Programs, methods, information processing devices, and systems

Abstract

The system allows for flexible configuration of the decision-making criteria required for order processing, and outputs accurate order information based on the configured criteria. [Solution] A program for operating a computer comprising a processor and memory, the program causing the processor to perform the following steps: receiving setting operations for a plurality of judgment items necessary for a first AI model to execute order processing via a user interface; having the first AI model learn the received plurality of judgment items and tuning the first AI model to output order information relating to the order content based on the judgment results for each of the plurality of judgment items; receiving order information relating to the order content and extracting character information relating to predetermined characters from the order information; and inputting the extracted character information into the first AI model and causing the first AI model to output order information.

Description

Technical Field

[0005] , , , , ,

[0001] The present disclosure relates to a program, a method, an information processing apparatus, and a system.

Background Art

[0002] In Patent Document 1, characters are recognized from image data of documents using an artificial intelligence for recognition and converted into character data. After extracting character data corresponding to a preset item from the converted character data, a technique for executing an operation related to the extracted character data based on a preset rule is disclosed.

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0003] However, the technique disclosed in Patent Document 1 does not assume a change in a preset item. Further, the technique disclosed in Patent Document 1 does not execute an operation related to the extracted character data using an artificial intelligence that has learned a preset item. Therefore, there is room for improvement in the technique disclosed in Patent Document 1 in terms of flexibly setting an item as needed and executing an appropriate operation corresponding to the set item.

[0004] An object of the present disclosure is to enable flexible setting of an item as needed and to realize an appropriate output corresponding to the set item.

Means for Solving the Problems

[0005] To solve the aforementioned problems, a program according to one aspect of the present disclosure is a program for operating a computer comprising a processor and memory, the program causing the processor to perform the following steps: receiving a setting operation for a plurality of judgment items necessary for a first AI model to execute order processing via a user interface; having the first AI model learn the received plurality of judgment items and tuning the first AI model to output order information relating to the order content based on the judgment result for each of the plurality of judgment items; receiving order information relating to the order content and extracting character information relating to predetermined characters from the order information; and inputting the extracted character information into the first AI model and causing the first AI model to output order information. [Effects of the Invention]

[0006] The system allows for flexible configuration of the criteria necessary for order processing, and can output accurate order information based on the configured criteria. [Brief explanation of the drawing]

[0007] [Figure 1] This is a block diagram showing the overall configuration of System 1. [Figure 2] This is a block diagram showing an example of the functional configuration of the terminal device 10. [Figure 3] This block shows a functional configuration example for server 20. [Figure 4] This diagram shows the data structure of the Decision Rule Table 2021. [Figure 5] This diagram shows the data structure of the Customer Master Table 2022. [Figure 6] This diagram shows the data structure of the product master table 2023. [Figure 7] This table shows the data structure of the Order Information Table 2024. [Figure 8] This table shows the data structure of the general-purpose master table 2025. [Figure 9]This flowchart shows an example of the order processing flow in System 1. [Figure 10] This figure shows an example of the screen in this disclosure. [Figure 11] This is a block diagram showing the basic hardware configuration of Computer 90. [Modes for carrying out the invention]

[0008] The embodiments of this disclosure will be described below with reference to the drawings. In all the drawings illustrating the embodiments, common components are denoted by the same reference numerals, and repeated explanations are omitted. The following embodiments are not intended to unduly limit the content of this disclosure as described in the claims. Not all components shown in the embodiments are necessarily essential components of this disclosure. Also, each drawing is a schematic diagram and is not necessarily a strict illustration.

[0009] Furthermore, in the following description, "processor" refers to one or more processors. At least one processor is typically a microprocessor such as a CPU (Central Processing Unit), but may be another type of processor such as a GPU (Graphics Processing Unit). At least one processor may be single-core or multi-core.

[0010] Furthermore, at least one processor may be a broad-sense processor, such as a hardware circuit that performs some or all of the processing (e.g., an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit)).

[0011] In the following description, the expression such as "xxx table" may be used to describe information from which an output is obtained for an input. This information may be data of any structure or a learning model such as a neural network that generates an output for an input. Therefore, "xxx table" can be referred to as "xxx information".

[0012] In the following description, the configuration of each table is an example. One table may be divided into two or more tables, or all or part of two or more tables may be one table.

[0013] In the following description, the "program" may be used as the subject to describe a process. However, since the program is executed by a processor to perform a defined process while appropriately using a storage unit and / or an interface unit, etc., the subject of the process may be the processor (or a device such as a controller having the processor).

[0014] The program may be installed in a device such as a computer, or may be in, for example, a program distribution server or a computer-readable (e.g., non-transitory) recording medium. In the following description, two or more programs may be realized as one program, or one program may be realized as two or more programs.

[0015] In the following description, an identification number is used as identification information for various objects, but other types of identification information (e.g., an identifier including letters and symbols) may be adopted.

[0016] In the following description, when describing elements of the same type without distinction, reference signs (or common signs among the reference signs) are used, and when describing elements of the same type by distinction, identification numbers (or reference signs) of the elements may be used.

[0017] In the following description, the control lines and information lines show those considered necessary for the description, and do not necessarily show all the control lines and information lines on the product. All components may be interconnected.

[0018] Each information processing device is composed of a computer having an arithmetic unit and a storage unit. The basic hardware configuration of the computer and the basic functional configuration of the computer realized by the hardware configuration will be described later. For each of the server 20 and the terminal device 10, descriptions overlapping with the basic hardware configuration and the basic functional configuration of the computer described later will be omitted.

[0019] <Overview> The system according to this embodiment receives, via a user interface, setting operations for a plurality of judgment items required when the first AI model executes an order receiving process. The first AI model is, for example, a large language model (Large Language Models: hereinafter, LLM). The system according to this embodiment causes the first AI model to learn the received plurality of judgment items, and tunes the first AI model so as to output order receiving information regarding the order receiving content based on the judgment results for each of the plurality of judgment items. The system according to this embodiment receives order information regarding the order content, and extracts character information regarding a predetermined character from the order information. The system according to this embodiment inputs the extracted character information into the first AI model, and causes the first AI model to output order receiving information. Note that the system according to this embodiment is applicable to various order receiving processes regardless of the order target, order receiving target, order form, and the like.

[0020] <1. Configuration Diagram of the Whole System> FIG. 1 is a block diagram showing an example of the overall configuration of the system 1. As shown in FIG. 1, the system 1 includes, for example, a terminal device 10, a server 20, and an AI system 30. The terminal device 10, the server 20, and the AI system 30 are communicatively connected via, for example, a network 80.

[0021] Figure 1 shows an example where System 1 includes one terminal device 10, but the number of terminal devices 10 included in System 1 is not limited to one. System 1 may include two or more terminal devices 10.

[0022] In this embodiment, a collection of multiple devices may be treated as a single server. The method of distributing the multiple functions required to implement the server 20 according to this embodiment to one or more hardware can be appropriately determined in view of the processing capacity of each hardware and / or the specifications required for the server 20. In addition, in this embodiment, the server 20 may also include the functions of the AI ​​system 30.

[0023] [1-1. Terminal Devices] Terminal device 10 is an information processing device used by users who utilize the order service provided by server 20. The user is, for example, a business that receives orders and delivers goods to the customer.

[0024] The terminal device 10 can be implemented as, for example, a stationary PC (Personal Computer), a laptop PC, or a head-mounted display. Alternatively, the terminal device 10 may be a portable computer such as a smartphone or a tablet device.

[0025] The terminal device 10 comprises a communication interface (IF) 12, an input device 13, an output device 14, a memory 15, storage 16, and a processor 19. The communication interface 12 is an interface for inputting and outputting signals so that the terminal device 10 can communicate with devices in system 1, such as a server 20. The input device 13 is a device for receiving input operations from the user (e.g., a touch panel, touchpad, pointing device such as a mouse, keyboard, etc.). The output device 14 is a device for outputting information to the user (display, speaker, etc.). The memory 15 is for temporarily storing programs and data processed by programs, etc., and is a volatile memory such as DRAM (Dynamic Random Access Memory). The storage 16 is for saving data, and is a flash memory, HDD (Hard Disk Drive), etc. The processor 19 is hardware for executing the instruction set written in the program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0026] [1-2. Server] Server 20 is an information processing device that provides order services. Server 20 is an information processing device implemented by, for example, a computer connected to network 80.

[0027] The server 20 comprises a communication interface 22, an input device 23, an output device 24, a memory 25, storage 26, and a processor 29. The communication interface 22 is an interface for inputting and outputting signals so that the server 20 can communicate with devices in system 1, such as the terminal device 10. The input device 23 is a device for receiving input operations from the person in charge of order services (hereinafter abbreviated as "person in charge"). The output device 24 is a device for outputting information to the person in charge. The memory 25 is for temporarily storing programs and data processed by programs, etc., and is a volatile memory such as DRAM. The storage 26 is for saving data, and is a flash memory, HDD, etc. The processor 29 is hardware for executing the instruction set written in the program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0028] [1-3. AI System] AI System 30 is a system that generates and outputs order information related to order details. Order information includes, for example, basic order information (e.g., order date, name of ordering company and contact person, etc.), product information (e.g., product name, quantity, unit price, total amount, etc.), and delivery / delivery date information (e.g., delivery date, delivery method, etc.). There are no particular limitations on the output format of the order information, but a structured digital data format is mainly envisioned, assuming integration with the internal systems of the ordering company, which is the user. Specifically, formats such as CSV (Comma Separated Values), XML (Extensible Markup Language), and JSON (JavaScript Object Notation) are envisioned. However, the content and output format of the order information are not limited to the examples above and can be flexibly adjusted according to the set judgment items and output prompts (both detailed below).

[0029] The AI ​​system 30 is physically built on one or more server computers. The AI ​​system 30 is connected to the server 20 via a network 80 and provides various functions through an API (Application Programming Interface). The AI ​​system 30 may be built on, for example, a cloud-based infrastructure. The AI ​​system 30 includes, for example, an AI-OCR (Optical Character Recognition) function and a first AI model.

[0030] The AI-OCR function is used to extract text information from order information in an unstructured format. Order information is information about the order content and includes image data related to unstructured documents such as faxes, emails, and PDF files containing the order content, as well as structured digital data in formats such as WebEDI (Web Electronic Data Interchange) (e.g., CSV, XML, JSON, etc.). The AI ​​system 30 extracts text information using the AI-OCR function when the acquired order information is in an unstructured format such as a fax.

[0031] On the other hand, if the acquired order information is in a format that does not require image recognition by AI-OCR functionality, it is possible to extract text information from the order information using RPA (Robotic Process Automation), conventional text analysis technology, rule-based dedicated programs, or conventional OCR software. In this embodiment, the RPA provided on server 20 is used to extract the text information. For example, AI system 30 may be equipped with RPA, or either server 20 or AI system 30 may extract the text information using a method other than RPA. Examples of "information in a format that does not require image recognition" include data exchanged via WebEDI, specific structured email bodies, text files or CSV files output from existing systems, or standardized documents such as order forms with fixed layouts.

[0032] In this embodiment, the server 20 determines whether the acquired order information is in a non-standard format or a standard format, but the AI ​​system 30 may also make this determination.

[0033] Textual information refers to information about specific characters included in the order information. These specific characters are, for example, characters that are essential or useful for identifying the order details, and may include, for example, the order date, basic information of the ordering party (company name, contact person's name, etc.), product name, total amount, unit price, quantity, delivery date, and remarks.

[0034] The AI-OCR function is implemented, for example, by multiple AI models different from the first AI model. In other words, the AI ​​system 30 includes, for example, an AI-OCR model (hereinafter collectively referred to as the AI-OCR model) in addition to the first AI model. The AI-OCR model consists of, for example, a character detection model that identifies areas containing predetermined characters from image data as order information, a character recognition model that converts predetermined characters displayed in the identified areas into text data, and a layout analysis model that understands the structure of items, tables, etc., displayed in the image data. These AI models are, for example, well-known machine learning models such as neural networks.

[0035] For example, the functions of the character detection model, character recognition model, and layout analysis model may each be implemented by a single AI model. Also, for example, AI system 30 does not need to have AI-OCR functionality. In this case, for example, a different AI system (not shown) or server 20 may have AI-OCR functionality. Also, for example, if the order information subject to the order service is only in a format that does not require image recognition, system 1 itself does not need to have AI-OCR functionality.

[0036] The first AI model is an AI model that takes text information as input data and generates and outputs order information. Specifically, for example, the first AI model has pre-trained multiple judgment items and judgment criteria corresponding to each of those judgment items, and it determines for each of the multiple judgment items whether a predetermined character contained in the input text information satisfies the judgment criteria. Then, based on the judgment results for each of the multiple judgment items, the first AI model identifies the order details and generates and outputs order information that shows the identified order details.

[0037] Therefore, order information can be described as information that indicates the content of an order, which is determined / identified according to judgment criteria for each of several judgment items, based on the content of the text information extracted from the order information. Accordingly, as mentioned above, order information may include various items, but is not limited to these, and its composition is determined by the set judgment items.

[0038] The judgment criteria are items necessary for the first AI model to execute order processing, and specifically, they are indicators for determining whether the text information contains the characters necessary to generate order information. The judgment criteria include, for example, items that constitute the order details shown in the order information, and key items for determining whether the text information contains the characters corresponding to those items. Examples of judgment criteria include product name, quantity, delivery date, customer name (name of a company with a history of placing orders), and unit price.

[0039] Criteria are rules and / or conditions that define how judgment items should be recognized, interpreted, or processed based on that recognition / interpretation. For example, when comparing specific judgment items (e.g., product name, customer name) with pre-registered master data (e.g., product master, customer master), the criteria would be conditions such as "Is the product name registered in the product master?" or "Does it match the default individual criterion ID in the customer master?".

[0040] In this embodiment, the judgment criteria include common criteria that apply to all customers and individual criteria that apply individually to each customer. The individual criteria are set to accommodate different business practices, variations in terminology, etc., for each customer. The judgment criteria may consist of, for example, only common criteria or only individual criteria. Also, for example, it is not necessary for the first AI model to pre-train the judgment criteria; it is sufficient for the first AI model to pre-train the judgment items.

[0041] In other words, the first AI model learns multiple judgment items and corresponding judgment criteria as its internal judgment logic, and is pre-tuned to output order information based on the judgment results for each of the multiple judgment items. To put it another way, the AI ​​system 30 functions as an AI agent that, when instructed to perform tasks such as generating and outputting order information, autonomously determines whether the content of the textual information satisfies the judgment criteria. Such an autonomous AI agent, for example, when given a goal, generates tasks to achieve that goal, collects information to enable the AI ​​model to execute the generated tasks, and outputs information that achieves the goal by repeatedly processing the AI ​​model to execute the tasks.

[0042] The tuning of the first AI model is triggered, for example, by the setting of multiple decision items and corresponding decision criteria by the person in charge. In other words, the person in charge can flexibly set and change the decision items and decision criteria according to the business content and order requirements of the user, the orderer. This enables order processing that can be adapted to changes in the company's business processes and diverse transaction forms.

[0043] The setting of judgment items and criteria is performed via a user interface. The user interface can be of any type; in this embodiment, the user interface is a settings screen displayed on the server 20's display 241 (see Figure 10). By performing the setting operation via a user interface, the person in charge can easily and accurately understand the settings and changes they have made, improving the convenience of setting judgment items and criteria.

[0044] Specifically, for example, server 20 accepts setting operations on the settings screen and sends the accepted decision items and criteria to AI system 30. Then, AI system 30 receives the decision items and criteria and inputs them into the first AI model, thereby starting the tuning of the first AI model.

[0045] It is not mandatory that the input of decision criteria to the first AI model be based on the settings made by the operator. For example, another AI model different from the first AI model may generate and output decision criteria, and the AI ​​system 30 may input these outputted decision criteria into the first AI model (see the second modified example below).

[0046] There are no particular limitations on the tuning method for the first AI model; for example, it may be fine-tuning or prompt tuning. In the case of fine-tuning, the judgment items and criteria received from the server 20 become the training dataset. In the case of prompt tuning, the AI ​​system 30 inputs a prompt to the first AI model that includes, for example, the judgment items and criteria received from the server 20, and an instruction statement that instructs the AI ​​to output order information based on the judgment results in accordance with the judgment items and criteria. The prompt may be generated, for example, using a template pre-stored in the AI ​​system 30, or it may be automatically generated based on the judgment items and criteria.

[0047] The first AI model may be a generative AI model such as an LLM, or a machine learning model other than a generative AI, such as a sequence labeling model. In this embodiment, the AI ​​system 30 is assumed to have an LLM as the first AI model. The number of LLMs may be one or more.

[0048] For example, when the AI ​​system 30 inputs text information into the first AI model and outputs order information, it may also input a prompt (hereinafter referred to as an output prompt) that includes an instruction to output the order information. This allows for more precise control over the behavior of the first AI model and increases the likelihood of obtaining the desired output result. In this embodiment, it is assumed that an output prompt is input to the first AI model (see Figure 4, etc.).

[0049] LLMs are single-modal natural language models built by training on large amounts of text data, and are used in many NLG (Natural Language Generation) tasks, such as generating responses to specific questions, automatically generating text, and summarizing text data. Examples of LLMs include the following: • OpenAI: GPT-4 Google: Gemini 1.5 Flash Anthropic:Claude 3.5 Sonnet

[0050] Each information processing device consists of a computer equipped with an arithmetic unit and a memory device. The basic hardware configuration of the computer and the basic functional configuration of the computer realized by said hardware configuration will be described later. For each of the terminal device 10, server 20, and AI system 30, explanations that overlap with the basic hardware configuration and basic functional configuration of the computer described later will be omitted.

[0051] <2. Configuration of terminal equipment> Figure 2 is a block diagram showing an example of the functional configuration of the terminal device 10. As shown in Figure 2, the terminal device 10 includes a communication unit 120, an input device 13, an output device 14, an audio processing unit 17, a microphone 171, a speaker 172, a camera 160, a location information sensor 150, a storage unit 180, and a control unit 190. Each block included in the terminal device 10 is electrically connected, for example, by a bus.

[0052] The communication unit 120 performs processing such as modulation and demodulation processing for the terminal device 10 to communicate with other devices. The communication unit 120 performs transmission processing on the signal generated by the control unit 190 and transmits it to an external source (for example, the server 20). The communication unit 120 performs reception processing on the signal received from an external source and outputs it to the control unit 190.

[0053] The input device 13 is a device for a user operating the terminal device 10 to input instructions or information. The input device 13 can be implemented, for example, by a touch-sensitive device 131 on which instructions are input by touching the operating surface. If the terminal device 10 is a PC, the input device 13 may be implemented by a reader, keyboard, mouse, etc. The input device 13 converts the instructions input by the user into electrical signals and outputs the electrical signals to the control unit 190. The input device 13 may also include, for example, a receiving port that accepts electrical signals input from an external input device.

[0054] The output device 14 is a device for presenting information to the user operating the terminal device 10. The output device 14 is implemented, for example, by a display 141. The display 141 displays data according to the control of the control unit 190. The display 141 is implemented, for example, by an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display.

[0055] The audio processing unit 17 performs, for example, digital-to-analog conversion processing of the audio signal. The audio processing unit 17 converts the signal received from the microphone 171 into a digital signal and provides the converted signal to the control unit 190. The audio processing unit 17 also provides the audio signal to the speaker 172. The audio processing unit 17 is implemented, for example, by an audio processing processor. The microphone 171 receives an audio input and provides the audio signal corresponding to that audio input to the audio processing unit 17. The speaker 172 converts the audio signal received from the audio processing unit 17 into audio and outputs the audio to the outside of the terminal device 10.

[0056] Camera 160 is a device that receives light using a photodetector and outputs it as a shooting signal.

[0057] The location information sensor 150 is a sensor that detects the position of the terminal device 10, and is, for example, a GPS (Global Positioning System) module. A GPS module is a receiving device used in a satellite positioning system. In a satellite positioning system, signals are received from at least three or four satellites, and the current position of the terminal device 10, which is equipped with a GPS module, is detected based on the received signals. The location information sensor 150 may also detect the current position of the terminal device 10 from the position of the wireless base station to which the terminal device 10 is connected.

[0058] The storage unit 180 is implemented by, for example, memory 15 and storage 16, and stores data and programs used by the terminal device 10. The storage unit 180 stores, for example, user information 181.

[0059] User information 181 includes, for example, information about the user who uses the terminal device 10. User information includes, for example, the user's name, age, address, date of birth, contact information, etc.

[0060] The control unit 190 is realized when the processor 19 reads a program stored in the memory unit 180 and executes instructions contained in the program. The control unit 190 controls the operation of the terminal device 10. By operating according to the program, the control unit 190 performs the functions of an operation reception unit 191, a transmission / reception unit 192, and a presentation control unit 193.

[0061] The operation reception unit 191 processes instructions or information input from the input device 13. Specifically, for example, the operation reception unit 191 receives instructions or information input from a touch-sensitive device 131 or the like.

[0062] Furthermore, the operation reception unit 191 receives voice instructions input from the microphone 171. Specifically, for example, the operation reception unit 191 receives voice signals input from the microphone 171 and converted into digital signals by the voice processing unit 17. The operation reception unit 191 obtains instructions from the user by, for example, analyzing the received voice signals and extracting predetermined nouns.

[0063] The transmitting / receiving unit 192 performs processing to enable the terminal device 10 to send and receive data with an external device such as the server 20 in accordance with a communication protocol. Specifically, for example, the transmitting / receiving unit 192 sends information input by the user or instructions from the user to the server 20. The transmitting / receiving unit 192 also receives information provided by the server 20.

[0064] The presentation control unit 193 controls the output device 14 in order to present information provided by the server 20 to the user. Specifically, for example, the presentation control unit 193 displays the information transmitted from the server 20 on the display 141. The presentation control unit 193 also outputs the information transmitted from the server 20 through the speaker 172.

[0065] <3. Functional Configuration of the Server> Figure 3 shows an example of the functional configuration of server 20. As shown in Figure 3, server 20 functions as a communication unit 201, a storage unit 202, and a control unit 203.

[0066] The communication unit 201 performs processing for the server 20 to communicate with external devices. The storage unit 202 stores, for example, the decision rule table 2021, the customer master table 2022, the product master table 2023, the order information table 2024, and the general-purpose master table 2025. The data tables stored in the storage unit 202 are not limited to these.

[0067] The Decision Rules Table 2021 is, for example, a table that stores decision items and decision criteria. The Customer Master Table 2022 is, for example, a table that stores customer master data, which is various information about the user's customers. The Product Master Table 2023 is, for example, a table that stores product master data, which is various information about the products that the user delivers to the ordering party. The Order Information Table 2024 is, for example, a table that stores order information and various related information. The General-Purpose Master Table 2025 is, for example, a table that stores general-purpose master data, which is various information that is commonly used throughout System 1. Details of these tables will be described later.

[0068] The control unit 203 is realized when the processor 29 reads a program stored in the memory unit 202 and executes instructions contained in the program. The program includes applications such as web browser applications. The program includes programming languages ​​such as JavaScript® that are executed on the web browser application stored in the terminal device 10. By operating according to the program, the control unit 203 performs the functions of the receiving control module 2031, the transmitting control module 2032, the presentation control module 2033, and the order processing module 2034.

[0069] The receive control module 2031 controls the process by which the server 20 receives signals from an external device according to a communication protocol. The transmit control module 2032 controls the process by which the server 20 transmits signals to an external device according to a communication protocol. The presentation control module 2033 controls the process of presenting various information to the user and the person in charge of the order service.

[0070] The order processing module 2034 provides the AI ​​system 30 with information necessary for, for example, extracting character information using the AI-OCR function and generating and outputting order information using the first AI model. The order processing module 2034 instructs the AI ​​system 30 to activate the AI-OCR function and execute various processes using the first AI model. The order processing module 2034 transmits the output data of the first AI model received from the AI ​​system 30 to the terminal device 10.

[0071] <4. Data Structure> Referring to Figures 4 to 8, the data structure of the various tables stored in the storage unit 202 of the server 20 will be explained. Note that the data structure described is just one example, and does not exclude any data not listed. Furthermore, even data listed in the same table may be stored in separate storage areas within the storage unit 202. In addition, the various tables described later may also be stored in the AI ​​system 30.

[0072] [4-1. Decision-Making Rule Table] Figure 4 shows the data structure of the decision rule table 2021. As shown in Figure 4, the decision rule table 2021 has columns such as Criteria Name, Applicable Target, Customer ID, Decision Item Name, Decision Logic, Priority, and Valid Flag, with Criteria ID as the key.

[0073] The Criteria ID is a column that stores an identifier to uniquely identify the judgment criterion. The Criteria Name is a column that stores the name of the judgment criterion. The Applicability is a column that stores information to identify whether the judgment criterion is a "common rule" that applies to all customers in general, or a "specific rule" that applies individually to a particular customer. The Customer ID is a column that stores the identification ID of the customer to whom the rule applies, if the Applicability is a specific rule. The Judgment Item Name is a column that stores the name of the judgment item to which the judgment criterion applies (e.g., product name, quantity, delivery date, customer name, etc.).

[0074] The decision logic is a column that stores specific rules and / or conditions that define how the decision item should be recognized, interpreted, or processed based on that recognition / interpretation. For example, it may store master reference conditions, fixed values, or prompt instructions.

[0075] Master reference conditions define whether the extracted text information matches various master data (e.g., customer master data, product master data, general master data), or how the information within that master should be referenced.

[0076] In the case of matching product names, the master reference condition is to check whether the product name extracted from the order information matches the product name registered in the product master table 2023. For example, if the order form contains the description "notebook PC," it will be compared with the "notebook PC" entry in the product master table 2023, and related information such as the product code and unit price will be retrieved.

[0077] When linking information about the ordering source, the master reference condition is determined by checking whether the name of the ordering source extracted from the order information matches the customer name registered in the customer master table 2022, and then referencing the customer ID, default individual criterion ID (rules applied to a specific ordering source), etc.

[0078] In the case of unit / code conversion, the master reference conditions define whether the units (e.g., "pieces", "sets") and specific codes (e.g., primary store codes) included in the order information match the corresponding canonical units and codes registered in the general master table 2025, or whether they should be converted according to the conversion rules.

[0079] A fixed value instructs the system to always apply a specific value to a particular decision item. For example, in the case of applying a default value, if the "delivery method" for an order of a particular product is always "home delivery," then a fixed value of "home delivery" is set for the decision item "delivery method," and the AI ​​model is instructed to apply this automatically. Alternatively, in the case of information supplementation, if there are items that are not included in the order information but are essential for business operations (e.g., order processing department, specific processing code, etc.), a default fixed value is set to apply to that item, and the AI ​​model is instructed to apply this automatically.

[0080] A prompt instruction is a command and / or rule that directly instructs the first AI model in natural language on how to interpret textual information and generate order information. For example, a prompt instruction might specify the output format of the order information, such as, "Extract the product name, quantity, and unit price from the following order information and output them in comma-separated CSV format." Another example is a prompt instruction that specifies the priority and / or interpretation method for ambiguous information, such as, "If there are multiple candidates for the product name, select the one that is most relevant." Yet another example is a prompt instruction that incorporates a specific business rule, such as, "If the quantity is not specified, complete it with '1'."

[0081] The priority column stores information indicating which criterion takes precedence when multiple criteria are applicable. The validity flag column stores information indicating whether the criterion is currently valid or not.

[0082] The various pieces of information stored in the decision rule table 2021 are changed and updated, for example, when the input device 23 receives an operation from the person in charge.

[0083] [4-2. Customer Master Table] Figure 5 shows the data structure of the customer master table 2022. As shown in Figure 5, the customer master table 2022 has columns for customer name, individual criteria ID, contact information, and transaction terms, with customer ID as the key. A customer is, for example, a company that has placed an order with a user, that is, a company with a history of being an orderer. In this embodiment, the degree of history is not considered when determining whether or not a company is a customer.

[0084] The Customer ID column stores an identifier to uniquely identify the customer. The Customer Name column stores the name of the customer. The Individual Criterion ID column stores an identifier for the individual criteria that are applied to the customer by default. The Contact column stores information about the customer's contact details. The Terms and Conditions column stores the terms and conditions of the transaction agreed upon with the customer.

[0085] [4-3. Product Master Table] Figure 6 shows the data structure of the product master table 2023. As shown in Figure 6, the product master table 2023 has columns for product name, unit price, and related information, with product code as the key, for example.

[0086] The Product Code column stores a code that uniquely identifies each product. The Product Name column stores the name of the product. The Unit Price column stores the unit price of the product. The Related Information column stores information such as the product's inventory status and / or pointers to related products.

[0087] [4-4. Order Information Table] Figure 7 shows the data structure of the order information table 2024. As shown in Figure 7, the order information table 2024 has columns for order information, text information, AI processing status, processing result, and application log, with order ID as the key.

[0088] The Order ID column stores an identifier that uniquely identifies each instance of order processing. The Order Information column stores order information uploaded to Server 20 (specifically, file path, source text data, etc.). The Character Information column stores character information extracted from the Order Information. The AI ​​Processing Status column stores information indicating the current state of processing by the AI ​​System 30. The Processing Result column stores order information generated and output by the First AI Model. The Application Log column stores a log of the judgment items and criteria applied in the order processing.

[0089] [4-5. General-purpose master table] Figure 8 shows the data structure of the general-purpose master table 2025. As shown in Figure 8, the general-purpose master table 2025 has columns for name and setting details, with code as the key, for example.

[0090] The Code column stores an identifier that uniquely identifies codes used commonly throughout System 1. Information stored in the Code column includes, for example, a unit code representing the quantity unit in an order, a status code corresponding to the status of each stage of order processing, a payment code representing the payment method, and an error code indicating the type of error that may occur during the processing of the order service. The Name column stores a human-readable display name and / or description corresponding to the Code. The Settings column stores specific setting values, conversion rules, conditions, supplementary information, etc., related to the item identified by the Code.

[0091] <5. Operation> The following describes an example of the order processing flow in System 1, with reference to Figure 9. Figure 9 is a flowchart showing an example of the order processing flow in System 1.

[0092] In step S101 of the flowchart shown in Figure 9, the server 20 accepts setting operations for multiple decision items via the user interface. The server 20 also accepts setting operations for decision criteria corresponding to each of the multiple decision items via the user interface.

[0093] Specifically, for example, the user operates the terminal device 10 and inputs multiple decision items necessary for order processing and their corresponding decision criteria through a user interface provided by the presentation control module 2033 via the presentation control unit 193. This input includes initial settings of the decision items and decision criteria, as well as changes to at least one of the decision items and decision criteria after the initial settings. The transmitting / receiving unit 192 transmits the received decision items and decision criteria to the server 20, for example.

[0094] The receiving control module 2031 receives, for example, the decision items and decision criteria transmitted from the terminal device 10. This allows the server 20 to accept the settings for the decision items and decision criteria. The transmitting control module 2032 then transmits, for example, the received decision items and decision criteria to the AI ​​system 30.

[0095] In step S102, the server 20 trains the first AI model with the multiple judgment items and corresponding judgment criteria it has received, and tunes the first AI model to output order information based on the judgment results for each of the multiple judgment items.

[0096] Specifically, for example, the AI ​​system 30 inputs the judgment items and criteria received from the server 20 into the first AI model and begins tuning the first AI model. As a result, the server 20 tunes the first AI model via the AI ​​system 30. Through this tuning, the first AI model functions as an AI agent that judges and outputs order information in accordance with the set judgment items and criteria.

[0097] In step S103, the server 20 receives order information and extracts character information from the order information.

[0098] Specifically, for example, order information is sent from the user in an unstructured format such as fax, email, or WebEDI. The order processing module 2034 determines, for example, whether the acquired order information is in an unstructured format or in a structured format that does not require image recognition by AI-OCR functionality.

[0099] If the order information is in an unstructured format, the order processing module 2034, for example, sends the order information to the AI ​​system 30 and instructs the AI ​​system 30 to extract text information using the AI-OCR function. The AI ​​system 30 receives the instruction from the server 20 and extracts text information from the order information using the AI-OCR function. On the other hand, if the order information is in a structured format, the order processing module 2034, for example, controls the RPA installed on the server 20 to have the RPA extract the text information. The order processing module 2034 acquires the text information extracted by the RPA and sends it to the AI ​​system 30. As a result, the server 20 extracts text information from the order information, sometimes via the AI ​​system 30.

[0100] The extracted textual information may then be chunked into sets of information sufficient to understand the order details (see the first modified example below). The order information and textual information are recorded in the order information table 2024, and the AI ​​processing status is updated to "extracted," etc.

[0101] In step S104, the server 20 inputs the extracted character information into the first AI model, causing the first AI model to output order information.

[0102] Specifically, for example, the AI ​​system 30 inputs the extracted character information along with an output prompt to the first AI model. At this time, the AI ​​system 30 performs, for example, the function of RAG (Retrieval-Augmented Generation). That is, for example, when causing the first AI model to generate order information, the AI ​​system 30 searches for character information and / or information related to judgment items from various data tables stored in the memory unit 202. The AI ​​system 30 provides, for example, the search results to the first AI model as contextual information.

[0103] The first AI model, while following pre-learned judgment items and criteria, refers to richer and more accurate information (contextual information) to determine whether a predetermined character in the input text information meets the judgment criteria for each of the multiple judgment items. Based on the judgment results for each of the multiple judgment items and the contextual information, the first AI model identifies the order details and generates order information.

[0104] The generated order information is output in a format (e.g., CSV, XML, JSON, etc.) that can be linked to the user's internal systems (e.g., core systems, sales management systems, inventory management systems, etc.). This linkage can be achieved through various methods such as API linkage, file linkage, and RPA linkage. The generated order information is also recorded in the Order Information Table 2024, and the AI ​​processing status is updated to "Processed," etc. Furthermore, a log of the decision criteria applied to order processing is also recorded in the Order Information Table 2024.

[0105] In the example shown in Figure 9, the server 20 executes the processes in step S102 and step S103 consecutively, but this is not the only case. The server 20 may, for example, execute each process up to step S102 separately from each process from step S103 onward. This is because the execution contexts for each process up to step S102 and each process from step S103 onward may differ.

[0106] <6. Screen Example> This disclosure describes an example of a screen displayed on the display 241 of the server 20. Figure 10 shows a settings screen 2411, which is a user interface for accepting setting operations for judgment items and judgment criteria. The settings screen allows, for example, the configuration of settings corresponding to the "Customer-Specific Rule Application" phase of the flowchart displayed at the top of the screen. Note that this screen example is merely an example, and various screen configurations and screen contents can be adopted.

[0107] As shown in Figure 10, the main part of the settings screen 2411 is composed of a grid-style table. In the row direction of the table, there are settings areas for "common rules" and settings for specific customers (e.g., "Customer CD: 10001 Customer A", "Customer CD: 10002 Customer B", etc.). This allows the person in charge to flexibly define rules that apply to all customers in general and rules that are specific to particular customers.

[0108] In the table, the "decision items" to be configured are displayed as headers in the column direction, making items such as "product code," "data source center code," "invoice number," and "primary store code" visible. Within each decision item's cell, the decision logic applied to that item and its specific settings are described.

[0109] For example, in the "Common Rules" cell for "Product Code," the decision logic "1. Master Search" is set, and the details of the parameters are specifically shown as follows: "Search Target Master: Past Order Data," "Search Target Column: Product Name," "Items Used: Product Name / Specifications," "Filter Column: Customer Name," "Value Used for Filter: Specified Customer Name," "AI Re-evaluation Criteria: Use Common Settings," and "Retrieve Target Column: Product ID." Also, in the "Common Rules" cell for "Invoice Number," the decision logic for "1. Fixed Value" is set to "<Invoice Number>," and in the "Invoice Number" cell for "Customer CD: 10001 Customer A," the decision logic for "1. Fixed Value" is set to "<Customer A>," which is a fixed value specific to that customer.

[0110] At the top of the settings screen are a search bar labeled "Search for customers...", a "+" icon for adding new rules, and an icon for saving settings. This allows the person in charge to intuitively manage decision items and criteria, improving the convenience of setting operations.

[0111] The settings performed on this settings screen are received by the server 20's receiving control module 2031 and sent to the server 20 as judgment items and judgment criteria. Subsequently, these settings are sent to the AI ​​system 30 and used for tuning the first AI model.

[0112] By providing a user interface like the one shown in this example screen, employees can easily customize the decision-making logic for order processing by the AI ​​system 30 to suit their own work content and / or order requirements. This enhances the flexibility and practicality of the AI ​​system 30.

[0113] <7.Summary> As described above, in this embodiment, the server 20 tunes the first AI model based on a plurality of judgment items and judgment criteria received via the user interface. Using the tuned first AI model, the server 20 executes a series of processes to generate order information from character information extracted from order information and output it to the user's internal system. At this time, the AI ​​system 30 performs RAG functionality and cooperates with various master tables to support the generation of more accurate order information.

[0114] This allows for flexible configuration of the decision-making criteria required for order processing, according to the nature of the work and the type of transaction. Furthermore, it enables the automatic output of accurate order information in accordance with the configured decision-making criteria. As a result, the burden of manual data entry and human judgment is significantly reduced, leading to increased efficiency in the overall order processing, improved data accuracy, and a reduction in human error.

[0115] Furthermore, this embodiment can handle various forms of order information (FAX, email, WebEDI, etc.) and can be easily integrated with existing in-house systems, making it easier to implement and operate in a wide range of business environments.

[0116] <8. Variation> [8-1. First variation: Input method of character information to the first AI model] Various input methods are conceivable when inputting character information into the first AI model. For example, the AI ​​system 30 may chunk the character information extracted from the order information into sets of information that allow for understanding the order content. Then, the AI ​​system 30 may input the chunked sets of information as a single unit into the first AI model. In other words, the AI ​​system 30 may chunk the character information to divide it into multiple sets of information and input each resulting set into the first AI model.

[0117] "A set of information sufficient to understand the order content" refers to the smallest meaningful unit of text extracted from order information, which the first AI model divides to efficiently and accurately understand and judge the order content. This "set" is not merely a string of characters, but a unit that completely contains one specific order element and its related information.

[0118] For example, if order information is in a tabular format such as a CSV file, an entire row is a typical example of "a set of information sufficient to understand the order details." That is, if a row in a CSV file contains order details for a specific product (e.g., "A123, Notebook PC, 5, 100000, 2023 / 07 / 10") in the format of "product code, product name, quantity, unit price, delivery date," then this entire row constitutes a set. This set contains textual information that is essential or useful for identifying the order for that product, such as product name, quantity, unit price, and delivery date.

[0119] Furthermore, for example, if the order information is image data related to non-standard documents such as faxes or PDF files, a line of details and item blocks containing information about a single product may constitute a "collection of information sufficient to understand the order contents." Item blocks include, for example, a customer information block containing the customer's address and contact information, and a payment terms block containing payment terms. For example, if a handwritten order form says "10 apples @ 100 yen," this "10 apples @ 100 yen" constitutes a single collection of information. In this case, the AI-OCR function plays a role in identifying collections of information not only from the text information but also from the layout.

[0120] For example, when accepting orders via email, paragraphs with specific meanings, bulleted list items, etc., within the email may constitute "a set of information sufficient to understand the order details." For instance, if the email states, "I would like to order the following items: 3 units of item X and 1 set of item Y," then "3 units of item X" and "1 set of item Y" are chunked as separate sets.

[0121] Thus, according to the first modification, the first AI model can process information by focusing on individual lines, item blocks, list items, etc., rather than interpreting the entire text information at once. This allows the AI ​​system 30 to efficiently and accurately extract information from various forms of order information and make accurate order decisions. As a result, the processing load on the first AI model is reduced and the accuracy of the order information is improved.

[0122] [8-2. Second variation: Automatic generation of judgment criteria] In the embodiments described above, the decision criteria were set by a person in charge via a user interface, but this disclosure is not limited thereto. This second modification relates to a configuration in which another external AI model automatically generates the decision criteria to be trained on the first AI model.

[0123] In other words, the server 20 may acquire at least one of past order information and past order information and input it into the second AI model, and have the second AI model output decision criteria. The server 20 may further train the first AI model with the decision criteria output from the second AI model.

[0124] The second AI model is responsible for automatically generating or extracting decision criteria for the first AI model to process orders. By using the second AI model, it becomes possible to automate or assist the task of setting decision criteria by the person in charge. The second AI model may be provided in, for example, AI system 30. Alternatively, for example, the second AI model may be provided in an AI system (not shown) different from AI system 30, or in server 20.

[0125] The second AI model acquires at least one of past order information and past order processing information as training data and analyzes them. Specifically, the second AI model learns the relationships between past order data and the corresponding final order processing results (e.g., confirmed information such as product code, quantity, unit price, and delivery date). Through this learning, the second AI model derives specific judgment rules and pattern trends that are necessary for the first AI model to output appropriate order information when new order information is input, and outputs them as judgment criteria.

[0126] The second AI model, like the first AI model, can utilize generative AI such as LLM. LLM excels at recognizing complex patterns and inferring rules from large amounts of text data, making it suitable for automatically discovering and generating decision criteria from historical unstructured order data. However, the second AI model is not limited to LLM; it could be any other machine learning model specialized in data analysis or rule mining. For example, models skilled at extracting specific correlations or regularities from historical structured data could be considered, such as decision trees, random forests, support vector machines, or neural network models optimized for specific pattern recognition.

[0127] The first AI model and the second AI model are, for example, independent AI models. However, they may be different instances running on the same physical hardware, or they may be completely different types of models. For example, it is possible to make the same underlying LLM function as the first and second AI models by fine-tuning it with different data for each task (order information generation and judgment criterion generation) or by operating it with different prompt sets. Alternatively, the first AI model could be an LLM, while the second AI model is a non-generative AI model specialized in trend analysis of historical data.

[0128] Specifically, for example, the order processing module 2034 acquires at least one of past order information and past order information. There are no particular limitations on how this information is acquired; for example, the order processing module 2034 may acquire it by reading at least one of past order information and past order information from the order information table 2024. Alternatively, for example, the order processing module 2034 may acquire it by receiving at least one of past order information and past order information transmitted from the terminal device 10 by user operation. For example, the order processing module 2034 transmits at least one of the acquired past order information and past order information to the AI ​​system 30.

[0129] The AI ​​system 30 inputs, for example, past order information and past order processing information received from the server 20 into the second AI model. The second AI model learns patterns and / or rules useful as decision criteria in order processing from the input information, and generates and outputs decision criteria based on the learning results. For example, the second AI model analyzes the relationship between past order data and actual order results and derives a decision criterion such as, "For this product name from this customer, it should be interpreted in this way."

[0130] The AI ​​system 30 tunes the first AI model by inputting the judgment criteria output from the second AI model along with the judgment items into the first AI model and training the first AI model. As a result, the server 20 tunes the first AI model via the second AI model and the AI ​​system 30.

[0131] According to this second modification, it becomes possible to automatically update and adjust the judgment criteria without human intervention, thereby improving operational efficiency. Furthermore, by continuously applying the latest judgment criteria that are more in line with actual conditions to the first AI model, the accuracy of order information generation can be continuously improved.

[0132] [8-3. Third Variation: Evaluation of Order Information and Proposed Revisions] In the embodiments described above, one process was described as being completed upon outputting order information, but this disclosure is not limited thereto. This third modification relates to an embodiment that incorporates an evaluation process for order information output by the first AI model and proposes modifications to the order information based on the evaluation results.

[0133] In other words, server 20 may accept an evaluation of the order information output from the first AI model. If the received evaluation indicates that the order information does not meet predetermined evaluation criteria, server 20 may have the first AI model propose a correction to the order information.

[0134] Specifically, for example, the order processing module 2034 accepts evaluations of the order information output from the first AI model. This "evaluation of order information" includes both evaluations by humans and evaluations by other AI models different from the first AI model.

[0135] In the case of evaluation by a person, for example, the person in charge checks the outputted order information and inputs the evaluation result into the terminal device 10, indicating whether or not the content meets the predetermined evaluation criteria (e.g., accuracy, completeness, consistency of format, etc.). The terminal device 10 then transmits the input evaluation result to the server 20. The order processing module 2034 then receives the evaluation result transmitted from the terminal device 10.

[0136] In the case of evaluation by a different AI model, for example, a third AI model provided in AI system 30 automatically analyzes the output order information and performs an evaluation based on pre-set evaluation criteria (e.g., consistency check by matching with various master data, validity check by comparing with similar past order information, syntax check, check for missing or missing required items, etc.). Note that the third AI model may be provided in an AI system not shown, or in server 20, that is different from AI system 30.

[0137] The third AI model is a machine learning model specifically designed for evaluation, distinct from the first AI model. Its role is to automatically evaluate the quality of the order information output by the first AI model. The third AI model could be, for example, an anomaly detection model, a classification model, or a model that efficiently executes rule-based verification logic, or it could be a generative AI model such as an LLM.

[0138] The evaluation results from the third AI model are output in the following format: evaluation score, identification of problems (e.g., mismatched product name, quantity outside range, etc.), and a flag indicating whether correction is necessary. The output evaluation results are sent to, for example, the server 20. The order processing module 2034 receives, for example, the evaluation results output from the third AI model. Such evaluation by the third AI model contributes to improving the efficiency and objectivity of the entire evaluation process by complementing or partially replacing manual evaluations performed by humans.

[0139] If the received evaluation result indicates that it does not meet the predetermined evaluation criteria (e.g., an evaluation by a human stating "correction is needed," or an evaluation by the third AI model stating "the evaluation score falls below the threshold"), the order processing module 2034 will, for example, feed the evaluation result back to the AI ​​system 30. The AI ​​system 30 will, for example, input the evaluation result into the first AI model and then prompt the first AI model to suggest corrections to the order information. More specifically, the AI ​​system 30 will, for example, provide the first AI model with the original input information and the evaluation result as feedback and instruct (by prompt) it to generate new order information that resolves the problems.

[0140] There are no particular limitations on the content of the proposals output by the first AI model, and a variety of content and formats are expected. For example, the proposals may include new order information that has been modified to meet the evaluation criteria. In addition to the new order information after modification, supplementary information that explicitly shows which items have been modified and how may also be included in the proposals. This will help the person in charge to quickly confirm and approve the modifications. In addition to the new order information after modification, the proposals may also include information that explains why these modifications were made and the rationale behind them. Alternatively, if the evaluation results indicate ambiguity and / or the possibility of multiple interpretations, the first AI model may output multiple proposed modifications to the order information.

[0141] By establishing a feedback loop for evaluating and proposing revisions to order information, the first AI model can continuously improve the quality of its output. This further enhances the accuracy of order information, reduces manual revision work, and ultimately leads to greater efficiency and quality improvement across the entire operation.

[0142] <9. Basic Computer Hardware Configuration> Figure 11 is a block diagram showing the basic hardware configuration of computer 90. Computer 90 comprises at least a processor 901, main memory 902, auxiliary storage 903, and a communication interface IF991. These are electrically connected to each other by a communication bus 921.

[0143] The processor 901 is hardware for executing the instruction set written in a program. The processor 901 consists of an arithmetic unit, registers, peripheral circuits, etc.

[0144] Main memory 902 is used to temporarily store programs and data processed by programs, etc. For example, it is a volatile memory such as DRAM (Dynamic Random Access Memory).

[0145] Auxiliary storage device 903 refers to a storage device for saving data and programs. Examples include flash memory, HDD (Hard Disc Drive), magneto-optical disk, CD-ROM, DVD-ROM, and semiconductor memory.

[0146] The IF991 communication interface is an interface for inputting and outputting signals for communication with other computers via a network using wired or wireless communication standards.

[0147] A network consists of various mobile communication systems, such as the internet, LANs, and wireless base stations. For example, a network includes 3G, 4G, and 5G mobile communication systems, LTE (Long Term Evolution), and wireless networks that can connect to the internet via designated access points (e.g., Wi-Fi®). When connecting wirelessly, communication protocols include, for example, Z-Wave®, ZigBee®, and Bluetooth®. When connecting via a wired connection, the network also includes connections made directly via USB (Universal Serial Bus) cables, etc.

[0148] Furthermore, by distributing all or part of each hardware configuration across multiple computers 90 and connecting them to each other via a network, a computer 90 can be virtually realized. Thus, the concept of computer 90 includes not only a computer 90 housed in a single enclosure or case, but also a virtualized computer system.

[0149] <10. Basic Functional Configuration of Computer 90> The functional configuration of the computer realized by the basic hardware configuration of computer 90 (Figure 11) is described below. The computer comprises at least one functional unit: a control unit, a memory unit, and a communication unit.

[0150] Furthermore, the functional units of computer 90 can also be realized by distributing all or part of each functional unit across multiple computers 90 interconnected via a network. The concept of computer 90 includes not only a single computer 90 but also a virtualized computer system.

[0151] The control unit is realized when the processor 901 reads various programs stored in the auxiliary storage device 903, loads them into the main memory device 902, and executes processing according to those programs. The control unit can realize various functional units that perform information processing depending on the type of program. In this way, the computer is realized as an information processing device that performs information processing.

[0152] The memory unit is implemented by the main memory 902 and the auxiliary memory 903. The memory unit stores data, various programs, and various databases. The processor 901 can also reserve memory areas corresponding to the memory unit in the main memory 902 or the auxiliary memory 903 according to the program. The control unit can also cause the processor 901 to perform operations such as adding, updating, and deleting data stored in the memory unit according to the various programs.

[0153] A database, specifically a relational database, is used to manage and link together tabular data sets called masters, which are structurally defined by rows and columns. In a database, tables are called tables, masters are called masters, the columns of tables are called columns, and the rows of tables are called records. In a relational database, relationships can be established and linked between tables and masters.

[0154] Typically, each table and master has a primary key column to uniquely identify records, but setting a primary key column is not mandatory. The control unit can instruct the processor 901 to add, delete, or update records in specific tables and masters stored in the memory unit, according to various programs.

[0155] Furthermore, by storing data, various programs, and various databases in the memory unit, it can be considered that an information processing device or information processing system according to one aspect of this disclosure has been manufactured.

[0156] Furthermore, the databases and masters in this disclosure may include any data structures (lists, dictionaries, associative arrays, objects, etc.) in which information is structurally defined. Data structures also include data that can be considered as data structures by combining data with functions, classes, methods, etc., written in any programming language.

[0157] The communication unit is implemented by the communication IF991. The communication unit provides the functionality to communicate with other computers 90 via the network. The communication unit can receive information transmitted from other computers 90 and input it to the control unit. The control unit can cause the processor 901 to perform information processing on the received information according to various programs. The communication unit can also transmit information output from the control unit to other computers 90.

[0158] Furthermore, each of the above-mentioned configurations, functions, processing units, processing means, etc., may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. The present invention can also be implemented by software program code that realizes the functions of the embodiment. In this case, a storage medium on which the program code is recorded is provided to a computer, and the processor of that computer reads the program code stored in the storage medium. In this case, the program code read from the storage medium itself realizes the functions of the embodiment described above, and the program code itself and the storage medium on which it is stored constitute the present invention. Examples of storage media used to supply such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs, optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, and the like.

[0159] Furthermore, the program code that implements the functions described in this embodiment can be implemented in a wide range of programming or scripting languages, such as assembler, C / C++, Perl, Shell, PHP, and Java (registered trademark).

[0160] Furthermore, the program code for the software that implements the functions of the embodiment may be distributed via a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the computer's processor may read and execute the program code stored in the storage means and the storage medium.

[0161] The functions realized by the components described herein may be implemented in a circuit or processing circuitry, including general-purpose processors, application-specific processors, integrated circuits, ASICs (Application Specific Integrated Circuits), CPUs (a Central Processing Unit), conventional circuits, and / or combinations thereof, programmed to realize the functions described herein. A processor is considered to be a circuit or processing circuitry, including transistors and other circuits. A processor may be a programmed processor that executes a program stored in memory.

[0162] In this specification, circuitry, unit, and means are hardware programmed to perform or execute the functions described herein. Such hardware may be any hardware disclosed herein, or any hardware known to be programmed to perform or execute the functions described herein.

[0163] If the hardware is a processor that is considered to be a type of circuitry, then the circuitry, means, or unit is a combination of hardware and software used to constitute the hardware and / or processor.

[0164] While several embodiments of this disclosure have been described above, these embodiments can be implemented in a variety of other forms, and various omissions, substitutions, and modifications are permitted without departing from the spirit of the invention. These embodiments and their variations are included within the scope and spirit of the invention, as well as within the scope of the claims and its equivalents.

[0165] <Note> The details described in each of the above embodiments are noted below.

[0166] [Note 1] A program for operating a computer comprising a processor and memory, the program causing the processor to perform the following steps: receiving a setting operation for a plurality of judgment items necessary for a first AI model to execute order processing via a user interface; training the first AI model with the received plurality of judgment items and tuning the first AI model to output order information relating to the order content based on the judgment results for each of the plurality of judgment items; receiving order information relating to the order content and extracting character information relating to predetermined characters from the order information; and inputting the extracted character information into the first AI model and causing the first AI model to output order information.

[0167] [Note 2] The program described in Appendix 1 accepts setting operations for each of multiple decision items via a user interface, and further trains the first AI model with the accepted decision criteria in the tuning step.

[0168] [Note 3] The criteria for judgment are the program described in Appendix 2, which includes common criteria that apply to all customers and individual criteria that apply to each customer individually.

[0169] [Note 4] A program described in any of the appendices 1 to 3, which, in the step of outputting order information, chunks the extracted predetermined character information into sets of information sufficient to grasp the order content, and inputs the chunked sets of information as a single unit into the first AI model.

[0170] [Note 5] A program described in any of the appendices 1 to 4, which causes the processor to further execute the steps of receiving an evaluation of the output order information, and, if the received evaluation indicates that it does not meet the predetermined evaluation criteria, having the first AI model propose a correction to the order information.

[0171] [Note 6] A program described in any of the appendices 1 to 5, which further has a processor perform the steps of: obtaining at least one of past order information and past order information; inputting at least one of the obtained past order information and past order information into a second AI model and having the second AI model output judgment criteria corresponding to each of multiple judgment items, and then tuning the program, wherein the output judgment criteria are further trained on the first AI model.

[0172] [Note 7] A method to be performed on a computer having a processor and memory, wherein the processor performs all steps in any of the programs described in Appendix 1 to Appendix 6.

[0173] [Note 8] An information processing device comprising a control unit and a storage unit, wherein the control unit executes all steps in any of the programs described in Appendix 1 to Appendix 6.

[0174] [Note 9] A system comprising one or more processors that execute all steps in any of the programs described in Appendix 1 to Appendix 6. [Explanation of Symbols]

[0175] 1... System 10…Terminal device 12…Communication IF 13…Input device 14…Output device 15…Memory 16…Storage 19… Processor 20... Server 22...Communication IF 23…Input device 24…Output device 25…Memory 2 hours… storage 29… Processor 30…AI system 80…Network

Claims

1. A program for operating a computer that includes a processor and memory, The program is provided to the processor: The first AI model receives, via a user interface, a set of multiple decision items for determining the content to be output as order information regarding the order details, which is necessary when the first AI model performs order processing, and a set of decision criteria for determining the content to be output as order information corresponding to each of the multiple decision items. The steps include: training the first AI model with the received multiple judgment items and multiple judgment criteria, tuning the first AI model to perform judgments for each of the multiple judgment items, and outputting the order information based on the results of those judgments; The steps include receiving order information regarding the order details and extracting character information related to a predetermined character from the said order information, A program that performs the steps of inputting the extracted character information into the first AI model, causing the first AI model to make the determination of whether or not a predetermined character contained in the character information satisfies the determination criteria, and outputting the order information from the first AI model.

2. The program according to claim 1, wherein the criteria for judgment include common criteria that apply to all customers and individual criteria that apply individually to each customer.

3. The program according to claim 1, wherein in the step of outputting the order information, the extracted predetermined character information is chunked into sets of information that allow the order contents to be understood, and the chunked sets of information are input to the first AI model as a single unit.

4. A step of receiving an evaluation of the outputted order information, The program according to claim 1, which, if the received evaluation indicates that it does not meet predetermined evaluation criteria, causes the processor to further perform the step of providing the first AI model with the text information and the evaluation as feedback and instructing it to generate new order information that meets the evaluation criteria, thereby causing the new order information to be output as a revised version of the order information.

5. Steps include obtaining past order information and past order information, The processor is further made to perform the following steps: input the acquired past order information and past order information into a second AI model, analyze the relationship between the past order information and the past order information, and output judgment criteria corresponding to each of the multiple judgment items from the second AI model based on the analysis results. The program according to claim 1, wherein in the tuning step, the output judgment criteria are further trained on the first AI model.

6. A method to be performed on a computer comprising a processor and memory, wherein the processor performs all steps of a program according to any one of claims 1 to 5.

7. An information processing device comprising a control unit and a storage unit, wherein the control unit executes all steps in the program described in any one of claims 1 to 5.

8. A system comprising one or more processors that perform all steps in the program described in any one of claims 1 to 5.

Images