AI-based business support system

Abstract

The objective is to obtain a business support system that utilizes AI, which allows chatbots (sites with automatic conversation programs equipped with AI functions) to be used at low cost, even when using existing business systems that use chat. [Solution] A business support system 1 that utilizes AI connects a user terminal 10, a chatbot site 30, and a business support site 50 via a communication network 60, and provides support from the chatbot site 30 to the user terminal 10. The user terminal 10 has a means for generating a chat screen, and the business support site 50 has a means for generating business support information previously discussed with the user, receiving user instruction information including questions from the user terminal, and transmitting user instruction prompts to the chatbot site 30 based on the business support information previously discussed with the user.

Description

[Technical Field]

[0001] The present invention relates to a business support system that utilizes AI. [Background technology]

[0002] In recent years, payroll calculations, attendance management, sales management, etc. have been computerized, and these tasks are often performed through interactive chat (sites that use natural language processing programs) (also known as existing business systems).

[0003] Chat is a system that allows users to exchange mainly text data in real time on the Internet. Patent Document 1 discloses a business support service system that uses chat.

[0004] Patent Document 1 determines an intent having an action name and a web hook that specify a processing operation corresponding to a conversational sentence, and further selects an external database via an API that uses as a parameter information for executing the processing operation specified by these entities, intents, etc. The above-mentioned intent is a mechanism for connecting functions within application software and taking over processing.

[0005] In other words, when a user wants to perform some operation on an object displayed on the screen, the app notifies the system. The system then searches for apps and functions within apps that can perform the operation on that object and presents them to the user as options. If the user finds a function they want to execute from the presented options, they can select it and execute it on the spot.

[0006] Meanwhile, chatbots (programs that automate certain tasks and processes in place of humans; also known as automated conversation programs with AI functions) have also come into use in recent years. Chatbots are programs in which computers incorporating artificial intelligence converse on behalf of humans.

[0007] For example, when making restaurant, taxi, hotel, or airline reservations, you can simply ask a chatbot, just like a secretary or butler, and when shopping on an e-commerce site, instead of the traditional rule-based recommendations, you can tap to respond to the chatbot's questions and suggestions to find the perfect product for your current mood. Patent Document 2 discloses a system using this chatbot. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Patent Publication No. 2021-68399 [Patent Document 2] Patent No. 7340063 Summary of the Invention [Problem to be solved by the invention]

[0009] However, the system in Patent Document 1 replies to questions with predetermined answers in text format, and in order to interactively obtain appropriate answers that correspond to the client's business style, the database data must be set up to include answers that correspond to the client's business style, which requires cost and effort.

[0010] On the other hand, Patent Document 2 uses a chatbot, but introducing a chatbot as is requires costs and man-hours. In other words, requesting a chatbot to execute a process requires costs and man-hours.

[0011] The present invention was made in consideration of the above-mentioned problems, and aims to provide a business support system that utilizes AI, which allows users to use chatbots (sites with automatic conversation programs equipped with AI functions) at low cost, even when using existing business systems that use chat. [Means for solving the problem]

[0012] The present invention has been made in consideration of the above problems, and the invention of claim 1 is a business support system that utilizes AI, which connects a user terminal, a chatbot site, and a business support site via a communication network and provides support from the chatbot site to the user terminal, wherein the user terminal has a means for generating a chat screen, the business support site has a means for generating business support information pre-negotiated with the user, receiving user instruction information including questions and requests from the user terminal, and for each reception, sending a user instruction prompt to the chatbot site based on the business support information pre-negotiated with the user, and the chatbot site has a chatbot output information sending unit that generates answers to the user instruction prompts in chat format data and sends it from the chatbot site to the user terminal and displays it on the chat screen.

[0013] The invention of claim 2 is characterized in that the business support site has a chatbot selection instruction unit that selects the chatbot site that can accomplish the user's business and sends the user instruction prompt to the chatbot site.

[0014] The invention according to claim 3 is characterized in that the user instruction prompt is input to the chatbot site activated based on an API (Application Programming Interface).

[0015] The invention according to claim 4 is characterized in that only a chat screen is displayed on the screen of the user terminal.

[0016] The invention according to claim 5 is characterized in that it comprises an RDB site, and the RDB site is a relational database. [Effects of the Invention]

[0017] As described above, according to the present invention, even if an existing business system using chat is used, it is possible to obtain a business support system that utilizes AI and allows the use of chatbots (sites for automatic conversation programs with AI functions) at low cost. [Brief explanation of the drawings]

[0018] The drawings illustrate specific embodiments of the invention according to the present disclosure, including essential features of the invention as well as alternative and preferred embodiments. [Figure 1] This is a schematic diagram of a business support system that utilizes AI. [Figure 2] This is a diagram showing the architecture of a business support system that utilizes AI. [Figure 3] This is an explanatory diagram of the blocks of a business support system that utilizes AI. [Figure 4] This is a sequence diagram of a business support system that utilizes AI. [Figure 5] FIG. 10 is an explanatory diagram of a chat screen. [Figure 6] FIG. 10 is an explanatory diagram of a chat screen relating to end-of-month processing. [Figure 7] FIG. 10 is an explanatory diagram of a chat screen related to handling an invoice. [Figure 8] FIG. 10 is an explanatory diagram of a chat screen related to inquiry management. [Figure 9] FIG. 10 is an explanatory diagram of a chat screen related to an EC site. [Figure 10] FIG. 10 is an explanatory diagram of a chat screen related to the sales management system. [Figure 11] FIG. 10 is an explanatory diagram of a chat screen related to the store management system. [Figure 12] This is an explanatory diagram when the API is public. [Figure 13] This is an explanatory diagram for the case where the API is not public. DETAILED DESCRIPTION OF THE INVENTION

[0019] Each embodiment will be described in detail below with reference to the accompanying drawings. In these examples, a description of already known technologies will be omitted. Furthermore, the following examples illustrate devices and methods for embodying the technical concept of the invention, and the technical concept of the present invention is not limited to the following. Various modifications can be made to the technical concept of the present invention within the scope of the claims. It should be noted that the drawings are schematic and may differ from the actual product.

[0020] FIG. 1 is a schematic diagram of a business support system 1 utilizing AI according to this embodiment. The AI-based business support system 1 of this embodiment is a system that connects a user terminal 10 (a personal computer or smartphone), a chatbot site 30 (a server system), an RDB site 40 (a relational database site), and a business support site 50 (including a server 52, a personal computer 54, etc.) via a cloud computing network 60 (Internet Protocol). The chatbot site 30 (comprised of a generative AI (artificial intelligence)—the same applies below) is inexpensively introduced into an existing business system, enabling interactive business instructions to be given to the chatbot site 30. More specifically, the chatbot site 30 not only allows users to ask questions and receive answers but also allows them to execute processes. In other words, by having the chatbot site 30 inexpensively perform the complex and tedious operations of existing business systems, business improvement and cost reduction are possible. In other words, all screens use chat screens, eliminating the need to create multiple new screens. In other words, all processing is completed using only the chat screen. For example, it becomes possible to ask the chatbot site 30 a question or have the chatbot site 30 perform a task.

[0021] The development cycle is shortened by introducing a generation AI support function into the existing system, listening to the business operations in order to have the generation AI work in the existing system, and implementing the necessary API before introducing the chatbot site 30. This allows for small-scale development to be started.

[0022] If the RDB can be referenced from the container of the AI-powered business support system 1, it is possible to link with an AI chatbot. There are no restrictions on the type of RDB. In addition, if an API is provided as a SaaS service, it is also possible to link with an AI chatbot. The AI-powered business support system 1 can be installed as a chatbot by embedding a script into an existing website, or by developing a chatbot in parallel with the existing system when developing a new system.

[0023] This concept is shown as an architecture in Figure 2. As shown in Figure 2, a user operates a user terminal 10 to input messages (prompts) such as questions, instructions for processing work, etc. on a chat screen, which are received by a server 52 of a business support site 50 as a front-end and sent as a request to an API. This causes the chatbot site 30 to process the generated AI, and the server 52 of the business support site 50 acquires the results and displays them in a predetermined format on the chat screen of the user terminal 10. At this time, a request is made to the API and acquired using a protocol based on that API.

[0024] In other words, the control of the human conversation portion is left to the generation AI (such as OpenAI (registered trademark)) API, and the AI ​​is instructed to call the appropriate Golang API processing.If the Golang API is a web system that already has a REST API implemented, the Golang API can be combined and called to reduce the Golang API implementation workload.

[0025] The server 52 of the business support site 50 has the program configuration shown in Fig. 3. As shown in Fig. 3, the server 52 of the business support site 50 includes a memory 522 (included in the RDB site 40) that stores user information, a memory 523 (included in the RDB site 40) that stores information necessary for business, a business consultant information processing unit 524 that registers business consultant information, which is information for executing optimally customized business, a chatbot selection instruction unit 526 that selects an appropriate chatbot site 30, and a user instruction prompt sending unit 529 that sends prompts from the user terminal 10. The user information in the memory 522 is stored by various information registration units (not shown).

[0026] The various information registration section provides a screen that allows a company employee to enter account information (ID, password), company ID, name, address, user code, business type, telephone number, email address, etc. when accessed from the user terminal 10, and stores this information in memory 522.

[0027] Furthermore, the various information registration unit transmits information exchanged with the user terminal 10 (such as invoices and sales figures) to the RDB site 40 and stores it.

[0028] The business consultant information processing unit 524 executes simplified processing based on the consultant information including business information previously discussed with the user, unnecessary items, and repeated items.

[0029] The chatbot selection instruction unit 526 selects the chatbot site 30 that is most suitable for accomplishing the user's task, and sends a user instruction prompt, which will be described later, to this chatbot site 30.

[0030] The user instruction prompt sending unit 529 receives user instruction information including questions and requests about work from the user terminal 10, and each time it receives it, generates a user instruction prompt by editing this user instruction information based on the consultant information.Then, it sends this to the chatbot site 30, and causes the chatbot site 30 to send this user instruction prompt as chat format data to the user terminal 10.

[0031] Fig. 4 is a sequence diagram illustrating the AI-based business support system 1. Note that, although there are multiple types of chatbot site 30 and RDB site 40 depending on the business, etc., in Fig. 4, they are shown as one.

[0032] The company's personnel operates the user terminal 10 to access the business support site 50 and register user information (d10). The user information is registered by a various information registration unit (not shown).

[0033] This user information consists of an account (ID, password), company ID (group ID), name, address, user code, type of business, telephone number, email address, registration number, etc. (only company ID and email address are acceptable).

[0034] Furthermore, the various information registration unit causes the user terminal 10 to transmit information exchanged in the AI-based business support system 1 and stores it in the RDB site 40 (d11a, d11b). This information is associated with user information.

[0035] For example, based on this information, information on problems in the current business of each user is accumulated, making it possible to build a business that solves the problems in the existing system.

[0036] Next, the various information registration unit of the server 52 of the business support site 50 associates the user information with the date, time, user code, reception number, etc., and registers it in the memory 522 (d12).

[0037] The business consultant information processing unit 524 of the server 52 of the business support site 50 generates consultant information that compiles unnecessary and repetitive items obtained by, for example, interviewing the company about its existing business operations such as invoices and sales management (existing systems) into the minimum necessary items (d14).

[0038] Meanwhile, the chatbot selection instruction unit 526 reads out this consultant information from the memory 523 and selects an appropriate chatbot site 30 (d16, d18: chatbot selection instruction process).

[0039] That is, a chatbot site 30 is selected from the RDB site 40 (d20), and information (URL, also called appropriate chatbot information) of the selected chatbot site 30 (there may be many) is sent to the business support site 50 (d22). This completes the preparation (d24).

[0040] The following explains how to start the chatbot site 30. An API (Application Programming Interface) is software that runs on a PC or smartphone (using a programming language to issue instructions to the PC or smartphone).

[0041] In other words, an API is a programming interface used when two applications or pieces of software exchange information.

[0042] By the way, there are two patterns for using APIs: "developing it yourself" and "using a provided service." In this embodiment, we will upgrade our own company's service using a provided API.

[0043] "I" refers to the interface used when software exchanges information with other software. More specifically, it defines the protocols and connection specifications used when connecting applications to each other.

[0044] That is, the server 52 of the business support site 50 listens to, for example, frequently used task programs in programs (existing systems) such as invoicing, inventory management, and sales management, identifies these task programs, and has already determined the chatbot site 30 that can execute these task programs.

[0045] Then, the user operates the user terminal 10 to access the business support site 50 and transmit the inquiry information (d30).

[0046] The business support site 50 receives the user instruction information (prompt) and compares it with the consultant information stored in the memory 523 to generate information (user instruction prompt) that will provide the minimum necessary answer to a question for the user of this company's group. Then, this is stored in the memory 523 (d32) and sent to the chatbot site 30 (d34).

[0047] The chatbot site 30 receives this user instruction prompt (question) and selects an answer from the Internet environment or the RDB site 40 (d36). At this time, the chatbot site 30 uses machine learning to learn what kind of answer is appropriate for this question in the future.

[0048] Then, this selection result is transmitted to the user terminal 10 (for history purposes) via the business support site 50 and displayed thereon (d40, d42).

[0049] That is, the business support site 50 uses the chatbot site 30 as a hub.

[0050] Furthermore, the chatbot site 30 will learn through machine learning what kind of answer will be appropriate for this question in the future by receiving user-instructed editing information.

[0051] Examples of screens displayed on the user terminal 10 are shown in FIGS.

[0052] See Figure 5. First, the information "Hello. Please enter your question" is displayed on the screen of the user terminal 10 based on the information sent from the chatbot site 30. The information "Please tell me which clients have not yet confirmed payments this month" is input to the user terminal 10. Then, the chatbot site 30 generates an answer and displays a list of clients on the screen of the user terminal 10.

[0053] As shown in Figure 6, in the case of end-of-month processing, if the user asks "Please do the usual end-of-month processing for this month!", the chatbot site 30 will respond with "Is the following okay for Suzuki's end-of-month processing? 1. Extract contract companies that need to process invoices this month 2. Enter all the necessary information for billing 3. Create an invoice and email it to the person in charge." Furthermore, if the user requests, "When you send an email, please include me in CC," the chatbot site 30 responds, "I have included Suzuki in CC and have completed sending the email."

[0054] That is, with the above configuration, a screen other than the usual one is provided to the user, with unnecessary information removed.

[0055] In the case of an invoice, as shown in Figure 7, when the user terminal 10 makes a request saying "I received the invoice," the chatbot site 30 responds with "Please upload an image of the invoice." The user terminal 10 replies with "Uploaded."

[0056] Chatbot Site 30 responds, "We have confirmed that AAA Co., Ltd.'s August invoice is consistent with the purchase order. Would you like us to register it in our customer management system?" In other words, it omits any unnecessary details from the response.

[0057] In the case of inquiry management, as shown in Figure 8, when a user asks, "How many inquiries were there about product XXX yesterday?", the chatbot site 30 responds, "The following inquiries were received. Type: Question 55, Type: Other 18."

[0058] Furthermore, in response to a user request such as "Type: Please reply saying that you will answer the question at a later date," the chatbot site 30 responds with "I understand. Is the following text okay? Thank you for your inquiry. We have received many inquiries regarding this matter, and the relevant department will provide a response at a later date. ..." In other words, only necessary responses are given.

[0059] See Figure 9. In the case of an e-commerce site (customer experience), in response to a question from the user terminal 10 saying, "I bought a shirt here the other day. I'd like to buy some matching pants as well," the chatbot site 30 displays the following information on the screen of the user terminal 10: "Thank you as always. The shirt you purchased is product number XXX, blue, size L, correct?" The user terminal 10 replies, "Yes, that's right." In response, the chatbot site 30 displays on the screen of the user terminal 10: "There are 18 recommended products. Would you like to see a list of all of them? Or, if you provide more detailed criteria, we can narrow down your search."

[0060] See Figure 10. This is the case of a sales management system. When the user terminal 10 asks, "Please tell me which products among this week's top 20 sales are below the reorder point," the chatbot site 30 responds, "Products with inventory below the reorder point (inventory quantity): cushion cover (10), pillow cover (20), kitchen mat (15)." The user terminal 10 instructs, "Thank you. Please tell me the contact information for each supplier." The chatbot site 30 responds, "The suppliers for each product are as follows."

[0061] In the case of a store management system, as shown in Figure 11, when a user asks, "Tell me which stores have seen a sales increase of 10% or more from last month," the chatbot site 30 responds with, "Sales increase: Store A +18%; Store B +13%; Sales decrease: Store X -11%." In other words, the chatbot site 30 learns and answers the minimum items required by the company.

[0062] The user terminal 10 issues an instruction saying, "I'm going to have a meeting with the managers of the three stores mentioned above, and Mr. Suzuki. Please register it on your calendar." The chatbot site 30 asks, "Please tell me the specific date and time of the meeting."

[0063] That is, if the API is publicly available, it is added to the outside of the existing system as shown in Figure 12. If the API is not publicly available, it is added to the API of the existing system as shown in Figure 13.

[0064] This allows the AI ​​to repeatedly learn from a huge amount of pre-registered data, improving the accuracy of its answers (eliminating unnecessary information). Because it finds answers to questions based on what it has learned, it is capable of natural conversations even on advanced topics.

[0065] In addition, because it is a cloud-based system, there is no need to install servers or hardware, so preparations for implementation are relatively simple and initial costs are low, which offers great benefits for companies.

[0066] Although each embodiment has been described in detail above, it is not limited to the specific embodiment, and various modifications and changes are possible within the scope of the claims. It is also possible to combine all or a plurality of components of the above-described embodiments. [Explanation of symbols]

[0067] 1. AI-based business support system 10 User terminal 30 Chatbot Sites 40 RDB sites 50 Business support sites 60 Cloud Network (Communication Network)

Claims

1. A business support system utilizing AI that connects a user terminal, a chatbot site, and a business support site via a communication network and provides support from the chatbot site to the user terminal, the user terminal comprises means for generating a chat screen, The business support site has a means for generating business support information pre-agreed with the user, receiving user instruction information including questions and requests from the user terminal, and transmitting user instruction prompts to the chatbot site based on the business support information pre-agreed with the user each time the information is received; The chatbot site is characterized by having a chatbot output information sending unit that generates answers to the user instruction prompts in chat format data, sends them from the chatbot site to the user terminal, and displays them on the chat screen.

2. The business support system utilizing AI as described in claim 1, characterized in that the business support site has a chatbot selection instruction unit that selects the chatbot site that can accomplish the user's business and sends the user instruction prompt to the chatbot site.

3. 2. The business support system utilizing AI according to claim 1, wherein the user instruction prompt is input to the chatbot site activated based on an API (Application Programming Interface).

4. The business support system utilizing AI according to claim 1, characterized in that only a chat screen is displayed on the screen of the user terminal.

5. 2. The business support system utilizing AI according to claim 1, further comprising an RDB site, the RDB site being a relational database.

Images