Order placement / acceptance support device

Abstract

To provide an order placement / acceptance support device capable of facilitating order placement / acceptance activities by sufficient support in an entire order placement / acceptance process.SOLUTION: An order placement / acceptance support device 1 includes: a large-scale language model unit 113 having a large-scale language model specialized in creating order placement / acceptance support information by being trained and adjusted so as to stochastically predict how words or sentences provided in a prompt are likely to occur as natural language on the basis of order placement / acceptance related information of an order placement / acceptance database 121 and so as to generate order placement / acceptance support information; an information reception unit 111 for receiving response information of a user 4 in a conversation with the user 4; an initial setting prompt generation unit 114 that instructs a prerequisite for generating order placement / acceptance information in the large-scale language model; and an order placement / acceptance prompt generation unit 115 that causes the large-scale language model to generate order placement / acceptance support information suited to an order placement / acceptance stage on the basis of the prerequisite.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an order placement and receipt support device. [Background technology]

[0002] In the field of supporting ordering and receiving, active efforts are being made to enable ordering parties to be provided with the most suitable suppliers.

[0003] For example, Patent Document 1 discloses a technology for determining the construction content and necessary products for a building based on the operation of a user who is the client, and selecting candidate contractors who can undertake the construction. Specifically, the technology discloses a construction content determination means and a product determination means that determine the construction content and products based on the operation of the user, a candidate contractor determination means that selects contractors who can undertake the construction based on the content, and a candidate contractor that will be the final order recipient is determined based on the operation of the user.

[0004] As a result, according to the technology of Patent Document 1, each means such as the construction content determination means determines candidate contractors to order from and allows the user to make a selection, making it easy for the user to select a client. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2024-31943 Summary of the Invention [Problem to be solved by the invention]

[0006] Incidentally, the technology of Patent Document 1 allows the user to select from candidate contractors by using each means such as the construction content determination means, but there is room for further improvement in that it does not provide sufficient support for the entire ordering process from the user's acceptance of the task to placing an order with a contractor.

[0007] The present invention has been made in view of the above circumstances, and its object is to provide an order placement and receipt support device that enables easy order placement and receipt activities by providing sufficient support throughout the entire order placement and receipt process. [Means for solving the problem]

[0008] As a result of extensive research into solving the above problems, the inventors discovered that the above object can be achieved by utilizing a large-scale language model specialized for creating order support information, and by conducting a conversation with a user following the order placement stages from the user's acceptance of the task to placing an order with a vendor. The inventors then completed the present invention. Specifically, the present invention provides the following:

[0009] The present invention includes: an order database that stores order-related information related to the order process from receiving a user's assignment to placing an order with a vendor; an initial setting storage unit storing prerequisite information instructing the conversation with the user to be conducted in accordance with the order placement and receipt stages indicating the respective stages of the order placement and receipt process; an information receiving unit that receives response information of the user in a conversation with the user; a large-scale language model unit that has a large-scale language model specialized for creating the order support information by combining one or more types of natural language processing based on the order-related information in the order database, and probabilistically predicting how likely words and sentences given in the prompts are to occur in natural language, and by performing learning and adjustment to generate order support information to be presented to the user in the order process by sentence generation and question answering; an initial setting prompt generation unit that, at the start of the order placement and receipt process, includes the prerequisite information of the initial setting storage unit in the prompt of the large-scale language model, thereby instructing that a conversation in accordance with the order placement and receipt stage be held with the user as a prerequisite for generating the order placement and receipt support information in the large-scale language model; an order prompt generation unit that causes the large-scale language model to generate the order support information suitable for the order stage based on the prerequisites by including the response information received by the information receiving unit after the start of the order process in the prompt of the large-scale language model unit; The ordering support device has the above.

[0010] According to the present invention, at the start of the order process, a large-scale language model specialized for creating order support information is initialized to conduct a conversation according to the order stage, and response information from the conversation with the user after the order process has started is included in the prompts, thereby making it possible to fully support order activities such as finding a supplier just like having a conversation with a human concierge, that is, by simply conducting simple interviews and natural language exchanges without the user having to write prompts to input into the large-scale language model. This makes it possible to easily perform order activities with sufficient support throughout the entire order process provided by a large-scale language model specialized for creating order support information. [Effects of the Invention]

[0011] According to the present invention, ordering activities can be easily carried out by providing sufficient support throughout the entire ordering process using a large-scale language model specialized for creating ordering support information. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is an explanatory diagram showing the flow of information in the order placement and receipt support device of this embodiment. [Figure 2] FIG. 2 is an explanatory diagram showing the flow of information in the order placement and receipt support device of this embodiment. [Figure 3] FIG. 3 is a block diagram showing an example of the hardware configuration of the order placement and receipt support device of this embodiment. [Figure 4] FIG. 4 is an explanatory diagram showing a display screen of a user terminal. [Figure 5] FIG. 5 is an explanatory diagram showing a display screen of a user terminal. [Figure 6] FIG. 6 is an explanatory diagram of the user database. [Figure 7] FIG. 7 is an explanatory diagram of the vendor database. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings.

[0014] (Ordering support device 1) As shown in Figure 1, the order support device 1 is configured to utilize a large-scale language model specialized for creating order support information, and to have a conversation with user 4 following the order stages from accepting user 4's assignment to placing an order with the vendor, and is configured to include an order database 121 that stores order-related information, an initial setting memory unit 16 that stores prerequisite information, an information receiving unit 111 that receives response information from user 4, a large-scale language model unit 113 that has a large-scale language model specialized for generating order support information, an initial setting prompt generation unit 114 that includes prerequisite information in prompts, and an order prompt generation unit 115 that includes response information in prompts.

[0015] (Ordering support device 1: ordering database 121) More specifically, the ordering support device 1 has an ordering database 121 that stores ordering-related information relating to the ordering process from receiving an assignment from the user 4 to placing an order with a vendor.

[0016] Here, "User 4" refers to an individual or company employee who places an order for a specific product or service. For example, this includes purchasing personnel in the retail industry, material procurement personnel in the manufacturing industry, outsourcing management personnel in the service industry, as well as managers of small and medium-sized enterprises, freelancers, members of start-up companies, and sales personnel.

[0017] "Challenges" refer to the problems and difficulties that User 4 is currently facing, and include various elements of challenges such as quality assurance, delivery date management, cost reduction, detailed specifications, supplier management, preventing ordering errors, negotiating contract terms, and inventory management. Specifically, when User 4 needs a specific product or service, the challenges include finding a vendor that best meets those requirements and negotiating with a vendor that meets conditions such as cost and delivery date.

[0018] "Vendor" refers to a company or individual that provides goods or services, and refers to providers in various industries and fields who can supply the goods or services that User 4 wants to order. Examples of vendors include system development companies, web production companies, design companies, consulting companies, translation companies, logistics companies, printing companies, and freelancers.

[0019] The "order process" is a series of steps from receiving a problem from user 4 to actually placing an order with a vendor, and includes multiple order stages that represent each stage of the order process. Specifically, the issue is first received from user 4 (issue stage). Next, an idea for solving the issue is formed (idea formation stage), and the decision on whether outsourcing or in-house production is appropriate is supported (outsourcing or in-house production stage). If outsourcing is determined to be appropriate, candidate vendors are researched (candidate vendor research stage), and support is provided in selecting a vendor that matches user 4's requirements (matching stage). This series of steps includes supporting business negotiations between the selected vendor and user 4 (negotiation stage), and supporting user 4 in placing an order with the vendor (order stage). Note that the order process may also include a payment stage after the order stage, in which user 4 pays the vendor, and an evaluation stage, in which user 4 evaluates the vendor after the transaction is completed.

[0020] The "Problem Stage" is the initial stage in which user 4 clarifies the problems they face and the solutions they need. Through this problem stage, the ordering support system 1 accurately understands user 4's needs and uses this understanding as a basis for appropriate idea generation and vendor selection. For example, if user 4 is a restaurant owner, the problem stage might involve presenting a problem such as "We need to procure kitchen equipment and interior design for the opening of a new restaurant." Or, if user 4 is an IT company representative, the problem might involve presenting a problem such as "We need to implement new software to strengthen system security." There are two types of problems: explicit and latent. For example, during a management networking event, an acquaintance might make a realization, such as, "I heard about a software-as-a-service (SaaS) product that was a success." Based on these problems, large-scale language models can be used not only to generate ideas for explicit orders (e.g., relocating an office because the number of employees has increased and it's becoming too small) but also to generate ideas for latent ordering possibilities (e.g., introducing groupware to revitalize the organization and increase employee retention).

[0021] The "idea formation stage" is a stage in which user 4 comes up with specific ideas and solutions to solve the problems they face, and considers what kind of products or services are needed or what approach is appropriate for solving their problems. For example, if user 4 is in the manufacturing industry and is concerned about improving the efficiency of their production line, in the idea formation stage, they will consider ideas such as introducing automated machinery, redesigning the production process, and improving the quality control system. Then, the order placement and receipt support device 1 supports the idea formation process, enabling user 4 to find the appropriate direction for efficiently solving their problems.

[0022] The "Outsourcing or In-House Stage" is the stage where user 4 decides whether to produce in-house (in-house) or outsource to an external vendor (outsourcing) to realize a solution. User 4 is supported in making the optimal choice by considering various factors, such as cost, quality, delivery time, and resource availability. For example, if user 4 is an apparel company considering producing a new clothing collection, the outsourcing or in-house stage will determine whether to produce in-house or outsource to an external manufacturer, taking into account factors such as the company's production capacity, potential for cost reduction, ease of quality control, and strict delivery deadlines. Similarly, an IT services company may consider whether to develop a new system in-house or outsource it to an external development company. In such cases, the order placement and purchase support device 1 provides information and advice to user 4 throughout the outsourcing or in-house stage to help user 4 make the appropriate decision, allowing user 4 to determine a strategy that most effectively utilizes their company's resources and goals.

[0023] The "candidate vendor research stage" is a stage in which, if outsourcing is determined to be appropriate, user 4 searches for and evaluates vendors that may meet user 4's requirements. A list of vendors that can provide appropriate products and services according to user 4's needs is created, and each vendor's reliability, pricing, quality standards, delivery track record, etc. are investigated. For example, if user 4 is a construction company and needs to procure building materials for a new construction project, the candidate vendor research stage creates a list of vendors that can supply building materials, and evaluates the quality, cost efficiency, delivery reliability, etc. of the materials provided by each vendor. In such cases, the order placement and receipt support device 1 supports user 4 in efficiently collecting and evaluating information on candidate vendors through the candidate vendor research stage, so that user 4 can collect information to narrow down the appropriate candidates for final vendor selection.

[0024] The "matching stage" is a stage in which the vendor that best meets the specific requirements of user 4 is selected based on the information collected in the candidate vendor research stage. The optimal match is achieved by comparing user 4's needs with the characteristics of the products and services offered by the candidate vendors. The vendor selection process prioritizes vendors registered in the vendor database 1212 shown in Figure 2. If a suitable vendor is not registered, unregistered vendors are searched for via the Internet 5. For example, if user 4 is a hotel manager ordering furniture and interior items for guest room renovations, the matching stage selects the optimal furniture supply vendor based on requirements such as style, quality, price, and delivery time. In such cases, the ordering support device 1 efficiently identifies the vendor that best meets user 4's needs and assists user 4 in proceeding to a contract with an appropriate vendor through the matching stage, enabling the final selection.

[0025] Furthermore, when selecting a vendor, the large-scale language model unit 113 analyzes order-related information, such as company profiles, order details, and order history, of the user 4 (the orderer) and the vendor (the recipient), stored in the order database 121, using natural language processing. Using natural language processing technology, the large-scale language model unit 113 understands the user 4's response information and order-related information, compares them with each other's needs and capabilities, and identifies the characteristics of the vendor that best suits the order details. Based on this, a candidate list is generated. Furthermore, past success stories, industry trends, and market data are analyzed to improve the matching accuracy between the user 4 and the vendor. This not only ensures that requirements are matched, but also allows for the formation of an optimal partnership from a strategic perspective. The candidate list thus created is then presented to the user 4. It is preferable that the candidate list also includes a detailed profile of each candidate, along with the reasons for their selection. It is also preferable that the candidate list also includes a report on any unnecessary cost factors for each candidate.

[0026] The "negotiation stage" is a stage in which specific transaction terms are negotiated between the vendor selected in the matching stage and user 4, with the aim of reaching an agreement on detailed contract terms such as price, delivery date, payment terms, quality guarantee, and after-sales service. For example, if user 4 is a restaurant owner and signs a contract with a vendor to receive a regular supply of ingredients, in the negotiation stage, user 4 negotiates with the vendor about the quality standards of the ingredients, pricing, delivery schedule, emergency response measures, and the like. In such a case, the ordering support device 1 supports user 4 in conducting effective negotiations and concluding an appropriate contract through the negotiation stage, so that optimal contract terms can be determined for both user 4 and the vendor, and a mutually satisfactory agreement can be reached.

[0027] The "ordering stage" is the stage where user 4 officially places an order with a vendor based on the conditions agreed upon in the negotiation stage, and involves creating a purchase order, signing a contract, and final confirmation of the order contents. For example, if user 4 is a retail store manager and wants to purchase new seasonal products, in the ordering stage, user 4 sends a purchase order to the selected vendor that lists the product type, quantity, price, delivery date, etc. In such a case, the order placement and purchase support device 1 supports user 4 in smoothly completing procedures such as creating a purchase order and signing a contract throughout the ordering stage, so that user 4 can officially request the supply of goods or services based on the conditions agreed upon between user 4 and the vendor, thereby completing the ordering process.

[0028] "Order-related information" refers to information exchanged between User 4 and vendors during the order-receiving process, as well as information necessary for managing and executing the order-receiving process. It encompasses a variety of information necessary for the smooth flow of the order-receiving process. Specifically, this information includes user information (basic information about User 4 and the company), vendor information (information about vendors), past order data, product and service specifications, pricing information, inventory information, delivery date information, and payment terms information. Order-related information also includes business rules, regulations, manuals, and templates related to order-receiving. Examples include ordering procedures and approval flows, quotation and purchase order templates, and order-related contracts and regulations. Furthermore, order-related information also includes past analytical data and best practices for streamlining the order-receiving process. Specifically, this information includes lead time analysis, analysis of optimal inventory levels, and data on optimizing order timing.

[0029] "User information" refers to basic information about the user and information related to the transaction in the order placement and purchase process. Specifically, examples include the user's company name, such as the official name of the company, contact information such as the company's location and registered address, telephone number, fax number, and email address, the name of the person in charge of the order placement and purchase, company size such as the number of employees and amount of capital, the industry and business details to which the company belongs, past order placement and purchase performance and transaction history, credit information regarding the company's credibility and reliability, current issues and requests for specific transactions, and priority of contract terms that are emphasized in transaction terms and contracts.

[0030] "Vendor information" refers to information related to vendors involved in the order placement and purchase process. Examples of such information include the vendor's official name, the vendor's physical location or business address, the vendor's contact information (such as telephone number, fax number, and email address), the name of the primary contact person for the transaction, the types of products and services offered by the vendor, the vendor's industry classification, information indicating the vendor's size (such as the number of employees and annual sales), the year the vendor was founded, past transaction history and information about major business partners, whether the vendor has quality control certifications (such as ISO), information about the vendor's technology and expertise, past delivery deadline compliance rates and special notes regarding delivery dates, information about the vendor's credibility (such as payment history and credit rating), and catalogs and specifications for the products and services offered by the vendor. Vendor information may also include information about actual problems and how they were resolved.

[0031] The order database 121, which stores the above-mentioned order-related information, is provided in the order storage unit 12. The order storage unit 12 is built on a group of servers with sufficient performance to quickly process and access large amounts of data in order to efficiently operate a large-scale language model specialized for generating order support information. Furthermore, the order database 121 is built using a database management system to support efficient data management and access. Details of the order database 121 will be described later.

[0032] In the above-mentioned order placement and receipt process, it is preferable to use management theories such as SWOT analysis and 3C analysis to organize the issues and then use a large-scale language model to recommend candidate orders.

[0033] Here, "SWOT analysis" is an analytical method for identifying and evaluating the strengths, weaknesses, opportunities, and threats of an organization or project. Through this analysis, companies can understand their competitive advantage and gain useful insights when planning their strategy. "Strengths" refer to a company's competitive advantages, resources, capabilities, and other points where it excels compared to other companies. "Weaknesses" refer to elements that put it at a disadvantage in competition or areas that need improvement. "Opportunities" identify opportunities for growth and profit improvement available within the market or industry. "Threats" are external risks that hinder a company's growth, such as the competitive environment and regulations.

[0034] "3C analysis" is a method of analyzing three elements: company, customers, and competitors, and is useful for clarifying a company's position in the market, deepening understanding of customer needs, and formulating competitive strategies.

[0035] SWOT and 3C analyses are preferably conducted primarily at the early stages of the order placement process. Specifically, when accepting a problem from User 4 during the problem stage, SWOT and 3C analyses are used to understand the underlying factors of the problem and grasp the company's current situation. This makes it possible to more clearly grasp the essence of the problem and its impact.

[0036] In the ideation stage, SWOT and 3C analyses are used to determine which approaches are effective when considering solutions to problems. SWOT analysis helps form a strategy that takes into account the company's internal and external environments. 3C analysis, on the other hand, helps understand the company's position in the market, comparisons with competitors, and customer needs. These analyses then provide information for the outsourcing or in-house stage, where the options of outsourcing or in-house development are evaluated, and are used to inform decision-making in subsequent stages.

[0037] (Ordering support device 1: initial setting storage unit 16) The ordering support device 1 has an initial setting storage unit 16 that stores prerequisite information that instructs the device 1 to have a conversation with the user 4 according to the ordering stages that indicate the respective stages of the ordering process.

[0038] Here, "prerequisite information" refers to instruction information for appropriately advancing the conversation with User 4 at each stage of the order process. Specifically, it includes information on how to proceed with the conversation at each stage of the order process (issue stage, idea generation stage, outsourcing or in-house production stage, candidate vendor research stage, matching stage, business negotiation stage, and order stage), as well as the information gathering items and decision criteria required at each stage.

[0039] For example, the prerequisite information for the problem stage may include questions to clarify the problem faced by user 4, questions to analyze the causes and background of the problem, and questions to set goals for solving the problem. The prerequisite information for the idea generation stage may include questions to consider various solutions to solve the problem, questions to generate ideas using techniques such as brainstorming or mind mapping, and questions to analyze the advantages and disadvantages of each solution. The prerequisite information for the outsourcing or in-house development stage may include questions to analyze the advantages and disadvantages of outsourcing and questions to analyze the advantages and disadvantages of in-house development.

[0040] The prerequisite information for the candidate vendor research stage may include questions for collecting information about candidate vendors and questions for comparing and considering service content, fees, track record, etc. The prerequisite information for the matching stage may include questions for selecting a vendor that matches User 4's requirements and questions for contacting the vendor and arranging a business meeting. The prerequisite information for the business negotiation stage may include questions for supporting business negotiations between the selected vendor and User 4, questions for confirming contract details, negotiating quotes, etc. Note that "quote negotiation" may include first receiving a "rough quote," then finalizing the terms through business negotiations, etc., to arrive at a "formal quote," and then comparing the service content and quote amounts of multiple vendors to make a decision. The prerequisite information for the ordering stage may include questions for supporting User 4 in placing an order with the vendor, and questions for supporting the creation and sending of a purchase order, etc.

[0041] The prerequisite information also includes conditions and instructions for terminating the current order stage and moving on to the next order stage, in order to have a conversation with User 4 according to the order stages, which indicate each stage of the order process. Specifically, the prerequisite information includes questions to confirm that the purpose of the current stage has been achieved, such as "Have you fully understood the details of the issue?" and "Has all the necessary information been provided?", as confirmation of the transition conditions; instructions to encourage moving on to the next stage, such as "Next, let's form an idea for a solution" and "Let's gather information to decide whether to outsource or develop in-house," as guidance for the next stage; confirmation statements to obtain User 4's consent, such as "Are you sure you want to proceed to the next stage?", as confirmation and approval; and conditional statements to present the information and conditions required for the next stage, such as "The following information is required to move on to the idea generation stage," as instructions for the transition conditions.

[0042] The prerequisite information including these phrases will be included in the initial setting prompt that the large-scale language model first loads when the user 4 uses the order placement and receipt support device 1, and will provide the guidelines necessary to smoothly advance the order placement and receipt process through dialogue with the user 4. The prerequisite information also clearly indicates the transition to the next stage when the current stage is completed, and by providing specific instructions on the actions required in the next stage, it will be possible for the large-scale language model to smoothly advance the conversation with the user 4.

[0043] The first processing operation when the user 4 uses the order placement and receipt support device 1 is to automatically load an initial setting prompt into the large-scale language model. As a first automatic loading method, the initial setting storage unit 16 of the order placement and receipt support device 1 is configured as an initial setting server, and when the user 4 accesses the order placement and receipt support device 1, the initial setting prompt generation unit 114 of the order placement and receipt support device 1 sends a request to the initial setting server to obtain prerequisite information, and the initial setting server receives the request, reads out the prerequisite information, and loads it into the large-scale language model.

[0044] As a second automatic loading method, a JavaScript file describing the contents of the initial setting information is embedded in an HTML file so that it can be loaded on the client side (user terminal 2). When a user 4 accesses the web page of the order placement and receipt support device 1, the JavaScript file is automatically executed on the browser and the contents of the initial setting information are loaded. The loaded initial setting information is then sent to the large-scale language model and loaded into the large-scale language model. In this case, the initial setting storage unit 16 can be configured with a memory or storage.

[0045] (Ordering support device 1: Initial setting storage unit 16: Specific example of prompt for initial setting) An example of an initial setup prompt containing prerequisite information is as follows:

[0046] (Premise) You are an ordering agent AI that fully supports the ordering process. On behalf of User 4, you will act as a concierge, assisting and acting on behalf of User 4 at each stage of the ordering process, from identifying the issue to vendor selection, negotiations, and placing an order. At each stage, you will actively seek feedback from User 4 and adjust the process as appropriate based on that feedback to support the entire ordering process. Always be mindful of ensuring the smooth progression of the process and maximizing User 4's profits. Respond flexibly to changes in market conditions and User 4's needs and provide the optimal solution.

[0047] (Ordering process) The stages are: 1. Understanding the issues, 2. Supporting idea generation, 3. Supporting the decision of whether to outsource or manufacture in-house, 4. Researching and recommending potential vendors, 5. Matching with vendors and negotiating on your behalf, and 6. Creating ordering documents and placing the order on your behalf.

[0048] (Handling the order process) Please take the following measures at each stage. Stage 1 is to understand the awareness of the problem. Specifically, ask User 4 about the details and background of the problem and collect clear information. Analyze the essence of the problem and the direction of resolution. Once the problem has been understood, proceed to the next stage, idea formation. Stage 2 is to support idea formation. Specifically, propose the best idea that will lead to solving User 4's problem. Work with User 4 to select the best idea and obtain User 4's final approval. Once the idea has been decided, proceed to the next stage, deciding whether to outsource or produce in-house.

[0049] Stage 3 is support for deciding whether to outsource or in-house. Specifically, it makes a comprehensive assessment of User 4's situation and suggests whether outsourcing or in-house production is appropriate. If outsourcing is chosen, the system proceeds to the next stage, candidate vendor research. If in-house production is chosen, the order process ends. Stage 4 is candidate vendor research and recommendation. Specifically, it researches multiple optimal vendors based on User 4's requirements. It performs a detailed comparative evaluation of the vendors and proposes the final candidate to User 4. Once a candidate vendor has been decided, the system proceeds to the next stage, matching and negotiation.

[0050] Stage 5 is matching with vendors and negotiating on behalf of User 4. Specifically, the system adjusts terms and negotiates prices with the vendor on behalf of User 4. The system obtains User 4's approval for the final order terms. Once the order terms are approved, the system proceeds to the next stage of creating purchase order documents and placing an order. Stage 6 is creating purchase order documents and placing an order on behalf of User 4. Specifically, the system creates the documents required for placing an order (request for proposal, quotation, purchase order, etc.) based on the approved terms. The system places an order with the vendor on behalf of User 4, completing the process.

[0051] (Ordering support device 1: information receiving unit 111) The order placement and receipt support device 1 has an information receiving unit 111 that receives response information from the user 4 in a conversation with the user 4. Here, "response information" refers to an answer or response provided by the user 4 to the order placement and receipt support device 1, and is collected during the dialogue between the user 4 and the order placement and receipt support device 1 and used as data necessary for the progress of the order placement and receipt process. Examples of such information include information about the problem, such as the cause and background of the problem, methods for solving the problem, and the skills and experience required to solve the problem; information about solutions, such as the advantages and disadvantages of each solution, the feasibility and risks of each solution, and the costs and time required for each solution; and information about the vendors, such as a list of candidate vendors, the service content, fees, track record, and strengths and weaknesses of each vendor.

[0052] The information receiving unit 111 has an input function for receiving answer information input by the user 4 to the user terminal 2. In addition to the input function, the information receiving unit 111 also has a voice input function that enables the user 4 to input questions or requests by voice, a file attachment function that enables the user 4 to attach files related to the questions or requests, and a model transmission function that transmits the input questions or requests to the large-scale language model unit 113.

[0053] The information receiving unit 111 also has an activation function. The activation function enables data communication only with authenticated user terminals 2 and administrator terminals 7. The administrator terminal 7 is a terminal device operated by a person in charge of managing the order placement and receipt support device 1. The activation function verifies whether an activation signal sent from a specific user terminal 2 or administrator terminal 7 matches pre-set authentication information. Examples of authentication methods include authentication modes that combine one or more authentication elements, such as password authentication, biometric authentication, one-time password authentication, and smart card authentication using a card with an embedded IC chip. This allows the order placement and receipt support device 1 to communicate data only with authenticated user terminals 2 and administrator terminals 7, making it less likely for information to leak.

[0054] The activation function may be used for authentication with a specific vendor terminal 3. The activation function may also be provided between the control unit 11 and the servers of the order storage unit 12 and initial setting storage unit 16, allowing data communication only between the authenticated control unit 11 and the order storage unit 12 and initial setting storage unit 16. In this case, when the order storage unit 12 and initial setting storage unit 16 are installed in a location away from the control unit 11 and are accessible via data communication such as the Internet 5, unauthorized access from outside can be prevented.

[0055] (Ordering support device 1: large-scale language model unit 113) The order placement support device 1 includes a large-scale language model unit 113 having a large-scale language model specialized for generating order placement support information. The large-scale language model unit 113 combines one or more types of natural language processing based on order-related information in the order placement database 121 to probabilistically predict how likely words and sentences given in prompts are to appear in natural language. The large-scale language model unit 113 learns and adjusts to generate order placement support information to be presented to the user 4 during the order placement process through sentence generation and question answering. The large-scale language model unit 113 may be configured to provide the order placement support information to others via an API. By widely sharing the high-value order placement support information it generates with other businesses and systems, the efficiency of the order placement process can be expanded throughout the market. The large-scale language model unit 113 may also use a large-scale language model provided by another organization (such as a company or a university) via an API. Furthermore, the large-scale language model of the large-scale language model unit 113 may be linked to large-scale language models of other organizations via an API.

[0056] Here, "natural language processing" refers to a process that enables a computer to understand text and speech data written in natural language and execute processing appropriate to the purpose. Specifically, examples include morphological analysis, which breaks natural language down into "morphemes," the smallest units that make up the language, and assigns information such as parts of speech; syntactic analysis, which analyzes the grammatical structure of natural language to clarify the structure and meaning of a sentence; semantic analysis, which analyzes the meaning of natural language to understand the meaning of words and sentences and make logical judgments and inferences; contextual analysis, which understands natural language while taking into account the context before and after a sentence; and intent analysis, which extracts the intention of a speaker or writer from a conversation or sentence using natural language. Thus, "natural language processing" processes natural language by combining processes such as morphological analysis, syntactic analysis, semantic analysis, contextual analysis, and intent analysis, and enables the generation of order support information that supports the order-taking activities of this embodiment, as well as machine translation, automatic summarization, question-answering systems, and speech recognition.

[0057] A "prompt" is a word or sentence input to the large-scale language model unit 113, and serves as the starting point for the large-scale language model to generate order placement and receipt support information. The generation of order placement and receipt support information using a large-scale language model based on a prompt is a major difference from normal (conventional) machine learning. In normal machine learning, a model learns from training data and predicts output data for input data. For example, a machine learning model for image recognition learns from training data of images of cats and dogs and classifies the input image as either a cat or a dog. On the other hand, when generating order placement and receipt support information using a large-scale language model, the model learns not only from training data but also from prompts that provide instructions and information regarding the output data that the model is desired to generate.

[0058] In more detail, while conventional machine learning models can only generate content contained in the training data, large-scale language models can use prompts to generate new content not contained in the training data. For example, even if the training data only contains the conversations of female office workers in their 20s, prompts can also generate the conversations of male office workers in their 50s. Furthermore, while conventional machine learning models can only generate variations of the content contained in the training data, large-scale language models can use prompts to generate new variations of output data not contained in the training data. For example, even if the training data only contains conversations from the user's perspective, prompts can also generate conversations from the vendor's perspective. Furthermore, while conventional machine learning models can only generate new content by recursively combining content contained in the training data, large-scale language models can use prompts to generate creative content not contained in the training data. For example, even if the training data only contains explanations of existing problem-solving methods, prompts can generate problem-solving ideas from completely new perspectives.

[0059] A "large-scale language model" is a type of probabilistic model used in natural language processing, which is a model for probabilistically predicting how likely a given word or sentence is to occur in natural language. Specifically, a language model calculates the occurrence probability of a given word sequence or sentence, or compares the occurrence probabilities of multiple word sequences or sentences, making it possible to automatically generate the most likely word or sentence based on the context when predicting the next word or sentence, or to generate a sentence that meets specific conditions.

[0060] Furthermore, the large-scale language model is specialized for creating order placement and receipt support information. As a result, the large-scale language model unit 113 can improve generation accuracy and efficiency by specializing in creating order placement and receipt support information using natural language processing based on the order placement and receipt related information in the order placement and receipt database 121. Specifically, by focusing on learning vocabulary, grammar, example sentences, templates, etc. related to the order placement and receipt related information, more realistic and specific order placement and receipt support information can be generated, and by optimizing the generation process, it can be generated in a shorter time.

[0061] (Ordering support device 1: Initial setting prompt generation unit 114) The order placement and receipt support device 1 has an initial setting prompt generation unit 114 that instructs as a precondition for generating order placement and receipt support information. The initial setting prompt generation unit 114 has a function of including precondition information from the initial setting storage unit 16 in the prompt of the large-scale language model at the start of the order placement and receipt process, thereby instructing as a precondition for generating order placement and receipt support information in the large-scale language model that a conversation in accordance with the order placement and receipt stages be held with the user 4.

[0062] It is preferable that the initial setting prompt generation unit 114 be implemented as a separate functional module from the large-scale language model unit 113 equipped with the large-scale language model. The reason for this is that the initial setting prompt generation unit 114 has the role of generating appropriate prompts based on the order receiving process, and the large-scale language model unit 113 has the role of performing natural language processing. By making these independent modules, the roles of each unit are clarified, facilitating system design and maintenance. Furthermore, by implementing the initial setting prompt generation unit 114 as an independent module, it is possible to flexibly respond to changes in the order receiving process or updates to the large-scale language model.

[0063] The initial setting prompt generation unit 114 has an information reception interface for receiving prerequisite information from the initial setting memory unit 16, a prompt generation logic for generating an initial setting prompt corresponding to each stage of the order placement and receipt process based on the received prerequisite information, and a prompt transmission interface for transmitting the generated initial setting prompt to the large-scale language model unit 113.

[0064] The initial setup prompt generation unit 114 may be constructed using a rule-based logic circuit or a machine learning model. In the rule-based case, the initial setup prompt is generated using explicit conditional branching and transition rules. On the other hand, in the case of a machine learning model, more flexible prompt generation is possible using a model learned from data on the order placement and receipt process. With this configuration, the initial setup prompt generation unit 114 generates an appropriate initial setup prompt at the start of the order placement and receipt process and transmits it to the large-scale language model unit 113, thereby enabling the effective start of a conversation appropriate to the order placement and receipt process.

[0065] Furthermore, the initial setting prompt generation unit 114 may be implemented using a script or program that dynamically generates an initial setting prompt based on the control logic of the order placement and receipt process, written in a scripting language or a programming language. For example, prerequisite information may be acquired from the initial setting storage unit 16, and JavaScript code for generating an initial setting prompt based on the prerequisite information may be incorporated into an HTML file. When the user 4 accesses the web page of the order placement and receipt support device, the script may be executed on the browser. The script may acquire prerequisite information from the initial setting storage unit 16, generate an initial setting prompt, send it to the large-scale language model unit 113, and generate order placement and receipt support information in the large-scale language model. In this case, the initial setting prompt generation unit 114 exists as a functional module that spans both the client side (JavaScript) and the server side (initial setting storage unit 16).

[0066] (Ordering support device 1: ordering prompt generation unit 115) The order placement and purchase support device 1 has an order placement and purchase prompt generation unit 115 that causes the large-scale language model unit 113 to generate order placement and purchase support information that is suitable for the order placement and purchase stage based on the prerequisites by including the response information received by the information receiving unit 111 after the start of the order placement and purchase process in the prompt of the large-scale language model unit 113.

[0067] It is preferable that the order prompt generation unit 115 exists as a module independent of the large-scale language model unit 113. The reason for this is that, like the initial setting prompt generation unit 114, the order prompt generation unit 115 has the role of generating appropriate prompts based on response information from the user 4, which is different from the natural language processing role of the large-scale language model unit 113. For this reason, treating the two as independent modules makes it possible to clearly distinguish their respective roles and facilitate system design and maintenance. Furthermore, implementing the order prompt generation unit 115 as an independent module allows for flexible response to changes in the order process or updates to the large-scale language model.

[0068] The order prompt generation unit 115 has an interface for accepting response information received from the information receiving unit 111, and prompt generation logic for generating a prompt appropriate for the current stage of the order process based on the received response information. The prompt generation logic analyzes the response information to understand the current situation in the order process and the intention of the user 4, and selects an appropriate prompt template based on that understanding. Conditional branching is used for this selection, and a prompt template that matches the response information is selected. The prompt templates are prepared in advance. This prompt generation logic generates an appropriate prompt appropriate for the current stage of the order process, and sends it to the large-scale language model unit 113.

[0069] The order prompt generation unit 115 also includes a prompt transmission interface for transmitting the generated prompt to the large-scale language model unit 113. The order prompt generation unit 115 also includes a state management function for managing the current stage and progress of the order process. The state management function tracks the current stage and progress of the order process and provides this information to the prompt generation logic. This allows the prompt generation logic to determine which stage a prompt should be generated for and select an appropriate prompt template. For example, if the order process is in the "idea generation stage," the prompt generation logic selects a template containing questions and instructions related to idea generation. Furthermore, by embedding information and variables related to the current situation in the selected template, the logic can generate a prompt appropriate for the current stage and determine whether to transition to the next stage based on the response information from the user 4.

[0070] (Ordering support device 1: Overall configuration) As described above, the order support device 1 includes an order database 121 that stores order-related information related to the order process from receiving an assignment from the user 4 to placing an order with a vendor, an initial setting storage unit 16 that stores prerequisite information that instructs the device to have a conversation with the user 4 according to the order stages that indicate each stage of the order process, an information receiving unit 111 that receives response information from the user 4 in the conversation with the user 4, and a combination of one or more types of natural language processing based on the order-related information in the order database 121 to probabilistically predict how likely words and sentences given in a prompt are to occur in natural language, and generate order support information to be presented to the user 4 in the order process by sentence generation, question and answering, etc. The large-scale language model unit 113 has a large-scale language model specialized for creating order placement and receipt support information, and is trained and adjusted to generate order placement and receipt support information according to the preconditions, an initial setting prompt generation unit 114 includes precondition information in the initial setting storage unit 16 in the prompt of the large-scale language model at the start of the order placement and receipt process, thereby instructing that a conversation in accordance with the order placement and receipt stage be held with the user 4 as a precondition for generating the order placement and receipt support information in the large-scale language model, and an order placement and receipt prompt generation unit 115 includes response information received by the information receiving unit 111 after the start of the order placement and receipt process in the prompt of the large-scale language model unit 113, thereby causing the large-scale language model to generate order placement and receipt support information that is suitable for the order placement and receipt stage based on the preconditions.

[0071] According to the above configuration, at the start of the order process, the large-scale language model specialized for creating order support information is initially set to have a conversation according to the order stage, and by including response information from the conversation with user 4 after the order process has started in the prompts, it becomes possible to fully support order activities such as finding a supplier just like having a conversation with a human concierge, that is, by simply conducting simple interviews and natural language exchanges without having to write prompts for user 4 to input into the large-scale language model. This makes it possible to easily perform order activities with sufficient support throughout the entire order process from the large-scale language model specialized for creating order support information.

[0072] The order placement and receipt support device 1 may be configured to provide an API (Application Programming Interface) service for incorporating it into other companies' systems (marketplaces, SaaS (Software as a Service), intranets, Slack (registered trademark), etc.).

[0073] Furthermore, according to the above configuration, as shown in FIG. 2, the order support device 1 uses order-related information such as user information and vendor information in the order database 121 having each database 1211 to 1216, and the large-scale language model unit 113 learns, and the response information of the information receiving unit 111 corresponds to an explanatory variable of machine learning, and in response to prompts from the initial setting prompt generation unit 114 and the order prompt generation unit 115, the large-scale language model unit 113 generates order support information as an objective variable of machine learning.

[0074] As a result, the order placement and receipt support device 1 receives response information and generates order placement and receipt support information suitable for the order placement and receipt process using a large-scale language model specialized for creating order placement and receipt support information, thereby making it possible to fully support order placement and receipt activities. In particular, by using a large-scale language model instead of ordinary machine learning to generate order placement and receipt support information, it is possible to generate new content and variations not included in the training data, making it possible to generate flexible and diverse order placement and receipt support information according to the task content and situation, as well as to generate creative content not included in the training data.

[0075] Furthermore, because the accuracy of large-scale language models can be improved through learning, it is possible to learn from past successes and continuously generate more effective order support information. In other words, it is possible to analyze the results of the order process and generate order support information that reflects improvements, or to reflect feedback from users 4 and vendors and generate order support information that provides even greater satisfaction to users 4.

[0076] In this way, the order placement and receipt support device 1 uses a large-scale language model to provide advanced functions that are difficult to achieve with conventional machine learning approaches. Specifically, it utilizes the unique capabilities of large-scale language models, such as complex text generation using natural language processing and context-based content generation. Conventional machine learning models typically perform prediction, classification, clustering, and other tasks using primarily numerical and categorical data, and are limited in their ability to process large amounts of text data and generate new text. On the other hand, large-scale language models are adept at learning language patterns from large amounts of text data and generating new text based on given prompts. Therefore, in situations that require the use of diverse and complex language, such as order placement and receipt support information, an approach using a large-scale language model is more appropriate, and it is possible to provide functions that are difficult to achieve with conventional machine learning methods alone.

[0077] As shown in FIG. 3, the order placement and receipt support device 1 includes an information receiving unit 111, a large-scale language model unit 113, an initial setting prompt generating unit 114, an order placement and receipt prompt generating unit 115, an external call prompt generating unit 112, a communication control unit 116, a terminal control unit 117, and a communication unit 13. Each unit 111 to 117 except for the communication unit 13 is included in the control unit 11, which is a computer. Some or all of the units included in the control unit 11 may be configured as either hardware or software. Input to the information receiving unit 111 is performed via an input unit 22, such as a keyboard, of the user terminal 2 operated by the user 4. Furthermore, the order placement and receipt support information created by the large-scale language model unit 113 is displayed on an output unit 21, such as a display, of the user terminal 2.

[0078] The communication unit 13 is connected to the user terminal 2 operated by the user 4, the administrator terminal 7 operated by the administrator, and the vendor terminal 3 operated by the vendor's staff via the Internet 5 so that data can be transmitted between them. The administrator operating the administrator terminal 7 monitors the interactions between the user terminal 2 and the order placement and receipt support device 1 and can intervene as a human concierge to provide support depending on the situation. This enables a concierge service in which the order placement and receipt support device 1 and a human collaborate. Data communication between the communication unit 13 and the user terminal 2 and the administrator terminal 7 is not limited to the Internet 5. It may also be possible to use an information communication network such as a dedicated line or a local area network (LAN), or Bluetooth (registered trademark), a wireless communication standard for short-range communication. Data communication between the communication unit 13 and the user terminal 2 and the administrator terminal 7 may also be via a dedicated line to prevent information leakage. Furthermore, when data communication is performed via the Internet 5, the communication unit 13, the user terminal 2, and the administrator terminal 7 preferably have a virtual private network (VPN) function. If a VPN function is incorporated into the communication unit 13, the user terminal 2, and the administrator terminal 7, for example, data communication from the user terminal 2 passes through a VPN connection, so that the data is protected from external unauthorized access and secure communication is ensured. The user terminal 2, the administrator terminal 7, and the vendor terminal 3 are information processing devices such as general personal computers, laptop computers, smartphones, and tablet terminals.

[0079] The terminal control unit 117 has a user interface function that sets the display screen of the user terminal 2 to a screen suitable for viewing order support information and inputting response information. Specifically, as shown in Fig. 4, the terminal control unit 117 forms a user interface on the display screen of the user terminal 2, including a summary screen 211 that summarizes and displays information necessary to check the response information and order support information at each stage of the order process, a consultation box screen 212 for the user 4 to input response information, and a support box screen 213 that displays the order support information.

[0080] For example, the consultation box screen 212 displays a message such as, "Problem: The user is facing the challenge of declining sales due to intensifying competition in online sales with traditional sales strategies. Requirement: Approaching new customer segments through advertising and increasing brand awareness..." Then, when the user terminal 2 inputs a message by voice or keyboard, such as, "Please introduce me to a web advertising vendor that suits my business," this message is displayed on the consultation box screen 212, and then the "matching results" are displayed on the support box screen 213 as order support information from the large-scale language model unit 113. In this way, the communication control unit 116, through its user interface function, allows the user 4 to effectively grasp the information necessary for the order process and facilitates appropriate responses and decisions through conversation with the order support device 1.

[0081] As shown in FIG. 3 , the order placement and receipt support device 1 includes an input device 15 connected to an input receiving unit 1192 and a display device 14 connected to a display control unit 1191. Examples of the input device 15 include a keyboard, a mouse, a touch panel, and a voice input device. Examples of the display device 14 include a liquid crystal display device. This allows the order placement and receipt support device 1 to be configured using information processing devices such as general personal computers, laptop computers, smartphones, and tablet terminals. The order placement and receipt support device 1 may lack at least one of the input device 15 and the display device 14. In this case, the input device 15 and the display device 14 are provided in an external terminal (not shown), allowing the order placement and receipt support device 1 to be used as an order placement and receipt support information creation server. Furthermore, the order placement and receipt support device 1 may be used as a user terminal 2, and order placement and receipt support information may be created through conversation with the user 4 while the user 4 receives support information generated by the large-scale language model unit 113.

[0082] In this embodiment, the case where the functions of the order placement and receipt support device 1 are installed in an information processing device is described, but the present invention is not limited to this, and the functions of the order placement and receipt support device 1 may be installed in cloud computing. In this case, cloud computing allows computer resources to be added as needed, making it possible to process large amounts of unique operation information and common operation information, enabling faster and more efficient processing, and also making it possible to easily expand processing capacity to accommodate a significant increase in the number of users 4.

[0083] The order placement and receipt support device 1 also has a communication unit 13 capable of communicating with external terminals including the vendor terminal 3, a communication control unit 116 that controls the communication unit 13 based on order placement and receipt support information, and an external call prompt generation unit 112 that, when communication with an external terminal is established, causes the large-scale language model unit 113 to generate order placement and receipt support information suitable for calls with the external terminal by including call information indicating the content of the call with the external terminal in a prompt. Here, examples of the "external terminal" include personal terminals held by vendor personnel or users 4, servers of vendor companies, cloud platforms that provide specific services such as dining reservation services or travel reservation services, POS systems used in restaurants and hotels, and order management systems used by large retail chains.

[0084] The communication control unit 116 has a function of managing communication functions within the order placement and receipt support device 1 and appropriately controlling communication with external terminals. Specifically, the communication control unit 116 has a communication interface management function that manages the communication interface (wired or wireless) and sends and receives data in cooperation with the communication unit 13, a data transmission and reception management function that manages received data and sent data and maintains data consistency and security, a communication protocol management function that supports different communication protocols and formats data in accordance with these protocols, a communication security management function that ensures the security of communication data, and a communication log management function that records communication details.

[0085] The external call prompt generation unit 112 has a function of, when communication with an external terminal is established, analyzing the content of the call and providing a prompt for generating appropriate order placement and receipt support information to the large-scale language model unit 113. Specifically, it has a call information receiving interface function of receiving call data such as the content of the call, tone, order of remarks, and keywords from the external terminal, a call content analysis module function of analyzing the received call data using natural language processing technology and voice recognition technology to understand the content and topic of the call, prompt generation logic of creating a prompt appropriate for generating order placement and receipt support information based on the analysis result, a prompt transmission interface function of transmitting the generated prompt to the large-scale language model unit 113, and a communication control cooperation function of cooperating with the communication control unit 116 and appropriately controlling communication with the external terminal via the communication unit 113.

[0086] According to the above configuration, the order support device 1 is able to communicate with external terminals using the order support information of the large-scale language model, making it possible for user 4 to make inquiries to companies, vendors, and service providers that provide the tools selected by user 4, and to provide secretarial services for the complicated tasks that user 4 has to handle, such as making dinner reservations, arranging business trips, arranging conference rooms, and ordering consumables, etc.

[0087] The order placement support device 1 also has a function of creating a request for proposal when placing an order with a vendor. Specifically, the initial setting storage unit 16 stores request creation instruction information that instructs the creation of a request for proposal for officially conveying requirements organization information including business details, requirements, and order conditions to the vendor based on the content of a conversation with the user 4 at the order placement stage, and the initial setting prompt generation unit 114 is configured to include the request creation instruction information in the initial setting storage unit 16 in the prompt of the large-scale language model, thereby causing the large-scale language model to generate a request for proposal as order placement support information.

[0088] A Request for Proposal (RFP) is a formal document that solicits proposals from vendors for a specific project or service, and is created to provide vendors with detailed information necessary for their proposals, such as the project scope, requirements, expected deliverables, and submission deadline.

[0089] According to the above configuration, by including request creation instruction information in the prompt of the large-scale language model unit 113, the large-scale language model unit 113 extracts necessary information from the conversation with the user and automatically generates a request for proposal as order placement and receipt support information based on that information. Then, for example, as shown in FIG. 5, the request for proposal is sent to the selected vendor after being confirmed by the user 4 via the user terminal 2. This reduces the burden of creating a request for proposal and allows the order placement and receipt process to proceed more smoothly and efficiently.

[0090] In addition, the order support device 1 is configured such that the order database 121 stores user information related to the user 4 and vendor information related to the vendor as order-related information, the initial setting memory unit 16 stores recommended vendor proposal information that instructs the device to propose recommended vendors that will be promising partner candidates for the user 4 based on the user information, vendor information, and searched vendor information searched via the Internet 5 when the order stage is the candidate vendor research stage, and the initial setting prompt generation unit 114 includes the recommended vendor proposal information in the initial setting memory unit 16 in the prompt of the large-scale language model, thereby causing the large-scale language model to generate the recommended vendors proposed in the candidate vendor research stage as order support information.

[0091] According to the above configuration, the order placement and receipt support device 1 automatically suggests recommended vendors by utilizing user information, vendor information, and searched vendor information, thereby reducing the time and effort required for the user 4 to manually search for and evaluate vendors, enabling the user 4 to quickly select an appropriate vendor. In addition, a large-scale language model generates order placement and receipt related information based on the recommended vendor suggestion information stored in the order placement and receipt storage unit 12, enabling reliable decision-making based on data, thereby increasing the probability of selecting a vendor that best meets the needs of the user 4. Furthermore, because the reasons for selecting a recommended vendor are based on data, the transparency of the process is improved, making it easier to understand why a particular vendor was recommended, and this gives the user 4 confidence in the selection process.

[0092] The recommended vendor proposal information may also include a specific method for selecting a recommended vendor. For example, the recommended vendor proposal information may include instructions to select a recommended vendor who will be the optimal partner based on selection factors such as the strengths, track record, reputation, and price of the contractors in the vendor information and searched vendor information. Furthermore, the recommended vendor proposal information may include instructions to select multiple candidate vendors, such as 50 companies, analyze whether the vendor information of these vendors matches the desired conditions, and output an evaluation list with a rating of "yes" or "no" along with comments explaining the reasons for the evaluation.

[0093] In addition, the order placement and purchase support device 1 stores, in the initial setting memory unit 16, request form instruction information that instructs the creation of a request form for internal approval by the vendor to be contracted when the order placement and purchase stage is the ordering stage, and the initial setting prompt generation unit 114 is configured to include the request form instruction information in the initial setting memory unit 16 in the prompt of the large-scale language model, thereby causing the large-scale language model to generate the request form created in the ordering stage as order placement and purchase support information.

[0094] With the above configuration, the automatic generation of approval documents can significantly reduce the time required for manual document creation and minimize errors in documents caused by human input errors or inconsistent information. Furthermore, the rapid creation of approval documents can accelerate the approval process and shorten the lead time for projects and ordering activities.

[0095] (Order Database 121) Next, the order database 121 will be described in detail. The order database 121 is stored in the order storage unit 12. The order storage unit 12 is connected to the control unit 11 so that data can be communicated therewith. The order storage unit 12 may be configured with a hard disk, or may be configured with a combination of a hard disk and memory. In the case of a configuration that combines a hard disk and memory, some of the data and indexes used by the database can be cached in the memory as needed, thereby speeding up access to the database. The order storage unit 12 may be a data server that is connected to an information communication network such as the Internet 5, separate from the order support device 1. The order storage unit 12 may also be configured with multiple data servers, one for each database.

[0096] As shown in FIG. 2, the order database 121 includes a user database 1211 , a vendor database 1212 , an item information database 1213 , a chat log database 1214 , a business support database 1215 , and other databases 1216 .

[0097] As shown in FIG. 6 , the user database 1211 stores basic user information as well as transaction-related information in the form of a table. Specifically, the database has fields such as company name, industry, number of employees, issue area, order details, required conditions, and desired conditions. Each field is associated with a corresponding user element, creating a single user record. For example, if the company ID is "C001," the database stores user information including user elements such as company name: ABC Co., Ltd., industry: manufacturing, number of employees: 200, issue area: marketing, order details: new product promotion planning, required conditions: track record of competitors, and desired conditions: low budget. Furthermore, if the company ID is "C002," the database stores user information including user elements such as company name: XYZ Trading, industry: wholesale, number of employees: 50, issue area: human resources and labor, order details: introduction of a personnel evaluation system, required conditions: track record of small and medium-sized enterprises, and desired conditions: remote support available.

[0098] Registration of user information in the user database 1211 is usually performed at the introduction stage of the order placement and receipt support device 1 or when a user 4 first uses the order placement and receipt support device 1. For example, when the order placement and receipt support device 1 is introduced, when a new user 4 begins to use the order placement and receipt support device 1, at the start of a project or transaction, when information is updated periodically, or when the user 4 self-reports and provides information.

[0099] 7, vendor database 1212 has fields such as a vendor ID, company name, location, contact information, and website, and stores vendor elements associated with each field to form one piece of vendor information. For example, if the vendor ID is "Vendor A," user information including user elements such as company name A Co., Ltd., location Shibuya-ku, Tokyo, contact information 03-1234-5678, website wwwA, representative name Taro Yamada, date of establishment January 1, 2000, capital 100 million yen, number of employees 100, business content software development, main product / service CRM system, track record / implementation case studies implementation to major companies, technical capabilities / expertise AI development, customer list Company X, Company Y, certification / qualification ISO9001, etc. is stored.

[0100] Vendor information is registered in the vendor database 1212 at various times, such as when a new vendor is registered, when information on an existing vendor is updated, when vendor information is periodically reviewed, and during audits and compliance checks to verify the completeness and accuracy of vendor information.

[0101] The AIMITSU information database 1213 stores a wide range of information to support the order placement and receipt process. For example, the AIMITSU information database 1213 stores data on vendor performance evaluations by industry, market trends, and supply chain efficiency. Specifically, the AIMITSU information database 1213 stores performance information such as each vendor's past project success rate and feedback evaluations from clients (users 4), data on the transparency and efficiency of the vendor's supply chain, and price data.

[0102] The chat log database 1214 is a database that stores all communication records of exchanges between the order placement and receipt support device 1 and the user 4. The chat log database 1214 stores detailed records of the content of chat conversations between the user 4 and the order placement and receipt support device 1. For example, this information includes the date and time, the identification information of the user 4, and the text content of the messages exchanged. Specifically, the information stored in the chat log database 1214 includes all interactive exchanges, such as the content of the user 4's inquiries, the order placement and receipt support information that is the response from the order placement and receipt support device 1, and feedback and follow-up questions from the user 4.

[0103] The business support database 1215 stores information related to manuals and know-how to support the improvement of the efficiency and quality of the order placement and receipt process. Specifically, the business support database 1215 stores information such as detailed explanations of various business procedures, best practices for business execution, troubleshooting guidelines, and techniques and strategies for improving business efficiency. The business support database 1215 also aggregates and stores past case studies and project reviews, making them available for reference when similar situations arise. This allows the business support database 1215 to learn from previously successful methods and examples of failure to be avoided, enabling appropriate responses to problem solving.

[0104] Other databases 1216 store regulatory and compliance information important to meeting legal requirements in the order process, such as industry-specific regulations, legal compliance requirements, and safety standards, market research data such as consumer behavior analysis, competitor trends, and market growth forecasts, and information such as emergency response plans and crisis management protocols. In this way, other databases 1216 can provide a variety of information to address various special needs in the order process and to assist in training large-scale language models.

[0105] (Ordering activity support program) 3, in the order placement and receipt support device 1, the information receiving unit 111, the external call prompt generating unit 112, the large-scale language model unit 113, the initial setting prompt generating unit 114, the order placement and receipt prompt generating unit 115, the communication control unit 116, and the terminal control unit 117 may be configured as either hardware or software. Each of these units 111 to 117 constitutes at least a part of the control unit 11. When each of the units 111 to 117 is configured as software, the control unit 11 is configured to execute an order placement and receipt activity support program that generates order placement and receipt support information by utilizing a language model for natural language processing specialized for creating order placement and receipt support information.

[0106] As a specific example, the order placement and receipt activity support program includes an order placement and receipt database 121 storing order placement and receipt related information relating to the order placement and receipt process from the acceptance of the assignment by the user 4 to the placement of an order to the vendor, an initial setting storage unit 16 storing prerequisite information instructing the program to have a conversation with the user 4 according to the order placement and receipt stages indicating each stage of the order placement and receipt process, and a large-scale language model unit having a large-scale language model specialized for creating order placement and receipt support information by combining one or more types of natural language processing based on the order placement and receipt related information in the order placement and receipt database 121, and learning and adjustment are performed so as to generate order placement and receipt support information to be presented to the user 4 in the order placement and receipt process by sentence generation and question answering. 113, the computer (control unit 11) executes an information receiving process step (S1) of receiving response information from user 4 in a conversation with user 4, an initial setting prompt generation process step (S2) of including prerequisite information in the initial setting storage unit 16 in a prompt of the large-scale language model at the start of the order-receiving process to instruct, as a prerequisite for generating order-receiving support information in the large-scale language model, that a conversation in accordance with the order-receiving stage be held with user 4, and an order prompt generation process step (S3) of including the response information received in the information receiving process step (S1) in a prompt of the large-scale language model unit 113 after the start of the order-receiving process, to generate order-receiving support information suited to the order-receiving stage based on the prerequisite. The order-receiving activity support program may be configured to cause the computer to execute the functions of each unit 111 to 117 in the control unit 11 as processing steps.

[0107] According to the above configuration, at the start of the order process, the large-scale language model specialized for creating order support information is initially set to have a conversation according to the order stage, and by including response information from the conversation with user 4 after the order process has started in the prompts, it becomes possible to fully support order activities such as finding a supplier just like having a conversation with a human concierge, that is, by simply conducting simple interviews and natural language exchanges without having to write prompts for user 4 to input into the large-scale language model. This makes it possible to easily perform order activities with sufficient support throughout the entire order process from the large-scale language model specialized for creating order support information.

[0108] Furthermore, simply by installing the order placement and receipt activity support program in an information processing device such as a personal computer or tablet terminal, the information processing device can function as the order placement and receipt support device 1. The program may be distributed in a state recorded on a computer-readable recording medium such as a CD-ROM or USB memory, or may be distributed via a two-way communication network or communication line such as the Internet or a one-way communication network such as a television broadcast.

[0109] (Ordering activity support method) The order placement and receipt support device 1 is configured to cause a computer (control unit 11) to execute an order placement and receipt activity support method. Specifically, the order placement and receipt support device 1 includes an order placement and receipt database 121 storing order placement and receipt related information relating to the order placement and receipt process from the acceptance of the task by the user 4 to the placement of an order to the vendor, an initial setting storage unit 16 storing prerequisite information instructing the device to have a conversation with the user 4 according to the order placement and receipt stages indicating each stage of the order placement and receipt process, and by combining one or more types of natural language processing based on the order placement and receipt related information in the order placement and receipt database 121, the device probabilistically predicts how likely words and sentences given in the prompts are to occur in natural language, and learning and adjustment are performed to generate order placement and receipt support information to be presented to the user 4 in the order placement and receipt process by sentence generation and question answering, thereby providing a large-scale system with a large-scale language model specialized for creating order placement and receipt support information. The method is for causing a computer (control unit 11) having a large-scale language model unit 113 to execute an information reception process for receiving response information from user 4 in a conversation with user 4, an initial setting prompt generation process for instructing, at the start of the order process, that a conversation in accordance with the order stage be held with user 4 as a prerequisite for generating order support information in the large-scale language model by including prerequisite information in the initial setting memory unit 16 in the prompt of the large-scale language model, and an order prompt generation process for causing the large-scale language model to generate order support information suitable for the order stage based on the prerequisite by including, in the prompt of the large-scale language model unit 113, the response information received in the information reception process after the start of the order process.

[0110] It should be noted that within the scope of the concept of the present invention, those skilled in the art may conceive of various modifications and alterations. Therefore, it is understood that such modifications and alterations fall within the scope of the present invention. For example, those skilled in the art may appropriately add, delete, or modify components of the above-described embodiments, or may add, omit, or change the conditions of processes, as long as they maintain the essence of the present invention. [Explanation of symbols]

[0111] 1. Order placement support device 2. User terminal 3 Vendor terminal 4 User 5. Internet 7 Administrator terminal 11 Control section 12 Order storage unit 13 Communications Department 14 Display device 15 Input Devices 16 Initial setting storage section 111 Information Reception Department 112 External call prompt generation unit 113 Large-scale Language Modeling 114 Initialization prompt generator 115 Order Prompt Generation Unit 116 Communication control unit 117 Terminal control unit

Claims

1. an order database storing order-related information relating to the order process from receiving a user's assignment to placing an order with a vendor; an initial setting storage unit storing prerequisite information instructing the conversation with the user to be conducted in accordance with the order placement and receipt stages indicating the respective stages of the order placement and receipt process; an information receiving unit that receives response information of the user in a conversation with the user; a large-scale language model unit that has a large-scale language model specialized for creating the order support information by combining one or more types of natural language processing to probabilistically predict how likely words and sentences given in prompts are to occur in natural language based on the order-related information in the order database, and by performing learning and adjustment to generate order support information to be presented to the user in the order process by sentence generation and question answering; an initial setting prompt generation unit that, at the start of the order placement and receipt process, includes the prerequisite information of the initial setting storage unit in the prompt of the large-scale language model, thereby instructing that a conversation in accordance with the order placement and receipt stage be held with the user as a prerequisite for generating the order placement and receipt support information in the large-scale language model; an order prompt generation unit that causes the large-scale language model to generate the order support information suitable for the order stage based on the prerequisites by including the response information received by the information receiving unit after the start of the order process in the prompt of the large-scale language model unit; An order placement and receipt support device having the above.

2. The initial setting storage unit request creation instruction information for instructing the creation of a request for proposal for formally conveying to the vendor information on requirements organization including business details, requirements, and order conditions based on the content of the conversation with the user at the order placement and receipt stage is stored; The initial setting prompt generation unit The request creation instruction information of the initial setting storage unit is included in the prompt of the large-scale language model, thereby causing the large-scale language model to generate the request for proposal as the order placement and receipt support information.

2. The ordering support device according to claim 1.

3. The order database includes: user information relating to the user and vendor information relating to the vendor are stored as the order-receiving related information; The initial setting storage unit When the order placement stage is a candidate vendor research stage, recommended vendor proposal information is stored, which instructs the system to propose recommended vendors that are likely to be potential partner candidates for the user based on the user information, the vendor information, and searched vendor information searched via the Internet, The initial setting prompt generation unit The recommended vendor proposal information of the initial setting storage unit is included in the prompt of the large-scale language model, so that the recommended vendor proposed in the candidate vendor research stage is generated in the large-scale language model as the order placement and receipt support information.

2. The ordering support device according to claim 1.

4. The initial setting storage unit When the order placement stage is an order placement stage, approval document instruction information is stored to instruct the vendor to prepare an approval document for internal approval of the vendor to be contracted, The initial setting prompt generation unit The request form instruction information of the initial setting storage unit is included in the prompt of the large-scale language model, and the request form created in the ordering stage is generated in the large-scale language model as the order placement and receipt support information.

2. The ordering support device according to claim 1.

5. a communication unit capable of communicating with an external terminal; A communication control unit (normal machine learning) that controls the communication unit based on the order placement and receipt support information; an external call prompt generation unit that, when communication with the external terminal is established, causes the large-scale language model unit to generate the order placement and receipt support information suitable for the call with the external terminal by including call information indicating the content of the call with the external terminal in the prompt of the large-scale language model unit; The ordering support device according to claim 1 , further comprising:

Images