Intelligent customer service methods and systems, computer-readable recording media

Abstract

An intelligent customer service method involves an AI module of an intelligent customer service system receiving a question from a user's terminal device. If the AI ​​module determines that a corresponding answer exists in a question-and-answer database, it provides the answer to the user's terminal device. Furthermore, a first intent analysis module of the intelligent customer service system, based on the question and historical question-and-answer records stored in the database, infers at least one possible intent following the question. It then retrieves at least one answer related to the at least one possible intent from the database and provides the at least one possible intent and its related answer to the AI ​​module. The AI ​​module then combines the at least one possible intent and its related answer to generate and provide potential demand information to the user's terminal device.

Description

Technical Field

[0001] This invention relates to an online customer service method, and more particularly to an intelligent customer service method that applies artificial intelligence. Prior Technology

[0002] Existing online text-based customer service systems typically only provide corresponding answers to customer questions, i.e., one question and one answer. Therefore, customers need to keep asking questions and ask the right questions in order to get the answers they want. As a result, customers need to spend a lot of time communicating with the online customer service system, and may not even get the answers they want in the end.

[0003] Furthermore, the customer questions received by online text customer service systems are quite diverse. As a result, intelligent customer service systems often encounter many unknown questions that cannot be identified, leading to the end of online customer service (or the transfer to human service). Afterwards, such unknown questions usually require manual review of the online customer service system's dialogue (question and answer) records to identify the problem, and then adding the customer question (question) and its corresponding answer to the online customer service system's question and answer database. However, the number of unknown questions is quite considerable (at least 1,500 per day). If each one is reviewed and processed manually, it is not only impossible for ordinary human resources to handle, but it is also time-consuming, labor-intensive, and cannot effectively improve the service efficiency of intelligent customer service. Summary of the Invention

[0004] Therefore, the purpose of this invention is to provide an intelligent customer service method, an intelligent customer service system for implementing the method, and a computer-readable recording medium, which at least solves the problems of the prior art mentioned above. In addition to optimizing the processing efficiency of online intelligent customer service for unknown questions and answers, reducing the burden of manual work, and effectively improving the service efficiency and quality of intelligent customer service.

[0005] Therefore, the present invention provides an intelligent customer service method, applied to an intelligent customer service system capable of communicating with a user terminal device, and includes the following steps: a customer service AI module of the intelligent customer service system receives a question from the user terminal device, and when it determines that a first answer corresponding to the question exists in a question-and-answer database, the customer service AI module provides the first answer to the user terminal device; and a first intent analysis module of the intelligent customer service system, based on the question and the historical question-and-answer records stored in the question-and-answer database, infers at least one first possible intent following the question, obtains at least one first answer data related to the at least one first possible intent from the question-and-answer database, and provides the at least one first possible intent and the at least one first answer data related to it to the customer service AI module, so that the customer service AI module combines the at least one first possible intent and the at least one first answer data related to it to generate and provide a first possible demand information to the user terminal device.

[0006] In some embodiments of the present invention, when the first intent analysis module cannot obtain the at least one first answer information related to the at least one first possible intent from the question and answer database, the first intent analysis module classifies the at least one first possible intent and stores it in a category database; and a data update module of the intelligent customer service system accepts the at least one first answer information related to the at least one first possible intent, adds the at least one first answer information to the question and answer database, and associates the at least one first possible intent with the at least one first answer information.

[0007] In some embodiments of the present invention, when the customer service AI module determines that there is no answer corresponding to the question in the question-and-answer database, it determines that the question is an unknown question and stores an unknown question-and-answer record containing the unknown question in an unknown question-and-answer database; and the intelligent customer service method further includes the following steps: a second intent analysis module of the intelligent customer service system reads the unknown question-and-answer record from the unknown question-and-answer database, and based on the unknown question-and-answer record and the historical question-and-answer record stored in the question-and-answer database, infers a possible question of the unknown question and at least one second possible intent following the possible question; when the second intent analysis module determines that there is a second answer corresponding to the possible question in the question-and-answer database, it adds the possible question to the question-and-answer database and makes it correspond to the second answer; when the second intent analysis module determines that there is at least one second answer data related to the at least one second possible intent in the question-and-answer database, it associates the at least one second possible intent with the at least one second answer data stored in the question-and-answer database.

[0008] Therefore, when the customer service AI module receives the possible question from the user terminal device, the customer service AI module can obtain the second answer corresponding to the possible question from the question and answer database and provide it to the user terminal device. The first intent analysis module, based on the possible question and the historical question and answer records stored in the question and answer database, infers the at least one second possible intent following the possible question, and can obtain at least one second answer related to the at least one second possible intent from the question and answer database, and provide the at least one second possible intent and its related at least one second answer to the customer service AI module. The customer service AI module then combines the at least one second possible intent and its related at least one second answer to generate and provide a second possible demand information to the user terminal device.

[0009] In some embodiments of the present invention, when the second intent analysis module determines that there is no second answer corresponding to the possible question in the question-and-answer database, a data update module of the intelligent customer service system accepts the second answer corresponding to the possible question, and adds the possible question and the second answer to the question-and-answer database and makes the possible question correspond to the second answer.

[0010] In some embodiments of the present invention, when the second intent analysis module determines that there is no at least one second answer related to the at least one second possible intent in the question and answer database, it classifies the at least one second possible intent and stores it in a category database of its own; and the intelligent customer service method further includes the following steps: a data update module of the intelligent customer service system receives the at least one second answer related to the at least one second possible intent, adds the at least one second answer to the question and answer database, and associates the at least one second possible intent with the at least one second answer.

[0011] In some embodiments of the present invention, a second intent analysis module of the intelligent customer service system further determines, based on the historical question-and-answer records stored in the question-and-answer database, that when a question-and-answer record contains an answer corresponding to a question that is irrelevant, the second intent analysis module stores the question-and-answer record in an irrelevant answer database, and a data update module of the intelligent customer service system accepts a correct answer corresponding to the question and replaces the existing answer corresponding to the question in the question-and-answer database with the correct answer.

[0012] Furthermore, the present invention provides an intelligent customer service system that implements the above-described method. This system can communicate with a user terminal device and includes a communication unit, a storage unit, and a processing unit. The communication unit can establish a communication channel with the user terminal device. The storage unit has a question-and-answer database. The processing unit is electrically connected to the communication unit and the storage unit to communicate with the user terminal device through the communication unit and access the storage unit. The processing unit includes a customer service AI module and a first intent analysis module. The processing unit implements the intelligent customer service method described above using the customer service AI module and the first intent analysis module.

[0013] Furthermore, the present invention provides a computer-readable recording medium for implementing the above method, wherein a software program is stored comprising a customer service AI module, a first intent analysis module, a data update module and a second intent analysis module. When the software program is loaded and executed by a computer device, the computer device can perform the intelligent customer service method described above.

[0014] The advantages of this invention are as follows: In addition to providing corresponding answers to questions, it also infers the user's possible intentions based on the questions and provides the user with relevant information in a timely manner, helping the user obtain the information they need without having to ask further questions, thus improving the user's customer service experience; Furthermore, for unknown questions, it infers the possible questions and their subsequent possible intentions by referring to historical question and answer records, and promptly supplements or adds answers to possible questions and information corresponding to possible intentions to the system, thereby optimizing the efficiency of online intelligent customer service in handling unknown questions and answers, reducing the burden of manual work, and effectively improving the service efficiency and quality of intelligent customer service. Simple Explanation of the Diagram

[0015] Other features and effects of the present invention will be clearly shown in the embodiments with reference to the drawings, wherein: Figures 1 and 2 illustrate the main process steps of an embodiment of the intelligent customer service method of the present invention; and Figure 3 is a schematic diagram of the components included in an embodiment of the intelligent customer service system of the present invention. Implementation

[0016] Before the invention is described in detail, it should be noted that similar elements are represented by the same numbers in the following description.

[0017] 1 and 2 are the flow steps of an embodiment of an intelligent customer service method of the present invention, which is implemented by an intelligent customer service system 1 shown in FIG.

[0018] As shown in FIG. 3 , the intelligent customer service system 1 mainly includes a communication unit 11 , a storage unit 12 and a processing unit 13 , the communication unit 11 is a communication interface (e.g., a network card) and can be connected to the user terminal device 2 via, for example, an Internet network; A collective term for the logging media (e.g., memory module, hard disk, etc.) taken, which has a Q&A database 121 containing Q&A data used to respond to a user's question (question) and its corresponding answer (answer) and a historical Q&A record recording previous Q&A processes in response to a user. In addition, the storage unit 12 stores a software program comprising a customer service AI module 131 , a first intent analysis module 132 , a data update module 133 , and a second intent analysis module 134 .

[0019] The processing unit 13 is a processor having computing functions, such as a central processor, which is electrically connected to the communication unit 11 and the storage unit 12 to communicate with and access the storage unit 12 through the communication unit 11 . And the processing unit 13 implements the method flow shown in FIG.

[0020] Thereby, the processing unit 13 runs the software program, as in step S1 of FIG Since the question of the user terminal device 2 , the customer service AI module 131 determines whether a first answer corresponding to the question exists in the question and answer database 121 in step S2 of FIG. 1. If, in step S3 of FIG.

[0021] Next, as shown in step S4 of Figure 1, the first intent analysis module 132 infers at least one possible first intent of the user based on the question and the historical question-and-answer records stored in the question-and-answer database. For example, when the user's question is: "My online banking password is locked, how do I unlock it?", the first intent analysis module 132 can infer that the user's next possible intent is to use functions such as checking account balance, transferring funds, and paying bills after solving the online banking password lock problem, based on the historical question-and-answer records stored in the question-and-answer database 121 (e.g., records of other users who solved the online banking password lock problem and then used functions such as checking account balance, transferring funds, and paying bills). Then, as shown in step S5 of Figure 1, the first intent analysis module 132 determines whether the question-and-answer database 121 contains at least one first answer related to the at least one possible first intent (e.g., checking account balance, transferring funds, paying bills, etc.) (e.g., what steps should be taken). (If the operation is to use and complete functions such as checking account balance, transferring funds, paying bills, etc.), as shown in step S6 of Figure 1, the first intent analysis module 132 obtains at least one first answer information related to the at least one first possible intent from the question and answer database 121, and provides the at least one first possible intent and its related at least one first answer information to the customer service AI module 131. Thus, as shown in step S7 of Figure 1, the customer service AI module 131 will combine the at least one first possible intent and its related at least one first answer information to generate and provide a first possible demand information to the user terminal device 2.

[0022] Therefore, in addition to receiving answers to their questions, users also gain access to information about their primary potential needs. This information provides options for functions that users might want to use, such as checking account balances, transferring funds, or making payments. Users can then directly select a desired function from these options provided by the primary potential needs information without having to ask questions again. Once a function option is selected, the primary potential needs information will then provide the corresponding usage steps and / or operation methods to assist the user in using that function option (e.g., making payments). This improves the service efficiency of the intelligent customer service and increases user satisfaction with online customer service.

[0023] Furthermore, in step S5 above, when the first intent analysis module 132 determines that there is no at least one first answer related to the at least one first possible intent in the question-and-answer database 121, that is, when the first intent analysis module 132 cannot obtain the at least one first answer related to the at least one first possible intent from the question-and-answer database 121, as shown in step S8 of FIG1, the first intent analysis module 132 classifies the at least one first possible intent (for example, by referring to the historical question-and-answer records and question-and-answer data in the question-and-answer database 121 to determine the responsible unit to which the possible intent belongs) and stores it in a category database 122 located in the storage unit 12; for example, when the first possible intent is, for example, tax payment, the first intent analysis module 132 stores the first possible intent (tax payment) in the category database 122 related to tax payment business.

[0024] In this way, personnel involved in tax payment can later check the category database 122 to find out which online services and answer materials related to tax payment need to be added (i.e., the first answer material corresponding to the first possible intention (tax payment)), and upload the first answer material corresponding to the first possible intention (tax payment) to the intelligent customer service system 1 through a terminal computer device (not shown).

[0025] As shown in step S9 of Figure 1, the data update module 133 in the processing unit 13 will receive the first answer data related to the first possible intention (tax payment), and add the first answer data to the question and answer database 121 to associate the first possible intention with the first answer data; thereby, the first intention analysis module 132 can obtain the first answer data related to the first possible intention (tax payment) from the question and answer database 121 next time.

[0026] Furthermore, in step S2 above, if the customer service AI module 131 determines that there is no first answer corresponding to the question in the question-and-answer database 121, that is, the first intent analysis module 132 cannot obtain the answer corresponding to the question from the question-and-answer database 121, the customer service AI module 131 will reply with a message such as "no relevant information found" or "suggest transferring to a real customer service representative" to the user terminal device 2. As shown in step S10 of Figure 1, the customer service AI module 131 determines the question as an unknown question and stores an unknown question-and-answer record containing the unknown question (that is, the process of the customer service AI module 131 responding to the question) in an unknown question-and-answer database 123 in the storage unit 12.

[0027] Accordingly, as shown in step S11 of Figure 2, the second intent analysis module 134 of the processing unit 13 can then read the unknown question and answer record from the unknown question and answer database 123, and based on the unknown question and answer record and the historical question and answer record stored in the question and answer database 121, infer a possible question for the unknown question and at least one second possible intent that the user wants after the possible question. For example, assuming the question (unknown question) is "How to transfer money through online banking?", the second intent analysis module 134 will infer that the possible question for the unknown question is "How to apply for online banking" and that the second possible intent that the user wants after the possible question is "transfer money through online banking". Then, as shown in step S12 of Figure 2, the second intent analysis module 134 determines whether there is a second answer corresponding to the possible question in the question and answer database 121. If so, as shown in step S13 of Figure 2, the possible question (i.e., extended question) is added to the question and answer database 121 and made to correspond to the second answer.

[0028] If, in step S12, the second intent analysis module 134 determines that there is no second answer corresponding to the possible question in the question-and-answer database 121, as shown in step S14 of Figure 2, the second intent analysis module 134 categorizes the possible question (how to apply for online banking) (for example, by referring to historical question-and-answer records and data in the question-and-answer database 121 to determine the responsible unit to which the possible question belongs) and stores it in its respective category database 122. In this way, personnel related to the possible question (how to apply for online banking) can view the category database 122 to find out the possible question and provide the corresponding second answer accordingly, and upload the corresponding second answer to the intelligent customer service system 1 through a terminal computer device (not shown).

[0029] Therefore, as shown in step S15 of Figure 2, the data update module 133 will accept the second answer corresponding to the possible question, and add the possible question and the second answer to the question and answer database 121 and make the possible question correspond to the second answer.

[0030] As shown in step S16 of Figure 2, the second intent analysis module 134 determines whether there is a second answer data related to the second possible intent (e.g., transferring money via online banking) in the question and answer database 121. If so, as shown in step S17 of Figure 2, the possible intent (transferring money via online banking) is associated with the second answer data stored in the question and answer database 121. The second answer data includes information such as the relevant operation methods and / or operation steps (or processes) for transferring money via online banking.

[0031] If, in step S16, the second intent analysis module 134 determines that the question-and-answer database 121 does not contain the second answer information related to the second possible intent (via online banking transfer), then, as shown in step S18 of Figure 2, the second intent analysis module 134 categorizes the second possible intent (for example, by referring to historical question-and-answer records and data in the question-and-answer database 121 to determine the responsible unit to which the possible intent belongs) and stores it in the category database 122 located in the storage unit 12. In this way, personnel related to online banking transfer services can view the category database 122 to determine the need to add online services and answer information related to online banking transfers (i.e., the second answer information corresponding to the second possible intent (online banking transfer)), and upload the second answer information corresponding to the second possible intent (online banking transfer) to the intelligent customer service system 1 via a terminal computer device (not shown).

[0032] As shown in step S19 of Figure 2, the data update module 133 in the processing unit 13 will receive the second answer data related to the second possible intention (online banking transfer), and add the second answer data to the question and answer database 121 to make the second possible intention related to the second answer data.

[0033] Therefore, when the customer service AI module 131 receives the possible question (how to apply for online banking) from the user terminal device 2, the customer service AI module 131 can obtain the second answer corresponding to the possible question from the question and answer database 121 and provide it to the user terminal device 2. At the same time, the first intent analysis module 132, based on the possible question and the historical question and answer records stored in the question and answer database 121, infers the at least one second possible intent (transferring money through online banking) that follows the possible question. Then, it can obtain at least one second answer related to the at least one second possible intent (transferring money through online banking) from the question and answer database 121 and provide the at least one second possible intent and its related at least one second answer to the customer service AI module 131. This allows the customer service AI module 131 to combine the at least one second possible intent and its related at least one second answer to generate and provide a second possible demand information to the user terminal device 2.

[0034] Furthermore, the second intent analysis module 134, based on the historical question-and-answer records stored in the question-and-answer database 121, determines if an answer to a question in one of the question-and-answer records is irrelevant (e.g., the user's evaluation of the answer is irrelevant or unhelpful). The second intent analysis module 134 then stores the question-and-answer record in an irrelevant answer database 124 within the storage unit 12. This allows relevant personnel to check the irrelevant answer database 124 to confirm whether the answer to the question in the question-and-answer record needs correction, and upload the correct answer to the question to the intelligent customer service system 1 via a terminal computer device (not shown). Therefore, the data update module 133 accepts the correct answer to the question and replaces the existing answer to the question in the question-and-answer database with the correct answer.

[0035] It is worth mentioning that the first intent analysis module 132 and the second intent analysis module 134 use generative AI models to assist in predicting possible questions for the aforementioned unknown questions, as well as predicting possible subsequent intents for known questions and possible questions; and the customer service AI module 131 uses Retrieval Augmented Generation (RAG) technology to combine the aforementioned at least one first (second) possible intent and its related at least one first (second) answer data to provide the user terminal device 2 with more fluent information content regarding the first (second) possible demand. Since RAG technology is a known technology and not the focus of this case, it will not be described in detail here. The aforementioned generative AI model can be, for example, but not limited to, one of the following:

[0036] 1. Generative Adversarial Networks (GANs): This model consists of two neural networks, a generator and a discriminator, that compete with each other to generate more realistic data. GANs are commonly used for image generation and enhancement, but can also be applied to text generation and customer intent inference.

[0037] 2. Generative Pre-trained Transformer (GPT): A large-scale language model developed by OpenAI capable of generating high-quality text. GPT models can be used to analyze customers' language patterns to infer their intentions and needs.

[0038] 3. Variational Autoencoders (VAEs): These models generate new data by learning the latent representations of existing data. VAEs can be used to generate text and images, and also have applications in inferring customer intent.

[0039] 4. Diffusion Model: This is an emerging generative model that generates data through progressive denoising. Diffusion models excel in image generation but are also beginning to be applied to text generation and customer intent inference.

[0040] Therefore, based on the characteristics and application scenarios of these models, enterprises can choose the appropriate model according to their own needs to infer customer intentions and improve service quality and customer satisfaction.

[0041] In summary, the intelligent customer service system 1 of the above embodiments, in addition to providing answers to user questions, further infers the user's possible intentions based on the questions and promptly provides the user with information related to those intentions. This helps users obtain the necessary information without further questioning, thus improving their customer service experience. Furthermore, for unknown questions raised by users, the system infers possible questions and their subsequent possible intentions by referring to historical question-and-answer records, supplementing or adding answers to possible questions and information related to possible intentions to the system. This optimizes the efficiency of online intelligent customer service in handling unknown questions and answers, reduces the burden of manual operations, and effectively improves the service efficiency and quality of intelligent customer service, truly achieving the effects and objectives of the present invention.

[0042] However, the above description is merely an embodiment of the present invention and should not be construed as limiting the scope of the present invention. Any simple equivalent changes and modifications made in accordance with the scope of the patent application and the contents of the patent specification shall still fall within the scope of the patent of the present invention.

[0043] 1: Intelligent Customer Service System 11: Communication Unit 12: Storage Unit 121: Question and Answer Database 122: Category Database 123: Unknown Question and Answer Database 124: Irrelevant Answer Database 13: Processing Unit 131: Customer Service AI Module 132: First Intent Analysis Module 133: Data Update Module 134: Second Intent Analysis Module 2: User terminal device S1~S19: Steps

Claims

1. An intelligent customer service method, applied to an intelligent customer service system capable of communicating with a user terminal device, comprising the following steps: A customer service AI module of the intelligent customer service system receives a question from the user terminal device, and when it determines that a first answer corresponding to the question exists in a question-and-answer database, the customer service AI module provides the first answer to the user terminal device; and a first intent analysis module of the intelligent customer service system, based on the question and historical question-and-answer records stored in the question-and-answer database, infers at least one first possible intent following the question, obtains at least one first answer related to the at least one first possible intent from the question-and-answer database, and provides the at least one first possible intent and its related at least one first answer to the customer service AI module, so that the customer service AI module combines the at least one first possible intent and its related at least one first answer to generate and provide a first possible demand information to the user terminal device; wherein... When the customer service AI module determines that there is no answer corresponding to the question in the question-and-answer database, it determines that the question is an unknown question and stores an unknown question-and-answer record containing the unknown question in an unknown question-and-answer database. The intelligent customer service method further includes the following steps: A second intent analysis module of the intelligent customer service system reads the unknown question-and-answer record from the unknown question-and-answer database and, based on the unknown question-and-answer record and the historical question-and-answer records stored in the database, infers a possible question for the unknown question and at least one subsequent second possible intent for the possible question; When the second intent analysis module determines that there is a second answer corresponding to the possible question in the question-and-answer database, it adds the possible question to the database and makes it correspond to the second answer; When the second intent analysis module determines that there is at least one second answer related to the at least one second possible intent in the question-and-answer database, it establishes a connection between the at least one second possible intent and the at least one second answer stored in the question-and-answer database. Therefore, when the customer service AI module receives a possible question from the user terminal device, the customer service AI module can obtain the second answer corresponding to the possible question from the question and answer database and provide it to the user terminal device. The first intent analysis module, based on the possible question and the historical question and answer records stored in the question and answer database, infers the at least one second possible intent following the possible question, and can obtain at least one second answer information related to the at least one second possible intent from the question and answer database, and provide the at least one second possible intent and its related at least one second answer information to the customer service AI module. This allows the customer service AI module to combine the at least one second possible intent and its related at least one second answer information to generate and provide a second possible demand information to the user terminal device.

2. The intelligent customer service method as described in request item 1, wherein, When the first intent analysis module cannot obtain the at least one first answer information related to the at least one first possible intent from the question and answer database, the first intent analysis module classifies the at least one first possible intent and stores it in a category database of its own; and a data update module of the intelligent customer service system accepts the at least one first answer information related to the at least one first possible intent, adds the at least one first answer information to the question and answer database and associates the at least one first possible intent with the at least one first answer information.

3. The intelligent customer service method as described in request item 1, wherein, When the second intent analysis module determines that there is no second answer corresponding to the possible question in the question and answer database, a data update module of the intelligent customer service system accepts the second answer corresponding to the possible question, and adds the possible question and the second answer to the question and answer database and makes the possible question correspond to the second answer.

4. The intelligent customer service method as described in request item 1, wherein, When the second intent analysis module determines that there is no at least one second answer related to the at least one second possible intent in the question and answer database, it classifies the at least one second possible intent and stores it in a category database of its own; and the intelligent customer service method further includes the following steps: a data update module of the intelligent customer service system receives the at least one second answer related to the at least one second possible intent, adds the at least one second answer to the question and answer database, and associates the at least one second possible intent with the at least one second answer.

5. The intelligent customer service method as described in request item 1, wherein, The second intent analysis module of the intelligent customer service system also determines, based on the historical question-and-answer records stored in the question-and-answer database, that when a question-and-answer record contains an answer that is irrelevant to the question, the second intent analysis module stores the question-and-answer record in an irrelevant answer database. Furthermore, the data update module of the intelligent customer service system accepts a correct answer that corresponds to the question and replaces the existing answer corresponding to the question in the question-and-answer database with the correct answer.

6. An intelligent customer service system capable of communicating with a user terminal device, comprising: A communication unit capable of establishing a communication channel with the user terminal device; a storage unit having a question-and-answer database; The system includes a processing unit electrically connected to the communication unit and the storage unit, for communicating with the user terminal device through the communication unit and accessing the storage unit. The processing unit includes a customer service AI module and a first intent analysis module. The customer service AI module receives a question from the user terminal device and, if it determines that a first answer corresponding to the question exists in the question-and-answer database, provides the first answer to the user terminal device. The first intent analysis module, based on the question and the historical question-and-answer records stored in the question-and-answer database, infers at least one possible first intent following the question, retrieves at least one first answer related to the at least one possible first intent from the question-and-answer database, and provides the at least one possible first intent and its related at least one first answer to the customer service AI module. This allows the customer service AI module to combine the at least one possible first intent and its related at least one first answer to generate and provide a first possible demand information to the user terminal device. When the customer service AI module determines that there is no answer to a question in the question-and-answer database, it classifies the question as an unknown question and stores an unknown question-and-answer record containing the unknown question in an unknown question-and-answer database in the storage unit. The processing unit also includes a second intent analysis module. This second intent analysis module reads the unknown question-and-answer record from the unknown question-and-answer database and, based on the unknown question-and-answer record and the historical question-and-answer records stored in the database, infers a possible question for the unknown question and at least one subsequent second possible intent. When the second intent analysis module determines that there is a second answer in the question-and-answer database corresponding to the possible question, it adds the possible question to the database and assigns it the second answer. When the second intent analysis module determines that there is at least one second answer related to the at least one second possible intent in the question-and-answer database, it establishes a connection between the at least one second possible intent and the at least one second answer stored in the question-and-answer database. Therefore, when the customer service AI module receives a possible question from the user terminal device, the customer service AI module can obtain the second answer corresponding to the possible question from the question and answer database and provide it to the user terminal device. The first intent analysis module, based on the possible question and the historical question and answer records stored in the question and answer database, infers the at least one second possible intent following the possible question, and can obtain at least one second answer information related to the at least one second possible intent from the question and answer database, and provide the at least one second possible intent and its related at least one second answer information to the customer service AI module. This allows the customer service AI module to combine the at least one second possible intent and its related at least one second answer information to generate and provide a second possible demand information to the user terminal device.

7. The intelligent customer service system as described in claim 6, wherein, When the first intent analysis module cannot obtain the at least one first answer data related to the at least one first possible intent from the question and answer database, the first intent analysis module classifies the at least one first possible intent and stores it in the category database to which the at least one first possible intent belongs in the storage unit; and a data update module included in the processing unit accepts the at least one first answer data related to the at least one first possible intent, adds the at least one first answer data to the question and answer database, and associates the at least one first possible intent with the at least one first answer data.

8. The intelligent customer service system as described in claim 6, wherein, When the second intent analysis module determines that there is no second answer corresponding to the possible question in the question-and-answer database, the second intent analysis module classifies the possible question and stores it in a category database to which the possible question belongs in the storage unit. The data update module included in the processing unit accepts the second answer corresponding to the possible question and adds the possible question and the second answer to the question-and-answer database and makes the possible question correspond to the second answer.

9. The intelligent customer service system as described in claim 6, wherein, When the second intent analysis module determines that there is no at least one second answer data related to the at least one second possible intent in the question and answer database, the second intent analysis module classifies the at least one second possible intent and stores it in a category database to which the at least one second possible intent belongs in the storage unit; and a data update module included in the processing unit receives the at least one second answer data related to the at least one second possible intent, adds the at least one second answer data to the question and answer database, and associates the at least one second possible intent with the at least one second answer data.

10. The intelligent customer service system as described in claim 6, wherein, The processing unit further includes a second intent analysis module and a data update module. The second intent analysis module determines, based on the historical question-and-answer records stored in the question-and-answer database, that when a question-and-answer record contains an answer that is irrelevant to the question, the second intent analysis module stores the question-and-answer record in an irrelevant answer database. The data update module accepts a correct answer that corresponds to the question and replaces the existing answer corresponding to the question in the question-and-answer database with the correct answer.

11. A computer-readable recording medium storing a software program comprising a customer service AI module, a first intent analysis module, a data update module and a second intent analysis module, wherein when the software program is loaded and executed by a computer device, the computer device is capable of performing the intelligent customer service method as described in any one of claims 1 to 5.

Images