AI ​​generates predictions the time of day and event timings when customers are most likely to respond, such as weekday evenings, weekend mornings, just before a sale starts, within 24 hours of adding items to the cart, on their birthday, or the day before an event, and delivers at the optimal time.

[0089] The response data collection unit collects response data in real time, such as whether the message was opened, clicked, used a coupon, purchased a product, posted a review, read app notifications, website access history, banner click history, answered surveys, registered for events, shared on social media, and used chat support, and records it in the server's database. The optimization learning unit analyzes the collected response data, statistically analyzing things like the click-through rate and purchase rate for "new fashion announcements," the purchase rate of items in the cart for "cart abandonment reminder notifications," the review posting rate for "review requests," the coupon usage rate for "limited coupon announcements," and the registration rate for "event invitations," and reflects this analysis in the parameters of the generating AI, prompt design, segment classification algorithm, delivery channel selection logic, and delivery timing prediction model.

[0090] Specific examples of prompt texts to be fed into the generating AI include: "A woman in her 20s, who has purchased items in the fashion category 8 times in the past 6 months, and has items left in her cart. Generate a friendly tone of recommendation for new fashion items," "A man in his 40s, who has a history of purchasing items in the home appliance category, and has items awaiting restock. Generate a restock notification that includes specialized content," "A dormant customer with no purchase history in the past year, who is coupon-oriented. Generate a coupon notification to encourage re-entry that encourages reuse," "A customer who actively posts reviews, who has posted 5 reviews in the past 3 months. Generate a new product review request that expresses gratitude," and "An event-participating customer who has attended 3 events in the past year. Generate an invitation to the next event that emphasizes the benefits."

[0091] Thus, the system of the present invention automatically generates personalized messages and offers based on diverse customer behavior and attribute data, utilizing generative AI, delivers them through the optimal channel and timing, and continuously optimizes by collecting and learning response data in real time, thereby improving customer satisfaction, reducing churn rates, and maintaining long-term customer relationships.

[0092] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0093] Data generation model 58 is a form of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AI include text generation AI, image generation AI, and multimodal generation AI. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats from audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVMs), k-means clustering, convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each of the above parts is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example.Furthermore, processing performed by AI, including generative AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by AI, including generative AI.

[0094] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart device 14 or an external device, and the smart device 14 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0095] For example, the data collection unit is implemented by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the smart device 14 and processes it with the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using generation AI. For example, the serving unit is implemented by the output device 40 of the smart device 14 or the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the devices and control units is not limited to the examples described above and can be modified in various ways.

[0096] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0097] [Second Embodiment]

[0098] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0099] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0100] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0101] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and communication interface 44 are also connected to the bus 52.

[0102] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0103] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0104] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0105] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0106] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0107] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0108] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the identification processing unit 290 using these models.

[0109] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is realized by the data processing device 12 and the smart glasses 214. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0110] (Example 1)

[0111] Since the process flow is the same as in Example 1 described in the First Embodiment above, the explanation will be omitted.

[0112] (Application Example 1) Since the process flow is the same as in Application Example 1 described in the first embodiment above, the explanation will be omitted.

[0113] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0114] Furthermore, the processing by the data processing system 10 described above is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart glasses 214 or an external device, and the smart glasses 214 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0115] For example, the data collection unit is implemented by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the smart glasses 214 and processes it using the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using generation AI. For example, the serving unit is implemented by the speaker 240 of the smart glasses 214 or the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the device or control unit is not limited to the examples described above and can be modified in various ways.

[0116] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0117] [Third Embodiment]

[0118] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0119] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0120] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0121] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, display 343, and communication interface 44 are also connected to the bus 52.

[0122] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0123] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0124] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0125] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0126] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0127] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0128] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0129] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the headset terminal 314. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0130] (Example 1) Since the process flow is the same as in Example 1 described in the First Embodiment above, the explanation will be omitted.

[0131] (Application Example 1) Since the process flow is the same as in Application Example 1 described in the first embodiment above, the explanation will be omitted.

[0132] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0133] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset terminal 314, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset terminal 314. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the headset terminal 314 or an external device, and the headset terminal 314 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0134] For example, the data collection unit is implemented by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the headset terminal 314 and processes it using the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using generation AI. For example, the serving unit is implemented by the speaker 240 and display 343 of the headset terminal 314 or the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the device or control unit is not limited to the examples described above and can be modified in various ways.

[0135] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0136] [Fourth Embodiment]

[0137] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0138] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0139] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0140] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, controlled object 443, and communication interface 44 are also connected to the bus 52.

[0141] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0142] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0143] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0144] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0145] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0146] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0147] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0148] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0149] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the robot 414. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0150] (Example 1) Since the process flow is the same as in Example 1 described in the First Embodiment above, the explanation will be omitted.

[0151] (Application Example 1) Since the process flow is the same as in Application Example 1 described in the first embodiment above, the explanation will be omitted.

[0152] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0153] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the robot 414 or external devices, and the robot 414 acquires or collects information necessary for processing from the data processing device 12 or external devices.

[0154] For example, the data collection unit is implemented by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the robot 414 and processes it using the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using a generation AI. For example, the serving unit is implemented by the speaker 240 and controlled object 443 of the robot 414 and the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the device or control unit is not limited to the examples described above and can be modified in various ways.

[0155] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0156] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0157] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0158] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0159] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0160] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0161] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0162] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0163] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0164] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0165] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0166] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[0167] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0168] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0169] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0170] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0171] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0172] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0173] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0174] Furthermore, the following additional information is disclosed regarding the above explanation.

[0175] (Claim 1)

[0176] The system comprises a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a delivery management unit, a response data collection unit, and an optimization learning unit. The customer data acquisition unit extracts multifaceted customer data such as customer ID, attribute information, past purchase history, browsing history, cart addition history, review posting history, campaign response history, inquiry history, survey responses, and app usage history from a database on the server. The customer segment classification unit uses clustering algorithms and supervised learning models to classify customers into multiple segments such as high-frequency buyers, new customers, inactive customers, customers with a preference for specific categories, customers who leave their carts unattended, customers who actively post reviews, and customers who are inclined to use coupons, based on the extracted data. The personalized message generation unit uses a generational AI to classify customers into past segments or individual customers. This system is characterized by the following features: it takes user behavior history and attribute information as input and automatically generates optimal product suggestions, campaign announcements, coupon distribution messages, reminder notifications, review requests, restock announcements, and seasonal event announcements in natural language; the distribution management unit automatically distributes the generated messages and offers to each customer by selecting the most suitable channel and distribution timing from multiple channels such as email, app notifications, web push notifications, and SMS; the response data collection unit collects response data in real time after distribution, such as whether the message was opened, clicked, whether a coupon was used, whether a product was purchased, whether a review was posted, whether an app notification was read, and website access history; and the optimization learning unit analyzes the collected response data and feeds it back as learning data for the generative AI to optimize the content of future suggestions, distribution timing, and channel selection. (Claim 2)

[0177] The system according to claim 1, characterized in that the personalized message generation unit provides multidimensional data such as each customer's past purchase history, browsing history, cart addition history, review posting history, campaign response history, inquiry history, survey responses, and app usage history as input to a generation AI, the generation AI estimates each customer's interests and purchasing tendencies, and based on the estimation results, automatically generates optimal product suggestions, campaign announcements, coupons, reminders, review requests, restock announcements, seasonal event announcements, etc., in natural language for each customer, and further dynamically adjusts the style and expression of the generated messages according to the customer's attributes and past response tendencies. (Claim 3)

[0178] The system according to claim 1, characterized in that the distribution management unit analyzes historical data for each customer, such as past response rates by distribution channel, response rates by day of the week and time of day, response rates by offer type, cart abandonment period, review posting frequency, coupon usage history, and app notification read rate, uses generative AI to predict the distribution channel and timing that are most likely to elicit a response from each customer, and automatically determines and delivers the distribution channel and timing of messages and offers based on the prediction results. [Explanation of Symbols]

[0179] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots

Claims

1. A system comprising a customer data acquisition unit, a customer segment classification unit, a personalized message generation unit, a delivery management unit, a response data collection unit, and an optimization learning unit, wherein the customer data acquisition unit extracts data over a predetermined period from a customer database including customer attribute information, past usage history, and past response history; the customer segment classification unit classifies customers into multiple segments using a machine learning algorithm based on the extracted data; the personalized message generation unit analyzes past usage trends and response trends for each segment using a generative AI and automatically generates marketing messages and offers optimized for individual customers; the delivery management unit automatically delivers the generated messages and offers by selecting the optimal channel and delivery timing for each customer from multiple channels such as email, SNS, and app notifications; the response data collection unit collects response data such as whether the message was opened, clicked, and whether the offer was used in real time after delivery and stores it in a database; and the optimization learning unit analyzes the collected response data and feeds it back as learning data for the generative AI, reflecting it in optimizing the content of future messages and delivery timing.

2. The system according to claim 1, characterized in that the personalized message generation unit provides multidimensional data, including each customer's past purchase history, service usage history, inquiry history, and past campaign response history, to a generation AI, the generation AI estimates each customer's interests and purchasing tendencies, and based on the estimation results, automatically generates optimal product suggestions, service information, coupons, campaign information, etc., in natural language for each customer, and further dynamically adjusts the style and expression of the generated messages according to the customer's attributes and past response tendencies.

3. The system according to claim 1, characterized in that the distribution management unit analyzes historical data for each customer, such as past response rates by distribution channel, response rates by day of the week and time of day, and response rates by offer type, uses a generative AI to predict the distribution channel and timing that are most likely to elicit a response from each customer, and automatically determines the distribution channel and timing of messages and offers based on the prediction results and performs distribution.

Images